Stability of Functional Equations in Banach Algebras

Yeol Je Cho · Choonkil Park Themistocles M. Rassias · Reza Saadati Stability of Functional Equations in Banach Algebras...

Author: Yeol Je Cho | Choonkil Park | Themistocles M. Rassias | Reza Saadati

11 downloads 683 Views 3MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Yeol Je Cho · Choonkil Park Themistocles M. Rassias · Reza Saadati

Stability of Functional Equations in Banach Algebras


Yeol Je Cho • Choonkil Park Themistocles M. Rassias • Reza Saadati


123

Yeol Je Cho Department of Mathematics Education and the RINS Gyeongsang National University College of Education Jinju, Korea, Republic of South Korea Themistocles M. Rassias Department of Mathematics National Technical University of Athens Athens, Greece

ISBN 978-3-319-18707-5 DOI 10.1007/978-3-319-18708-2

Choonkil Park Department of Mathematics Hanyang University Seoul, Korea, Republic of South Korea Reza Saadati Department of Mathematics Iran University of Science and Technology Tehran, Iran

ISBN 978-3-319-18708-2 (eBook)

Library of Congress Control Number: 2015939079 Mathematics Subject Classification (2010): 39-XX, 26-XX, 41-XX, 46-XX, 47-XX Springer Cham Heidelberg New York Dordrecht London © Springer International Publishing Switzerland 2015 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. Printed on acid-free paper Springer International Publishing AG Switzerland is part of Springer Science+Business Media (www. springer.com)

To S.M. Ulam for the 75th anniversary of his stability problem for approximate homomorphisms

Preface

The main purpose of this book is to present some of the old and recent results on homomorphisms and derivations in Banach algebras, quasi-Banach algebras, C -algebras, C -ternary algebras, non-Archimedean Banach algebras, and multinormed algebras. In 1940, S. M. Ulam [321] proposed a stability problem on group homomorphisms in metric groups. In 1941, D. H. Hyers [133] proved the stability of additive mappings in Banach spaces associated with the Cauchy equation. In 1978, Th. M. Rassias [267] proved the stability of R-linear mappings associated with the Cauchy equation, and in 2002 C. Park [220] proved the stability of C-linear mappings in the spirit of Hyers, Ulam, and Rassias in Banach modules. Homomorphisms and derivations in Banach algebras, quasi-Banach algebras, C -algebras, C ternary algebras, non-Archimedean Banach algebras and multi-normed algebras are additive and R-linear or C-linear, and so we study the stability problems for additive functional equations and additive mappings. Using the direct method and the fixed point method, the authors have studied the stability and the superstability of homomorphisms and derivations in Banach algebras, quasi-Banach algebras, C -algebras, C -ternary algebras, non-Archimedean Banach algebras, and multinormed algebras, which are also associated with additive functional equations and additive functional inequalities. The book provides a survey of both the latest and new results especially on the following topics: (1) Stability theory for several new functional equations in Banach algebras and C -algebras via fixed point method and direct method. (2) Stability theory for several new functional inequalities in Banach algebras and C -algebras via fixed point method and direct method. (3) Stability theory of well-known new functional equations in non-Archimedean Banach algebras and non-Archimedean C -algebras. (4) Stability theory for several new functional equations and functional inequalities in multi-Banach algebras and multi-C -algebras via fixed point method and direct method. vii

viii

Preface

The book is intended to be accessible especially to graduate students who have a basic background with operator theory, functional analysis, functional equations, and analytic inequalities including an introduction to Banach algebras, quasi-Banach algebras, C -algebras, C -ternary algebras, non-Archimedean Banach algebras, and multi-normed algebras. In Chap. 1, we provide a brief introduction to concepts with historic remarks for functional equations and their stability and the definitions of Banach algebras, quasi-Banach algebras, C -algebras, C -ternary algebras, non-Archimedean Banach algebras, and multi-normed algebras. In Chap. 2, we study the stability of additive functional equations in Banach spaces as well as the stability and the superstability of isomorphisms, homomorphisms, derivations, and generalized derivations in Banach algebras and quasiBanach algebras associated with additive functional equations. In Chap. 3, we study the stability and the superstability of isomorphisms, homomorphisms, and derivations in C -algebras, Lie C -algebras, and JC -algebras, as well as the stability and the superstability of linear mappings in Banach modules over unital C -algebras. Moreover, we study Jordan -derivations, quadratic Jordan -derivations, .˛; ˇ; /-derivations on Lie C -algebras, square root functional equations, 3rd root functional equations, and positive-additive functional equations. In Chap. 4, we study the stability of C-linear mappings in Banach spaces and linear mappings in normed modules over a C -algebra as well as the stability of homomorphisms and derivations in proper CQ -algebras associated with functional inequalities. In Chap. 5, we study the stability and the superstability of C -ternary homomorphisms, C -ternary derivations, C -ternary 3-homomorphisms, and C -ternary 3-derivations in C -ternary algebras as well as investigate the stability of JB -triple homomorphisms and JB -triple derivations in JB -triples by using the direct method and the fixed point method. In Chap. 6, we study the stability of linear mappings in multi-Banach spaces as well as the stability and the superstability of isomorphisms, homomorphisms, and derivations in multi-Banach algebras, multi-C -algebras, proper multi-CQ algebras, and multi-C -ternary algebras. Moreover, we study the stability of ternary Jordan homomorphisms and ternary Jordan derivations in multi-C -ternary algebras. Finally, in Chap. 7, we study the stability of additive functional equations in non-Archimedean Banach spaces as well as the stability of homomorphisms and derivations in non-Archimedean C -algebras and non-Archimedean Lie C algebras. Jinju, South Korea Seoul, South Korea Athens, Greece Tehran, Iran March 2015

Yeol Je Cho Choonkil Park Themistocles M. Rassias Reza Saadati

Acknowledgments

We would like to express our thanks to the referees for reading the manuscript and providing valuable suggestions and comments which have helped to improve the presentation of the book. This work was supported by the Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology. Last but not least, it is our pleasure to acknowledge the superb assistance provided by the staff of Springer for the publication of the book. Yeol Je Cho Choonkil Park Themistocles M. Rassias Reza Saadati

ix

Contents

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1.1 Fixed Point Theorems . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1.2 Quasi-Banach Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1.3 C -Algebras.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1.4 C -Ternary Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1.5 Non-Archimedean Normed Algebras .. . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1.6 Multi–normed Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . .

1 7 9 10 12 13 14

2 Stability of Functional Equations in Banach Algebras . . . . . . . . . . . . . . . . . . 2.1 Stability of 1q f .qx C qy C qz/ D f .x/ C f .y/ C f .z/ . . . . . . . . . . . . . . . . . 2.1.1 Isomorphisms in Banach Algebras .. . . . .. . . . . . . . . . . . . . . . . . . . 2.1.2 Derivations in Banach Algebras .. . . . . . . .. . . . . . . . . . . . . . . . . . . . 2.2 Stability of m-Variable Functional Equations . . . .. . . . . . . . . . . . . . . . . . . . 2.2.1 Stability of Homomorphisms in Banach Algebras . . . . . . . . . 2.2.2 Stability of Derivations in Banach Algebras .. . . . . . . . . . . . . . . 2.3 Stability in Quasi-Banach Algebras . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 2.3.1 Stability of Homomorphisms in Quasi-Banach Algebras .. 2.3.2 Isomorphisms in Quasi-Banach Algebras . . . . . . . . . . . . . . . . . . 2.3.3 Stability of Generalized Derivations in Quasi-Banach Algebras .. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 2.4 Stability of Cauchy–Jensen Functional Equations . . . . . . . . . . . . . . . . . . . 2.4.1 Stability of Homomorphisms in Real Banach Algebras . . . 2.4.2 Stability of Generalized Derivations in Real Banach Algebras . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . .

19 20 20 23 24 24 29 31 32 35

3 Stability of Functional Equations in C -Algebras .. . .. . . . . . . . . . . . . . . . . . . . 3.1 Isomorphisms in Unital C -Algebras .. . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 3.1.1 C -Algebra Isomorphisms in Unital C -Algebras . . . . . . . . . 3.1.2 On the Mazur-Ulam Theorem in Modules over C -Algebras .. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 3.2 Apollonius Type Additive Functional Equations .. . . . . . . . . . . . . . . . . . . . 3.2.1 Homomorphisms and Derivations on C -Algebras . . . . . . . .

51 53 54

36 38 38 44

61 63 64 xi

xii

Contents

3.2.2 Homomorphisms and Derivations in Lie C -Algebras.. . . . 3.2.3 Homomorphisms and Derivations in JC -Algebras .. . . . . . . 3.3 Stability of Jensen Type Functional Equations in C -Algebras . . . . . 3.3.1 Stability of Homomorphisms in C -Algebras . . . . . . . . . . . . . . 3.3.2 Stability of Derivations in C -Algebras . . . . . . . . . . . . . . . . . . . . 3.3.3 Stability of Homomorphisms in Lie C -Algebras.. . . . . . . . . 3.3.4 Stability of Lie Derivations in C -Algebras . . . . . . . . . . . . . . . . 3.4 Generalized Additive Mapping . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 3.4.1 Hyers–Ulam Stability of Functional Equations in Banach Modules over a C -Algebra .. . . . . . . . . . . . . . . . . . . . 3.4.2 Homomorphisms in Unital C -Algebras . . . . . . . . . . . . . . . . . . . 3.5 Generalized Additive Mappings in Banach Modules . . . . . . . . . . . . . . . . 3.5.1 Odd Functional Equations in d–Variables . . . . . . . . . . . . . . . . . . 3.5.2 Stability of Odd Functional Equations in Banach Modules over a C -Algebra . . . .. . . . . . . . . . . . . . . . . . . . 3.5.3 Isomorphisms in Unital C -Algebras . . .. . . . . . . . . . . . . . . . . . . . 3.6 Jordan -Derivations and Quadratic Jordan -Derivations . . . . . . . . . . 3.6.1 Stability of Jordan -Derivations.. . . . . . .. . . . . . . . . . . . . . . . . . . . 3.6.2 Stability of Quadratic Jordan -Derivations . . . . . . . . . . . . . . . . 3.6.3 Stability of Jordan -Derivations: The Fixed Point Method .. . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 3.7 .˛; ˇ; /-Derivations on Lie C -Algebras: The Direct Method . . . . . 3.8 Square Roots and 3rd Root Functional Equations: The Direct Method.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 3.8.1 Stability of the Square Root Functional Equation .. . . . . . . . . 3.8.2 Stability of the 3rd Root Functional Equation .. . . . . . . . . . . . . 3.9 Square Root and 3rd Root Functional Equations: The Fixed Point Method .. . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 3.9.1 Stability of the Square Root Functional Equation .. . . . . . . . . 3.9.2 Stability of the 3rd Root Functional Equation .. . . . . . . . . . . . . 3.10 Positive-Additive Functional Equation . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 3.10.1 Stability of the Positive-Additive Functional Equations: The Fixed Point Method . . . .. . . . . . . . . . . . . . . . . . . . 3.10.2 Stability of the Positive-Additive Functional Equations: The Direct Method . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 3.11 Stability of -Homomorphisms in JC -Algebras .. . . . . . . . . . . . . . . . . . . 3.11.1 -Homomorphisms in JC -Algebras . . .. . . . . . . . . . . . . . . . . . . . 3.11.2 Stability of -Homomorphisms in JC -Algebras . . . . . . . . . . 4 Stability of Functional Inequalities in Banach Algebras . . . . . . . . . . . . . . . . 4.1 Stability of Additive Functional Inequalities in Banach Algebras . . 4.1.1 Stability of C-Linear Mappings in Banach Spaces . . . . . . . . . 4.1.2 Stability of Homomorphisms in Proper CQ -Algebras . . . . 4.1.3 Stability of Derivations in Proper CQ -Algebras . . . . . . . . . .

67 69 70 70 79 81 83 85 85 92 98 99 100 106 110 110 114 119 124 130 131 135 137 137 141 143 144 148 151 153 160 165 166 166 171 178

Contents

4.2

xiii

Stability of Functional Inequalities over C -Algebras .. . . . . . . . . . . . . . 4.2.1 Functional Inequalities in Normed Modules over C -Algebras . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 4.2.2 On Additive Functional Inequalities in Normed Modules over C -Algebras .. . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 4.2.3 Generalization of Cauchy-Rassias Inequalities via the Fixed Point Method .. . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . .

184 185 192 196

5 Stability of Functional Equations in C -Ternary Algebras . . . . . . . . . . . . . 5.1 C -Ternary 3-Homomorphism and C -Ternary 3-Derivations .. . . . . 5.1.1 C -Ternary 3-Homomorphisms in C -Ternary Algebras . . 5.1.2 C -Ternary 3-Derivations in C -Ternary Algebras . . . . . . . . 5.2 Apollonius Type Additive Functional Equations .. . . . . . . . . . . . . . . . . . . . 5.2.1 Homomorphisms in C -Ternary Algebras .. . . . . . . . . . . . . . . . . 5.2.2 Derivations in C -Ternary Algebras . . . .. . . . . . . . . . . . . . . . . . . . 5.2.3 Homomorphisms in JB -Triples . . . . . . . .. . . . . . . . . . . . . . . . . . . . 5.2.4 Derivations in JB -Triples .. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 5.2.5 C -Ternary Homomorphisms: Fixed Point Method .. . . . . . . 5.2.6 C -Ternary Derivations: The Fixed Point Method . . . . . . . . . 5.2.7 JB -Triple Homomorphisms: The Fixed Point Method .. . . 5.2.8 JB -Triple Derivations: Fixed Point Method . . . . . . . . . . . . . . . 5.3 Bi--Derivations in JB -Triples . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . .

201 202 203 207 209 210 212 213 214 215 216 217 218 219

6 Stability of Functional Equations in Multi-Banach Algebras .. . . . . . . . . . 6.1 Stability of m-Variable Additive Mappings . . . . . .. . . . . . . . . . . . . . . . . . . . 6.1.1 Stability of Homomorphisms in Multi-Banach Algebras . . 6.1.2 Stability of Derivations in Multi-Banach Algebras.. . . . . . . . 6.2 Ternary Jordan Homomorphisms and Derivations in Multi-C -Ternary Algebras . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 6.2.1 Stability of Homomorphisms in Multi-C -Ternary Algebras . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 6.2.2 Stability of Derivations in Multi-C -Ternary Algebras . . . . 6.3 Generalized Additive Mappings and Isomorphisms in Multi-C -Algebras.. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 6.3.1 Stability of Odd Functional Equations in Multi-Banach Modules over a Multi-C -Algebra .. . . . . . . . . 6.3.2 Isomorphisms in Unital Multi-C -Algebras .. . . . . . . . . . . . . . . 6.4 Additive Functional Inequalities in Proper Multi-CQ -Algebras . . . 6.4.1 Stability of C-Linear Mappings in Multi-Banach Spaces . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 6.4.2 Stability of Homomorphisms in Proper Multi-CQ -Algebras .. . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 6.4.3 Stability of Derivations in Proper CQ -Algebras . . . . . . . . . .

229 230 230 238 241 242 249 252 252 262 264 264 269 277

xiv

Contents

6.5

Stability of Homomorphisms and Derivations in Multi-C –Ternary Algebras .. . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 282 6.5.1 Stability of Homomorphisms .. . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 282 6.5.2 Stability of Derivations in Multi-C -Ternary Algebras . . . . 297

7 Stability of Functional Equations in Non-Archimedean Banach Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 7.1 Stability of Jensen Type Functional Equations: The Fixed Point Approach . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 7.1.1 Stability of Homomorphisms in Non-Archimedean C -Algebras.. . . . . . . .. . . . . . . . . . . . . . . . . . . . 7.1.2 Stability of Derivations in Non-Archimedean C -Algebras .. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 7.1.3 Stability of Homomorphisms in Non-Archimedean Lie C -Algebras . . . .. . . . . . . . . . . . . . . . . . . . 7.1.4 Stability of Non-Archimedean Lie Derivations in C -Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 7.2 Stability for m-Variable Additive Functional Equations . . . . . . . . . . . . . 7.2.1 Stability of Homomorphisms and Derivations in Non-Archimedean C -Algebras .. . . . .. . . . . . . . . . . . . . . . . . . . 7.2.2 Stability of Homomorphisms and Derivations in Non-Archimedean Lie C -Algebras .. . . . . . . . . . . . . . . . . . . .

303 304 304 314 316 318 320 320 324

References .. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 327 Index . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 341

Chapter 1

Introduction

In this chapter, we recall some definitions and results which will be used later on in the book. The study of functional equations has a long history. In 1791 and 1809, Legendre [184] and Gauss [121] attempted to provide a solution of the following functional equation: f .x C y/ D f .x/ C f .y/ for all x; y 2 R, which is called the Cauchy functional equation. A function f W R ! R is called an additive function if it satisfies the Cauchy functional equation. In 1821, Cauchy [67] first found the general solution of the Cauchy functional equation, that is, if f W R ! R is a continuous additive function, then f is R–linear, that is, f .x/ D mx, where m is a constant. Further, we can consider the biadditive function on R R as follows: A function f W R R ! R is called a biadditive function if it is additive in each variable, that is, f .x C y; z/ D f .x; z/ C f .y; z/ and f .x; y C z/ D f .x; y/ C f .x; z/ for all x; y; z 2 R. It is well–known that every continuous biadditive function f W R R ! R is of the form f .x; y/ D mxy for all x; y 2 R, where m is a constant. © Springer International Publishing Switzerland 2015 Y.J. Cho et al., Stability of Functional Equations in Banach Algebras, DOI 10.1007/978-3-319-18708-2_1

1

2

1 Introduction

Since the time of Legendre and Gauss, several mathematicians had dealt with additive functional equations in their books [4–6, 178, 313] and a number of them have studied Lagrange’s mean value theorem and related functional equations, Pompeiu’s mean value theorem and associated functional equations, two-dimensional mean value theorem and functional equations as well as several kinds of functional equations. We know that the mean value theorems have been motivated to study the functional equations (see the book “Mean Value Theorems and Functional Equations” by Sahoo and Riedel [305]) in 1998. In 1940, Ulam [321] proposed the following stability problem of functional equations: Given a group G1 , a metric group G2 with the metric d.; / and a positive number ", does there exist ı > 0 such that, if a mapping f W G1 ! G2 satisfies d.f .xy/; f .x/f .y// ı for all x; y 2 G1 , then there exists a homomorphism h W G1 ! G2 such that d.f .x/; h.x// " for all x 2 G1 ? Since then, several mathematicians have dealt with special cases as well as generalizations of Ulam’s problem. In fact, in 1941, Hyers [133] provided a partial solution to Ulam’s problem for the case of approximately additive mappings in which G1 and G2 are Banach spaces with ı D " as follows: Let X and Y be Banach spaces and let " > 0. Then, for all g W X ! Y with sup kg.x C y/ g.x/ g.y/k "; x;y2X

there exists a unique mapping f W X ! Y such that sup kg.x/ f .x/k " x2X

and f .x C y/ D f .x/ C f .y/ for all x; y 2 X. This proof remains unchanged if G1 is an Abelian semigroup. Particularly, in 1968, it was proved by Forti ([115], Proposition 1) that the following theorem can be proved:

1 Introduction

3

Theorem F (Forti). Let .S; C/ be an arbitrary semigroup and E be a Banach space. Assume that f W S ! E satisfies kf .x C y/ f .x/ f .y/k ":

(A)

Then the limit f .2n x/ n!1 2n

g.x/ D lim

(B)

exists for all x 2 S and g W S ! E is the unique function satisfying kf .x/ g.x/k ";

g.2x/ D 2g.x/:

Finally, if the semigroup S is Abelian, then G is additive. Here, the proof method which generates the solution g by the formula like (B) is called the direct method. If f is a mapping of a group or a semigroup .S; / into a vector space E, then we call the following expression: Cf .x; y/ D f .x y/ f .x/ f .y/ the Cauchy difference of f on S S. In the case that E is a topological vector space, we call the equation of homomorphism stable if, whenever the Cauchy difference Cf is bounded on S S, there exists a homomorphism g W S ! E such that f g is bounded on S. In 1980, Rätz [298] generalized Theorem F as follows: Let .X; / be a powerassociative groupoid, i.e., X is a nonempty set with a binary relation x1 x2 2 X such that the left powers satisfy xmCn D xm xn for all m; n 1 and x 2 X. Let .Y; j j/ be a topological vector space over the field Q of rational numbers with Q topologized by its usual absolute value j j. Theorem R (Rätz). Let V be a nonempty bounded Q-convex subset of Y containing the origin and assume that Y is sequentially complete. Let f W X ! Y satisfy the following conditions: for all x1 ; x2 2 X, there exist k 2 such that n

n

n

f ..x1 x2 /k / D f .xk1 xk2 /

(C)

f .x1 / C f .x2 / f .x1 x2 / 2 V:

(D)

for all n 1 and

Then there exists a function g W X ! Y such that g.x1 / C g.x2 / D g.x1 x2 / and f .x/ g.x/ 2 V, where V is the sequential closure of V for all x 2 X. When Y is a Hausdorff space, then g is uniquely determined.

4

1 Introduction

Note that the condition (C) is satisfied when X is commutative and it takes the place of the commutativity in proving the additivity of g. However, as Rätz pointed out in his paper, the condition n

n

n

.x1 x2 /k D xk1 xk2

for all x1 ; x2 2 X, where X is a semigroup, and, for all k 1, does not imply the commutativity. In the proofs of Theorems F and R, the completeness of the image space E and the sequential completeness of Y, respectively, were essential in proving the existence of the limit which defined the additive function g. The question arises whether the completeness is necessary for the existence of an odd additive function g such that f g is uniformly bounded, given that the Cauchy difference is bounded. For this problem, in 1988, Schwaiger [306] proved the following: Theorem S (Schwaiger). Let E be a normed space with the property that, for each function f W Z ! E, whose Cauchy difference Cf D f .xCy/f .x/f .y/ is bounded for all x; y 2 Z and there exists an additive mapping g W Z ! E such that f .x/g.x/ is bounded for all x 2 Z. Then E is complete. Corollary 1. The statement of Theorem S remains true if Z is replaced by any vector space over Q. In 1950, Aoki [17] generalized Hyers’ theorem as follows: Theorem A (Aoki). Let E1 and E2 be two Banach spaces. If there exist K > 0 and 0 p < 1 such that kf .x C y/ f .x/ f .y/k K.kxkp C kykp / for all x; y 2 E1 , then there exists a unique additive mapping g W E1 ! E2 such that kf .x/ g.x/k

2K kxkp 2 2p

for all x 2 E1 . In 1978, Th. M. Rassias [267] formulated and proved the stability theorem for the linear mapping between Banach spaces E1 and E2 subject to the continuity of f .tx/ with respect to t 2 R for each fixed x 2 E1 . Thus Rassias’ Theorem implies Aoki’s Theorem as a special case. Later, in 1990, Th. M. Rassias [274] observed that the proof of his stability theorem also holds true for p < 0. In 1991, Gajda [119] showed that the proof of Rassias’ Theorem can be proved also for the case p > 1 by just replacing n by n in (B). These results are stated in a generalized form as follows (see Rassias and Šemrl [293]):

1 Introduction

5

Theorem RS (Th.M. Rassias and P. Semrl). Let ˇ.s; t/ be nonnegative function for all nonnegative real numbers s; t and positive homogeneous of degree p, where p is real and p ¤ 1, i.e., ˇ.s; t/ D p ˇ.s; t/ for all nonnegative ; s; t. Given a normed space E1 and a Banach space E2 , assume that f W E1 ! E2 satisfies the inequality kf .x C y/ f .x/ f .y/k ˇ.kxk; kyk/ for all x; y 2 E1 . Then there exists a unique additive mapping g W E1 ! E2 such that kf .x/ g.x/k ıkxkp for all x 2 E1 , where ( ı WD

ˇ.1;1/ ; 22p ˇ.1;1/ ; 22p

p < 1; p > 1:

The proofs for the cases p < 1 and p > 1 were provided by applying the direct methods. For p < 1, the additive mapping g is given by (B), while, in case p > 1, the formula is g.x/ D lim 2n f n!1

x : 2n

Corollary 2. Let f W E1 ! E2 be a mapping satisfying the hypotheses of Theorem RS and suppose that f is continuous at a single point y 2 E1 . Then the additive mapping g is continuous. Corollary 3. If, under the hypotheses of Theorem RS, we assume that, for each fixed x 2 E1 , the mapping t ! f .tx/ from R to E2 is continuous, then the additive mapping g is R–linear. Remark 4. (1) For p D 0, Theorem RS, Corollaries 2 and 3 reduce to the results of Hyers in 1941. If we put ˇ.s; t/ D ".sp C tp /, then we obtain the results of Rassias [267] in 1978 and Gajda [119] in 1991. (2) The case p D 1 was excluded in Theorem RS. Simple counterexamples prove that one can not extend Rassias’ Theorem when p takes the value one (see Z. Gajda [119], Rassias and Šemrl [293] and Hyers and Rassias [135] in 1992). A further generalization of the Hyers-Ulam stability for a large class of mappings was obtained by Isac and Rassias [139] by introducing the following: Definition 5. A mapping f W E1 ! E2 is said to be -additive if there exist ˚ 0 and a function W RC ! RC satisfying .t/ D0 t!C1 t lim

6

1 Introduction

such that kf .x C y/ f .x/ f .y/k ˚Œ.kxk/ C .kyk/ for all x; y 2 E1 . In [139], Isac and Rassias proved the following: Theorem IR (Isac and Rassias). Let E1 be a real normed vector space and E2 be a real Banach space. Let f W E1 ! E2 be a mapping such that f .tx/ is continuous in t for each fixed x 2 E1 . If f is -additive and satisfies the following conditions: (1) .ts/ .t/.s/ for all s; t 2 R; (2) .t/ < t for all t > 1, then there exists a unique R–linear mapping T W E1 ! E2 such that kf .x/ T.x/k

2 .kxk/ 2 .2/

for all x 2 E1 . Remark 6. (1) If .t/ D tp with p < 1, then, from Theorem IR, we obtain Rassias’ Theorem [267]. (2) If p < 0 and .t/ D tp with t > 0, then Theorem IR is implied by the result of Gajda in 1991. Since the time the above stated results have been proven, several mathematicians (cf. [1, 3, 14, 44, 46, 49–65, 69–80, 82–85, 87–90, 95–99, 101–107, 109, 116– 118, 120, 122, 124, 125, 129–134, 136, 137, 140–149, 151–153, 156–158, 160– 162, 164, 173–179, 187, 189, 190, 195–201, 207, 208, 212, 214, 217, 219, 221– 223, 226, 228, 230, 236, 239–241, 262, 266, 268, 269, 275–288, 296–303, 309, 311–322, 324, 330, 331] and also very recent survey papers [42, 60, 61]) have extensively studied stability theorems for several kinds of functional equations in various spaces, for example, Banach spaces, 2-Banach spaces, Banach nLie algebras, quasi-Banach spaces, Banach ternary algebras, non-Archimedean normed and Banach spaces, metric and ultra metric spaces, Menger probabilistic normed spaces, probabilistic normed space, p-2-normed spaces, C -algebras, C ternary algebras, Banach ternary algebras, Banach modules, inner product spaces, Heisenberg groups and others. Further, we have to pay attention to applications of the Hyers-Ulam-Rassias stability problems, for example, (partial) differential equations, Fréchet functional equations, Riccati differential equations, Volterra integral equations, group and ring theory and some kinds of equations (see [66, 142, 150, 154, 159, 176, 185, 186, 192–194, 259–261, 327, 329]). For more details on recent development in Ulam’s type stability and its applications, see the papers of Brillouët-Belluot et al. [53] and Ciepliński [86] in 2012 (see also [3–22, 47, 78, 79, 81, 110–112, 138, 148–150, 154, 157–163, 165, 166, 183, 204, 205, 213, 215, 216, 245, 246, 249, 255, 257, 263, 264, 270–295, 302, 304, 328]).

1.1 Fixed Point Theorems

7

A functional equation is called stable if any function satisfying the functional equation “approximately” is near to a true solution of the functional equation. We say that a functional equation is superstable if every approximate solution is an exact solution of it (see some recent papers [40, 41, 55, 59]).

1.1 Fixed Point Theorems In this section, we present some fixed point theorems which will play important roles in proving our main theorems. All stability results for functional equations were proved by applying direct method. Since the direct method sometimes does not work. In consequence, the fixed point method for studying the stability of functional equations was used for the first time by Baker in 1991 (see [43]). Next, in 2003, Radu [265] gave a lecture at Seminar on Fixed Point Theory Cluj-Napoca and proved a stability of functional equation by fixed method. Then, in 2003, C˘adariu and Radu [62, 64] considered Jensen functional equation and proved a stability result via fixed point method. Jung and Chang [155] proved the stability of a cubic type functional equation with the fixed point alternative. Since then, some authors [151–153, 156, 157, 162, 164, 191, 211, 234, 251, 256] considered some important functional equations and proved the stability results via fixed point method introduced by Baker and Radu. The Banach fixed point theorem [45] (also known as the Banach contraction principle) is an important tool in the theory of metric spaces because it guarantees the existence and uniqueness of fixed points of certain self mappings of metric spaces and provides a constructive method to find those fixed points. The theorem is named after Banach (1892–1945) and was first stated by him in 1922. Theorem 1.1 (Banach [45]). Let .X; d/ be a complete metric space and T W X!X be a contraction, i.e., there exists ˛ 2 Œ0; 1/ such that d.Tx; Ty/ ˛d.x; y/ for all x; y 2 X. Then there exists a unique a 2 X such that Ta D a. Moreover, for all x 2 X, lim T n x D a

n!1

and, in fact, for all x 2 X, d.x; a/

1 d.x; Tx/: 1˛

8

1 Introduction

Theorem 1.2 ([62, 265]). Let .X; d/ be a complete metric space and J W X ! X be a strictly contractive mapping, i.e., there exists a Lipschitz constant L < 1 such that d.Jx; Jy/ Ld.x; y/ for all x; y 2 X. Then we have (1) The mapping J has a unique fixed point x 2 X; (2) The fixed point x is globally attractive, i.e., lim J n x D x

n!1

for all x 2 X; (3) The following inequalities hold: d.J n x; x / Ln d.x; x /; 1 d.J n x; J nC1 x/; 1L 1 d.x; Jx/ d.x; x / 1L

d.J n x; x /

for all x 2 X and n 1. Following Luxemburg [188], the concept of a generalized complete metric space may be introduced as in this quotation: Let X be a nonempty set. A function d W X X ! Œ0; 1 is called a generalized metric on X if, for any x; y; z 2 X, (1) d.x; y/ D 0 if and only if x D y; (2) d.x; y/ D d.y; x/; (3) d.x; z/ d.x; y/ C d.y; z/. This concept differs from the usual concept of a complete metric space by the fact that not every two points in X have necessarily a finite distance. One might call such a space a generalized complete metric space. Next, Diaz and Margolis [95] proved a theorem of the alternative for any contraction mapping T on a generalized complete metric space X. The conclusion of the theorem, speaking in general terms, asserts that: either all consecutive pairs of the sequence of successive approximations (starting from an element x0 of X) are infinitely far apart or the sequence of successive approximations, with initial element x0 converges to a fixed point of T (what particular fixed point depends, in general, on the initial element x0 ). Theorem 1.3 ([62, 95]). Let .X; d/ be a complete generalized metric space and J W X ! X be a strictly contractive mapping with a Lipschitz constant L < 1. Then, for each x 2 X, either

1.2 Quasi-Banach Algebras

9

d.J n x; J nC1 x/ D 1 for all n 0 or there exists a positive integer n0 such that (1) (2) (3) (4)

d.J n x; J nC1 x/ < 1 for all n n0 ; The sequence .J n x/ converges to a fixed point y of J; y is the unique fixed point of J in the set Y D fy 2 X W d.J n0 x; y/ < 1g; 1 d.y; y / 1L d.y; Jy/ for all y 2 Y.

1.2 Quasi-Banach Algebras Let X be a vector space on field C. A normed space X in which, for all x; y 2 X, xy 2 X and kxyk kxkkyk is called a complex normed algebra. A complete normed algebra is called a Banach algebra. Moreover, if there exists a unit element e such that ex D xe D x for all x 2 X, then kek D 1 and X is called a unital Banach algebra. Let X; Y be Banach algebras. A C-linear mapping H W X ! Y is called a homomorphism in Banach algebras if H satisfies H.xy/ D H.x/H.y/ for all x; y 2 X. A C-linear mapping ı W X ! X is called a derivation on X if ı satisfies ı.xy/ D ı.x/y C xı.y/ for all x; y 2 X. We recall some basic facts concerning quasi-Banach spaces and some preliminary results. Definition 1.4 ([48, 300]). Let X be a real linear space. A quasi-norm is a realvalued function on X satisfying the following: (1) kxk 0 for all x 2 X and kxk D 0 if and only if x D 0; (2) kxk D jj kxk for all 2 R and x 2 X; (3) There is a constant K 1 such that kx C yk K.kxk C kyk/ for all x; y 2 X. The pair .X; k k/ is called a quasi-normed space if k k is a quasi-norm on X. The smallest possible K is called the modulus of concavity of k k. Obviously, the balls with respect to k k define a linear topology on X. By a quasi-Banach space we mean a complete quasi-normed space, i.e., a quasi-normed space in which every k k-Cauchy sequence in X converges. This class includes Banach spaces and the most significant class of quasi-Banach spaces which are not Banach spaces are the Lp spaces for 0 < p < 1 with the quasi-norm k kp .

10

1 Introduction

A quasi-norm k k is called a p-norm .0 < p 1/ if kx C ykp kxkp C kykp for all x; y 2 X. In this case, a quasi-Banach space is called a p-Banach space. For any p-norm, the formula d.x; y/ WD kx ykp gives us a translation invariant metric on X. By the Aoki–Rolewicz theorem [300] (see also [48]), each quasi-norm is equivalent to some p-norm. Since it is much easier to work with p-norms than quasi-norms, henceforth we restrict our attention mainly to p-norms. Definition 1.5 ([10]). Let .A; k k/ be a quasi-normed space. The quasi-normed space .A; k k/ is called a quasi-normed algebra if A is an algebra and there exists a constant C > 0 such that kxyk Ckxk kyk for all x; y 2 A. A quasi-Banach algebra is a complete quasi-normed algebra. If the quasi-norm k k is a p-norm, then the quasi-Banach algebra is called a p-Banach algebra.

1.3 C -Algebras Let U be a Banach algebra. Then an involution on U is a mapping u ! u from U into U which satisfies the following conditions: (1) u D u for all u 2 U; (2) .˛u C ˇv/ D ˛u C ˇv ; (3) .uv/ D v u for all u; v 2 U. If, in addition, ku uk D kuk2 for all u 2 U, then U is a C -algebra. Let U; V be C -algebras. A C-linear mapping H W U ! V is called a homomorphism in C -algebras if H satisfies H.xy/ D H.x/H.y/;

H.x / D H.x/

for all x; y 2 U. A C-linear mapping ı W U ! U is called a derivation on U if ı satisfies ı.xy/ D ı.x/y C xı.y/ for all x; y 2 U.

1.3 C -Algebras

11

Suppose that A is a complex Banach -algebra. Let C-linear mapping ı W D.ı/ ! A be a derivation on A, where D.ı/ is the domain of ı and D.ı/ is dense in A. If ı satisfies the additional condition ı.a / D ı.a/ for all a 2 A, then ı is called a -derivation on A. It is well-known that, if A is a C -algebra and D.ı/ is A, then the derivation ı is bounded. Now, we consider proper CQ -algebras, which arise as completions of C -algebras (see [15–39]) as follows: Let A be a Banach module over the C -algebra A0 with an involution and C -norm k k0 such that A0 A. We say that .A; A0 / is a proper CQ algebra if (1) A0 is dense in A with respect to its norm k k; (2) An involution , which extends the involution of A0 , is defined in A with the property .xy/ D y x for all x; y 2 A whenever the multiplication is defined; (3) kyk0 D supx2A;kxk1 kxyk for all y 2 A0 . Definition 1.6. Let .A; A0 / and .B; B0 / be proper CQ -algebras. (1) A C-linear mapping h W A ! B is called a proper CQ -algebra homomorphism if h.xy/ D h.x/h.y/ for all x; y 2 A whenever the multiplication is defined; (2) A C-linear mapping ı W A ! A is called a derivation on A if ı.xy/ D ı.x/y C xı.y/ for all x; y 2 A whenever the multiplication is defined. A C -algebra C endowed with the Lie product Œx; y WD C -algebra (see [224, 225, 227]).

xyyx 2

on C is called a Lie

Definition 1.7. Let A and B be Lie C -algebras. A C-linear mapping H W A ! B is called a Lie C -algebra homomorphism if H.Œx; y/ D ŒH.x/; H.y/ for all x; y 2 A. Definition 1.8. Let A be a Lie C -algebra. A C-linear mapping ı W A ! A is called a Lie derivation if ı.Œx; y/ D Œı.x/; y C Œx; ı.y/ for all x; y 2 A.

12

1 Introduction

1.4 C -Ternary Algebras Ternary algebraic structures appear more or less naturally in various domains of theoretical and mathematical physics, for example, the quark model inspired a particular brand of ternary algebraic system. One such attempt has been proposed by Nambu in 1973 and is now known under the name of “Nambu mechanics” [316] (see also [332]). A C -ternary algebra is a complex Banach space A, equipped with a ternary product .x; y; z/ 7! Œx; y; z of A3 into A, which is C-linear in the outer variables, conjugate C-linear in the middle variable and associative in the sense that Œx; y; Œz; w; v D Œx; Œw; z; y; v D ŒŒx; y; z; w; v and satisfies kŒx; y; zk kxk kyk kzk;

kŒx; x; xk D kxk3

(see [332]). If a C -ternary algebra .A; Œ; ; / has the identity, i.e., an element e 2 A such that x D Œx; e; e D Œe; e; x for all x 2 A, then it is routine to verify that A, endowed with x ı y WD Œx; e; y and x WD Œe; x; e, is a unital C -algebra. Conversely, if .A; ı/ is a unital C -algebra, then Œx; y; z WD x ı y ı z makes A into a C -ternary algebra. A C-linear mapping H W A ! B is called a C -ternary algebra homomorphism if H.Œx; y; z/ D ŒH.x/; H.y/; H.z/ for all x; y; z 2 A. A C-linear mapping ı W A ! A is called a C -ternary derivation if ı.Œx; y; z/ D Œı.x/; y; z C Œx; ı.y/; z C Œx; y; ı.z/ for all x; y; z 2 A (see [231]). Ternary structures and their generalization, the so-called n-ary structures, are important in view of their applications in physics (see [171]). Suppose that J is a complex vector space endowed with a real trilinear composition J J J 3 .x; y; z/ 7! fxy zg 2 J which is complex bilinear in .x; z/ and conjugate linear in y. Then J is called a Jordan triple system if fxy zg D fzy xg and ffxy zgu vg C ffxy vgu zg fxy fzu vgg D fzfyx ug vg:

1.5 Non-Archimedean Normed Algebras

13

We are interested in Jordan triple systems having a Banach space structure. A complex Jordan triple system J with a Banach space norm k k is called a J triple if, for all x 2 J , the operator xx is hermitian in the sense of Banach algebra theory. Here the operator xx on J is defined by .xx /y WD fxx yg. This implies that xx has the real spectrum .xx / R. A J -triple J is called a JB -triple if every x 2 J satisfies .xx / 0 and kxx k D kxk2 . A C-linear mapping H W J ! L is called a JB -triple homomorphism if H.fxyzg/ D fH.x/H.y/H.z/g for all x; y; z 2 J . A C-linear mapping ı W J ! J is called a JB -triple derivation if ı.fxyzg/ D fı.x/yzg C fxı.y/zg C fxyı.z/g for all x; y; z 2 J (see [225]).

1.5 Non-Archimedean Normed Algebras By a non-Archimedean field we mean a field K equipped with a function (valuation) j j from K into Œ0; 1/ such that (1) jrj D 0 if and only if r D 0; (2) jrsj D jrj jsj; (3) jr C sj maxfjrj; jsjg for all r; s 2 K. Clearly, j1j D j 1j D 1 and jnj 1 for all n 1. By the trivial valuation we mean the mapping j j taking everything but 0 into 1 and j0j D 0. Let X be a vector space over a field K with a non-Archimedean non-trivial valuation j j. A function k k W X ! Œ0; 1/ is called a non-Archimedean norm if it satisfies the following conditions: (1) kxk D 0 for all x 2 X if and only if x D 0; (2) For all r 2 K; x 2 X, krxk D jrjkxk; (3) The strong triangle inequality (ultrametric) holds, i.e., kx C yk maxfkxk; kykg for all x; y 2 X. Then .X; k k/ is called a non-Archimedean normed space. From the fact that kxn xm k maxfkxjC1 xj k W m j n 1g

14

1 Introduction

for all n 1 with n > m, a sequence fxn g is a Cauchy sequence if and only if fxnC1 xn g converges to zero in a non-Archimedean normed space. By a complete non-Archimedean normed space we mean one in which every Cauchy sequence is convergent. For any nonzero rational number x, there exists a unique integer nx 2 Z such that xD

a nx p ; b

where a and b are integers not divisible by p. Then jxjp WD pnx defines a nonArchimedean norm on Q. The completion of Q with respect to the metric d.x; y/ D jx yjp is denoted by Qp , which is called the p-adic number field. A non-Archimedean Banach algebra is a complete non-Archimedaen algebra A which satisfies kabk kak kbk for all a; b 2 A. For more detailed definitions of non-Archimedean Banach algebras, the readers refer to [310]. If U is a non-Archimedean Banach algebra, then an involution on U is a mapping t ! t from U into U satisfying the following: (1) t D t for all t 2 U; (2) .˛s C ˇt/ D ˛s C ˇt ; (3) .st/ D t s for all s; t 2 U. If, in addition kt tk D ktk2 for all t 2 U, then U is a non-Archimedean C –algebra.

1.6 Multi–normed Algebras The notion of multi-normed space was introduced by Dales and Polyakov in [92]. This concept is somewhat similar to the operator sequence space and has some connections with the operator spaces and Banach lattices. Motivations for the study of multi-normed spaces and many examples are given in [91, 92, 206]. Let .E; k k/ be a complex normed space and let k 2 N. We denote by E k the linear space E ˚ ˚ E consisting of k-tuples .x1 ; ; xk /, where x1 ; ; xk 2 E. The linear operations on E k are defined coordinate-wise. The zero element of either E or E k is denoted by 0. We denote by Nk the set f1; 2; ; kg and by ˙k the group of permutations on k symbols. Definition 1.9. A multi-norm on fE k W k 2 Ng is a sequence .k kk / D .k kk W k 2 N/ such that k kk is a norm on E k for each k 2 N: (A1) k.x .1/ ; ; x .k/ /kk D k.x1 ; ; xk /kk for all 2 ˙k and x1 ; ; xk 2 E; (A2) k.˛1 x1 ; ; ˛k xk /kk .maxi2Nk j˛i j/ kx1 ; ; xk kk for all ˛1 ; ; ˛k 2 C and x1 ; ; xk 2 E;

1.6 Multi–normed Algebras

15

(A3) k.x1 ; ; xk1 ; 0/kk D k.x1 ; ; xk1 /kk1 for all x1 ; ; xk1 2 E; (A4) k.x1 ; ; xk1 ; xk1 /kk D k.x1 ; ; xk1 /kk1 for all x1 ; ; xk1 2 E. In this case, we say that ..E k ; k kk / W k 2 N/ is a multi-normed space. Lemma 1.10 ([206]). Suppose that ..E k ; k kk / W k 2 N/ is a multi-normed space and let k 2 N. Then (1) k.x; ; x/kk D kxk for all x 2 E; Pk (2) maxi2Nk kxi k k.x1 ; ; xk /kk iD1 kxi k k maxi2Nk kxi k for all x1 ; ; xk 2 E. It follows from (2) that, if .E; k k/ is a Banach space, then .E k ; k kk / is a Banach space for each k 2 N. In this case, ..E k ; k kk / W k 2 N/ is a multi-Banach space. Now, we state two important examples of multi-norms for an arbitrary normed space E (see [92]). Example 1.11. The sequence .k kk W k 2 N/ on fE k W k 2 Ng defined by kx1 ; ; xk kk WD max kxi k i2Nk

for all x1 ; ; xk 2 E is a multi-norm called the minimum multi-norm. The terminology “minimum” is justified by the property (2). Example 1.12. Let f.k k˛k W k 2 N/ W ˛ 2 Ag be the (non-empty) family of all multi-norms on fE k W k 2 Ng. For k 2 N, set kjx1 ; ; xk kjk WD sup k.x1 ; ; xk /k˛k ˛2A

for all x1 ; ; xk 2 E. Then .kj kjk W k 2 N/ is a multi-norm on fE k W k 2 Ng, which is called the maximum multi-norm. We need the following observation which can be easily deduced from the triangle inequality for the norm k kk and the property (2) of multi-norms. Lemma 1.13. Suppose that k 2 N and .x1 ; ; xk / 2 E k . For each j 2 f1; ; kg, let .xjn /n1 be a sequence in E such that limn!1 xjn D xj . Then, for each .y1 ; ; yk / 2 E k , we have lim .x1n y1 ; ; xkn yk / D .x1 y1 ; ; xk yk /:

n!1

Definition 1.14. Let ..E k ; k kk / W k 2 N/ be a multi-normed space. A sequence .xn / in E is a multi-null sequence if, for any " > 0, there exists n0 2 N such that sup k.xn ; ; xnCk1 /kk < " k2N

for each n n0 . We say that the sequence .xn / is multi-convergent to a point x 2 E .write limn!1 xn D x/ if .xn x/ is a multi-null sequence.

16

1 Introduction

Definition 1.15 ([92, 170, 202]). Let .A; k k/ be a normed algebra such that ..Ak ; k kk / W k 2 N/ is a multi-normed space. Then ..Ak ; k kk / W k 2 N/ is called a multi-normed algebra if k.a1 b1 ; ; ak bk /kk k.a1 ; ; ak /kk k.b1 ; ; bk /kk for all k 2 N and a1 ; ; ak ; b1 ; ; bk 2 A. Further, the multi-normed algebra ..Ak ; k kk / W k 2 N/ is called a multi-Banach algebra if ..Ak ; k kk / W k 2 N/ is a multi-Banach space. Example 1.16 ([92, 202, 252]). Let p; q with 1 p q < 1 and A D `p . The algebra A is a Banach sequence algebra with respect to coordinatewise multiplication of sequences. Let .k kk W k 2 N/ be the standard .p; q/-multi-norm on fAk W k 2 Ng. Then ..Ak ; k kk / W k 2 N/ is a multi-Banach algebra. Definition 1.17. Let .A; kk/ be a Banach -algebra with the involution . A multiC -algebra is a multi-Banach algebra such that k.a1 a1 ; ; ak ak /k D k.a1 ; ; ak /k2 : In a series of the papers [15–33, 36–38] and [318–320], many authors have considered a special class of quasi -algebras, called proper CQ -algebras, which arise as completions of C -algebras. They can be introduced in the following way: Let A be a linear space and A be a -algebra contained in A. We say that A is a quasi--algebra over A if the right and left multiplications of an element of A and an element of A are always defined and linear. An involution which extends the involution of A is defined in A with the property .ab/ D b a whenever the multiplication is defined. A quasi--algebra .A; A/ is said to be topological if there exists a locally convex topology on A such that (Q1) The involution a 7! a is continuous; (Q2) The mappings a 7! ab and a 7! ba are continuous for each b 2 A; (Q3) A is dense in A with topology . In a topological quasi--algebra, the associative law holds in the following two formulations: a.bc/ D .ab/c;

b.ac/ D .ba/c

for all b; c 2 A and a 2 A. A CQ -algebra is a topological quasi--algebra .A; A/ with the following properties: (CQ1) .A; kk / is a C -algebra with respect to the norm kk and the involution ; (CQ2) .A; k k/ is a Banach space and ka k D kak for all a 2 A;

1.6 Multi–normed Algebras

17

(CQ3) For all b 2 A, we have kbk D max

n

o sup kabk; sup kbak :

kak1

kak1

Bagarello and Trapani [34] showed that both .Lp .X; /; C0 .X// and .Lp .X; /; L .X// are CQ -algebras. Now, we define the multi-CQ -algebra. Let .A; A/ be a CQ -algebra. We say that f.Ak ; Ak / W k 2 Ng is a multi-CQ algebra if, for each k 2 N, the couple .Ak ; Ak / is a CQ -algebra, where fAk W k 2 Ng and fAk W k 2 Ng are a multi-Banach algebra and a multi-C -algebra, respectively. 1

Example 1.18. In [34], the authors showed that the couple .A; A/ is a CQ -algebra, where A D `p and A D c0 . Now, consider Example 1.16. Then f.Ak ; Ak / W k 2 Ng is a multi-CQ -algebra. Definition 1.19. Let ..Ak ; k kk / W k 2 N/ be a multi-Banach space. A multi-C ternary algebra is a complex multi-Banach space ..Ak ; k kk / W k 2 N/ equipped with a ternary product.

Chapter 2


Beginning around the year 1980, the topic of approximate homomorphisms and derivations and their stability theory in the field of functional equations and inequalities was taken up by several mathematicians (see Hyers and Rassias [135], Rassias [285] and the references therein). In this chapter, in the first section, we show that, if X and Y are normed spaces with the norms k kX and k kY , respectively, and f W X ! Y is a mapping such that 1 kf .x/ C f .y/ C f .z/kY f .qx C qy C qz/ Y q for all x; y; z 2 X and for a fixed nonzero rational number q, then f is Cauchy additive. Next, we approximate isomorphisms and derivations in Banach algebras by the direct method. In Sect. 2.2, we consider the m-variable additive functional equation: m m m m X X X X f mxi C xj C f xi D 2f mxi iD1

iD1

jD1; j¤i

iD1

for all m 2 N and m 2 and, by the fixed point method, we approximate homomorphisms and derivations in Banach algebras. In Sect. 2.3, we prove the Hyers-Ulam stability of homomorphisms in quasiBanach algebras and generalized derivations on quasi-Banach algebras for the following functional equation: n X iD1

0 f@

n X jD1

1 q.xi xj /A C nf

n X

! qxi

iD1

© Springer International Publishing Switzerland 2015 Y.J. Cho et al., Stability of Functional Equations in Banach Algebras, DOI 10.1007/978-3-319-18708-2_2

D nq

n X

f .xi /:

iD1

19

20

2 Stability of Functional Equations in Banach Algebras

In Sect. 2.4, we approximate homomorphisms in real Banach algebras and generalized derivations on real Banach algebras for the following Cauchy-Jensen functional equations f

x C y 2

x y Cz Cf C z D f .x/ C 2f .z/ 2

and 2f

x C y 2

C z D f .x/ C f .y/ C 2f .z/:

2.1 Stability of 1q f .qx C qy C qz/ D f .x/ C f .y/ C f .z/ Using the direct method, we investigate isomorphisms in Banach algebras and derivations on Banach algebras associated with the following functional equation 1 f .qx C qy C qz/ D f .x/ C f .y/ C f .z/ q

(2.1)

for a fixed nonzero rational number q.

2.1.1 Isomorphisms in Banach Algebras Here we consider isomorphisms in Banach algebras associated with the functional equation (2.1). Lemma 2.1. Let X and Y be normed spaces with norms kkX and kkY , respectively. Let f W X ! Y be a mapping with f .0/ D 0 such that 1 kf .x/ C f .y/ C f .z/kY f .qx C qy C qz/ Y q for all x; y; z 2 X, then f is Cauchy additive, i.e., f .x C y/ D f .x/ C f .y/: Proof. Letting z D 0 and y D x in (2.2), we get 1 kf .x/ C f .x/kY f .0/ D 0 Y q

(2.2)

2.1 Stability of 1q f .qx C qy C qz/ D f .x/ C f .y/ C f .z/

21

for all x 2 X. Hence f .x/ D f .x/ for all x 2 X. Letting z D x y in (2.2), we get kf .x/ C f .y/ f .x C y/kY D kf .x/ C f .y/ C f .x y/kY 1 f .0/ Y q D0 for all x; y 2 X. Thus f .x C y/ D f .x/ C f .y/ for all x; y 2 X. This completes the proof.

Here, we assume that A is a Banach algebra with the norm kkA and B is a Banach algebra with the norm k kB . Theorem 2.2. Let r ¤ 1, be nonnegative real numbers and f W A ! B be a bijective mapping with f .0/ D 0 such that 1 k f .x/ C f .y/ C f .z/kB f .q x C qy C qz/ B q

(2.3)

and 2r kf .xy/ f .x/f .y/kB .kxk2r A C kykA /

(2.4)

for all 2 T1 WD f 2 C W jj D 1g and x; y; z 2 A, then the bijective mapping f W A ! B is an isomorphism. Proof. Let D 1 in (2.3). By Lemma 2.1, the mapping f W A ! B is Cauchy additive. Letting z D 0 and y D x in (2.3), we get f .x/ f . x/ D f .x/ C f . x/ D 0 for all x 2 A and so f . x/ D f .x/ for all x 2 A. By (2.3), the mapping f W A ! B is C-linear. (i) Assume that r < 1. By (2.4), we have 1 kf .4n xy/ f .2n x/f .2n y/kB 4n 4nr 2r lim n .kxk2r A C kykA / n!1 4 D0

kf .xy/ f .x/f .y/kB D lim

n!1

22


for all x; y 2 A and so f .xy/ D f .x/f .y/ for all x; y 2 A. (ii) Assume that r > 1. By a similar method to the proof of the case (i), one can prove that the mapping f W A ! B satisfies f .xy/ D f .x/f .y/ for all x; y 2 A. Therefore, the bijective mapping f W A ! B is an isomorphism in Banach algebras. This completes the proof. Theorem 2.3. Let r ¤ 1, be nonnegative real numbers and f W A ! B be a bijective mapping satisfying f .0/ D 0 and (2.3) such that kf .xy/ f .x/f .y/kB kwkrA kxkrA

(2.5)

for all x; y 2 A, then the bijective mapping f W A ! B is an isomorphism. Proof. By (2.3), the mapping f W A ! B is C-linear. (i) Assume that r < 1. By (2.5), we have 1 kf .4n xy/ f .2n x/f .2n y/kB 4n 4nr lim n kwkrA kxkrA n!1 4 D0

kf .xy/ f .x/f .y/kB D lim

n!1

for all x; y 2 A and so f .xy/ D f .x/f .y/ for all x; y 2 A. (ii) Assume that r > 1. By a similar method to the proof of the case (i), one can prove that the mapping f W A ! B satisfies f .xy/ D f .x/f .y/ for all x; y 2 A. Therefore, the bijective mapping f W A ! B is an isomorphism. This completes the proof.

2.1 Stability of 1q f .qx C qy C qz/ D f .x/ C f .y/ C f .z/

23

2.1.2 Derivations in Banach Algebras We consider derivations on Banach algebras associated with the functional equation (2.1). Theorem 2.4. Let r ¤ 1, be nonnegative real numbers and f W A ! A be a mapping with f .0/ D 0 such that 1 k f .x/ C f .y/ C f .z/kA f .q x C qy C qz/ A q

(2.6)

and 2r kf .xy/ f .x/y xf .y/kA .kxk2r A C kykA /

(2.7)

for all 2 T1 and x; y; z 2 A, then the mapping f W A ! A is a derivation on A. Proof. By (2.7), the mapping f W A ! A is C-linear. (i) Assume that r < 1. By (2.7), kf .xy/ f .x/y xf .y/kA 1 kf .4n xy/ f .2n x/ 2n y 2n xf .2n y/kA n!1 4n 4nr 2r lim n .kxk2r A C kykA / n!1 4 D0

D lim

for all x; y 2 A and so f .xy/ D f .x/y C xf .y/ for all x; y 2 A. (ii) Assume that r > 1. By a similar method to the proof of the case (i), one can prove that the mapping f W A ! A satisfies f .xy/ D f .x/y C xf .y/ for all x; y 2 A. Therefore, the mapping f W A ! A is a derivation on A. This completes the proof. Theorem 2.5. Let r ¤ 1, be nonnegative real numbers and f W A ! A be a mapping satisfying f .0/ D 0 and (2.6) such that kf .xy/ f .x/y xf .y/kA kxkrA kykrA for all x; y 2 A, then the mapping f W A ! A is a derivation on A.

24


Proof. The proof is similar to the proofs of Theorems 2.2 and 2.4.

2.2 Stability of m-Variable Functional Equations In this section, using the fixed point method, we prove the Hyers-Ulam stability of homomorphisms and of derivations on Banach algebras for the following additive functional equation (see [108]): m m m m X X X X f mxi C xj C f xi D 2f mxi iD1

jD1; j¤i

iD1

iD1

for all m 2 N with m 2.

2.2.1 Stability of Homomorphisms in Banach Algebras For any mapping f W A ! B, we define D f .x1 ; ; xm / WD

m X iD1

m m m X X X f mxi C xj C f xi 2f mxi jD1; j¤i

iD1

iD1

for all 2 T1 WD f 2 C W j j D 1g and x1 ; ; xm 2 A. Now, we prove the Hyers-Ulam stability of homomorphisms in Banach algebras for the functional equation D f .x1 ; ; xm / D 0. Assume that A is a complex Banach algebra with norm k kA and B is a complex Banach algebra with norm k kB . Theorem 2.6. Let f W A ! B be a mapping for which there are functions ' W Am ! Œ0; 1/ and W A2 ! Œ0; 1/ such that limj!1 mj '.mj x1 ; ; mj xm / D 0;

(2.8)

kD f .x1 ; ; xm /kB '.x1 ; ; xm /;

(2.9)

kf .xy/ f .x/f .y/kB

.x; y/;

(2.10)

limj!1 m2j .mj x; mj y/ D 0

(2.11)

2.2 Stability of m-Variable Functional Equations

25

for all 2 T1 and x1 ; ; xm ; x; y 2 A. If there exists L < 1 such that '.mx; 0; ; 0/ mL'.x; 0; ; 0/ for all x 2 A, then there exists a unique homomorphism H W A ! B such that kf .x/ H.x/kB

1 '.x; 0; ; 0/ m mL

(2.12)

for all x 2 A. Proof. Consider the set X WD fg W A ! Bg and introduce the generalized metric on X: d.g; h/ D inffC 2 RC W kg.x/ h.x/kB C'.x; 0; ; 0/; 8x 2 Ag; which .X; d/ is complete. Now, we consider the linear mapping J W X ! X such that Jg.x/ WD m1 g.mx/ for all x 2 A. Now, we have d.Jg; Jh/ Ld.g; h/ for all g; h 2 X. Letting D 1, x D x1 and x2 D D xm D 0 in (2.9), we get kf .mx/ mf .x/kB '.x; 0; ; 0/

(2.13)

for all x 2 A and so 1 1 f .x/ f .mx/ '.x; 0; ; 0/ B m m for all x 2 A. Hence d.f ; Jf / m1 . By Theorem 1.3, there exists a mapping H W A ! B such that (1) H is a fixed point of J, i.e., H.mx/ D mH.x/

(2.14)

for all x 2 A. The mapping H is a unique fixed point of J in the set Y D fg 2 X W d.f ; g/ < 1g: This implies that H is a unique mapping satisfying (2.13) such that there exists C 2 .0; 1/ satisfying kH.x/ f .x/kB C'.x; 0; ; 0/ for all x 2 A;

26


(2) d.J n f ; H/ ! 0 as n ! 1. This implies the equality f .mn x/ D H.x/ n!1 mn

(2.15)

lim

for all x 2 A; 1 d.f ; Jf /, which implies the inequality d.f ; H/ (3) d.f ; H/ 1L implies that the inequality (2.12) holds.

1 mmL :

This

Thus it follows from (2.8), (2.9) and (2.14) that m m m m X X X X H mxi C xj C H xi 2H mxi iD1

iD1

jD1; j¤i

iD1

B

m m X 1 X nC1 n f m x C m x i j n!1 mn iD1

D lim

jD1; j¤i

Cf

m X

m X mn xi 2f mnC1 xi

iD1

B

iD1

1 '.mn x1 ; ; mn xm / D 0 n!1 mn

lim

for all x1 ; ; xm 2 A and so m m m m X X X X H mxi C xj C H xi D 2H mxi iD1

iD1

jD1; j¤i

(2.16)

iD1

for all x1 ; ; xm 2 A. By a similar method given in above, we get H.mx/ D H.m x/ for all 2 T1 and x 2 A. Thus one can show that the mapping H W A ! B is C-linear. It follows from (2.10) that 1 kf .mn xy/ f .mn x/f .mn y/kB mn 1 lim n .mn x; mn y/ n!1 m D0

kH.xy/ H.x/H.y/kB D lim

n!1

for all x; y 2 A and so H.xy/ D H.x/H.y/ for all x; y 2 A. Thus H W A ! B is a homomorphism satisfying (2.12). This completes the proof. Corollary 2.7. Let r < 1, be nonnegative real numbers and f W A ! B be a mapping such that kD f .x1 ; ; xm /kB .kx1 krA C kx2 krA C C kxm krA /

(2.17)


27

and kf .xy/ f .x/f .y/kB .kxkrA kykrA /

(2.18)

for all 2 T1 and x1 ; ; xm ; x; y 2 A. Then there exists a unique homomorphism H W A ! B such that kf .x/ H.x/kB

kxkrA m mr

for all x 2 A. Proof. The proof follows from Theorem 2.6 by taking '.x1 ; ; xm / D .kx1 krA C kx2 krA C C kxm krA /; .x; y/ WD .kxkrA kykrA / for all x1 ; ; xm ; x; y 2 A and L D mr1 .

Theorem 2.8. Let f W A ! B be a mapping for which there exist the functions ' W Am ! Œ0; 1/ and W A2 ! Œ0; 1/ such that lim mj '.mj x1 ; ; mj xm / D 0;

(2.19)

kD f .x1 ; ; xm /kB '.x1 ; ; xm /;

(2.20)

j!1

kf .xy/ f .x/f .y/kB

.x; y/;

lim m2j .mj x; mj y/ D 0

j!1

(2.21) (2.22)

for all 2 T1 and x1 ; ; xm ; x; y 2 A. If there exists L < 1 such that '.x; 0; ; 0/ mL '.mx; 0; ; 0/ for all x 2 A, then there exists a unique homomorphism H W A ! B such that kf .x/ H.x/kB

L '.x; 0; ; 0/ m mL

for all x 2 A. Proof. We consider the linear mapping J W X ! X such that Jg.x/ WD mg

x m

(2.23)

28


for all x 2 A. It follows from (2.13) that x x L ; 0; ; 0 '.x; 0; ; 0/ ' f .x/ mf m B m m for all x 2 A and so d.f ; Jf / H W A ! B such that

L . m

By Theorem 1.3, there exists a mapping

(1) H is a fixed point of J, i.e., H.mx/ D mH.x/

(2.24)

for all x 2 A. The mapping H is a unique fixed point of J in the set Y D fg 2 X W d.f ; g/ < 1g: This implies that H is a unique mapping satisfying (2.20) such that there exists C 2 .0; 1/ satisfying kH.x/ f .x/kB C'.x; 0; ; 0/ for all x 2 A; (2) d.J n f ; H/ ! 0 as n ! 1. This implies the equality lim mn f

n!1

x D H.x/ mn

for all x 2 A; 1 (3) d.f ; H/ 1L d.f ; Jf /, which implies the inequality d.f ; H/

L ; m mL

which implies that the inequality (2.23) holds. The rest of the proof is similar to the proof of Theorem 2.6.

Corollary 2.9. Let r > 1, be nonnegative real numbers and let f W A ! B be a mapping such that kD f .x1 ; ; xm /kB .kx1 krA C kx2 krA C C kxm krA /

(2.25)

kf .xy/ f .x/f .y/kB .kxkrA kykrA /

(2.26)

and


29

for all 2 T1 and x1 ; ; xm ; x; y 2 A. Then there exists a unique homomorphism H W A ! B such that kf .x/ H.x/kB

mr

kxkrA m

for all x 2 A. Proof. The proof follows from Theorem 2.6 by taking '.x1 ; ; xm / D .kx1 krA C kx2 krA C C kxm krA /; .x; y/ WD .kxkrA kykrA / for all x1 ; ; xm ; x; y 2 A and L D m1r .

2.2.2 Stability of Derivations in Banach Algebras Now, we prove the Hyers-Ulam stability of derivations on Banach algebras for the functional equation D f .x1 ; ; xm / D 0. Theorem 2.10. Let f W A ! A be a mapping for which there exist the functions ' W Am ! Œ0; 1/ and W A2 ! Œ0; 1/ such that lim mj '.mj x1 ; ; mj xm / D 0;

(2.27)

kD f .x1 ; ; xm /kA '.x1 ; ; xm /;

(2.28)

j!1

kf .xy/ f .x/y xf .y/kA lim m

2j

j!1

.x; y/;

.m x; m y/ D 0 j

j

(2.29) (2.30)

for all 2 T1 and x1 ; ; xm ; x; y 2 A. If there exists L < 1 such that '.mx; 0; ; 0/ mL'.x; 0; ; 0/ for all x 2 A. Then there exists a unique derivation ı W A ! A such that kf .x/ ı.x/kA

1 '.x; 0; ; 0/ m mL

(2.31)

for all x 2 A. Proof. By the same reasoning as in the proof of Theorem 2.6, there exists a unique C-linear mapping ı W A ! A satisfying (2.31). The mapping ı W A ! A is given by f .mn x/ n!1 mn

ı.x/ D lim

(2.32)

30


for all x 2 A. It follows from (2.29), (2.30) and (2.32) that kı.xy/ ı.x/y xı.y/kA 1 kf .m2n xy/ f .mn x/ mn y mn xf .mn y/kA m2n 1 lim 2n .mn x; mn y/ n!1 m D0

D lim

n!1

for all x; y 2 A and so ı.xy/ D ı.x/y C xı.y/ for all x; y 2 A. Thus ı W A ! A is a derivation satisfying (2.31). This completes the proof. Corollary 2.11. Let r < 1, be nonnegative real numbers and f W A ! A be a mapping such that kD f .x1 ; ; xm /kA .kx1 krA C C kxm krA /

(2.33)

kf .xy/ f .x/y xf .y/kA .kxkrA kykrA /

(2.34)

and

for all 2 T1 and x1 ; ; xm ; x; y 2 A. Then there exists a unique derivation ı W A ! A such that kf .x/ ı.x/kA

kxkrA m mr

for all x 2 A. Proof. The proof follows from Theorem 2.10 by taking '.x1 ; ; xm / WD .kx1 krA C kxm krA /; .x; y/ WD .kxkrA kykrA / for all x1 ; ; xm ; x; y 2 A and L D mr1 .

Remark 2.12. Let f W A ! B be a mapping for which there exist the functions ' W Am ! Œ0; 1/ and W A2 ! Œ0; 1/ such that lim mj '.mj x1 ; ; mj xm / D 0;

j!1

(2.35)

2.3 Stability in Quasi-Banach Algebras

31

kD f .x1 ; ; xm /kA '.x1 ; ; xm /; kf .xy/ f .x/y xf .y/kA lim m

j!1

2j

j

.x; y/;

j

.m x; m y/ D 0

(2.36) (2.37) (2.38)

for all 2 T1 and x1 ; ; xm ; x; y 2 A. If there exists L < 1 such that '.mx; 0; ; 0/ mL '.x; 0; ; 0/ for all x 2 A. Then there exists a unique derivation ı W A ! A such that kf .x/ ı.x/kA

L '.x; 0; ; 0/ m mL

(2.39)

for all x 2 A. Corollary 2.13. Let r > 1, be nonnegative real numbers and f W A ! A be a mapping such that kD f .x1 ; ; xm /kA .kx1 krA C kxm krA /

(2.40)

kf .xy/ f .x/y xf .y/kA .kxkrA kykrA /

(2.41)

and

for all 2 T1 and x1 ; ; xm ; x; y 2 A. Then there exists a unique derivation ı W A ! A such that kf .x/ ı.x/kA

kxkrA mr m

for all x 2 A. Proof. Consider Remark 2.12 and take '.x1 ; ; xm / WD .kx1 krA C kxm krA /; .x; y/ WD .kxkrA kykrA / for all x1 ; ; xm ; x; y 2 A and L D m1r .

2.3 Stability in Quasi-Banach Algebras Let q be a positive rational number and n be a nonnegative integer. We consider the Hyers-Ulam stability of homomorphisms in quasi-Banach algebras and generalized derivations on quasi-Banach algebras for the following functional equation:

32

2 Stability of Functional Equations in Banach Algebras n n n n X X X X f q.xi xj / C nf qxi D nq f .xi /: iD1

jD1

iD1

(2.42)

iD1

This is applied to investigate some isomorphisms in quasi-Banach algebras (see [180, 233, 238]).

2.3.1 Stability of Homomorphisms in Quasi-Banach Algebras Let q be a positive rational number. For any mapping f W A ! B, we define Df W An ! B by Df .x1 ; ; xn / n n n n X X X X f q.xi xj / C nf qxi nq f .xi / WD iD1

jD1

iD1

iD1

for all x1 ; ; xn 2 X. Lemma 2.14. Let f W A ! B be a mapping satisfies the functional equation (2.42). Then the mapping f is Cauchy additive and R-linear.

Proof. The proof is easy (see also [229, 267]).

Now, we prove the Hyers-Ulam stability of homomorphisms in quasi-Banach algebras. Theorem 2.15. Assume that r > 2 if nq > 1 and 0 < r < 1 if nq < 1. Let be a positive real number and f W A ! B be an odd mapping such that kDf .x1 ; ; xn /kB

n X

kxj krA

(2.43)

jD1

and kf .xy/ f .x/f .y/kB .kxkrA C kykrA /

(2.44)

for all x; y; x1 ; ; xn 2 A. If f .tx/ is continuous in t 2 R for each fixed x 2 A, then there exists a unique homomorphism H W A ! B such that kf .x/ H.x/kB for all x 2 A.

..nq/pr

1

.nq/p / p

kxkrA

(2.45)


33

Proof. Letting x1 D D xn D x in (2.43), we get knf .nqx/ n2 qf .x/kB nkxkrA

(2.46)

for all x 2 A and so x kxkrA f .x/ nqf nq B .nq/r for all x 2 A. Since B is a p-Banach algebra, we have x x p m .nq/ f .nq/l f .nq/l .nq/m B

m1 X

.nq/j f

jDl

x x p jC1 .nq/ f .nq/j .nq/jC1 B

(2.47)

m1 p X .nq/pj pr kxkA .nq/pr jDl .nq/prj

for all m 1, l with m > l and x 2 A. It follows from (2.47) that the sequence x f.nq/d f . .nq/ d /g is a Cauchy sequence for all x 2 A. Since B is complete, the sequence x d f.nq/ f . .nq/ d /g converges. So one can define a mapping H W A ! B by H.x/ WD lim .nq/d f . d!1

x / .nq/d

for all x 2 A. Moreover, letting l D 0 and m ! 1 in (2.47), we get (2.46). It follows from (2.43) that x xn 1 kDH.x1 ; ; xn /kB D lim .nq/d Df ; ; d!1 .nq/d .nq/d B n .nq/d X kxj krA d!1 .nq/dr jD1

lim D0 for all x1 ; ; xn 2 A. Thus we have

DH.x1 ; ; xn / D 0 for all x1 ; ; xn 2 A. By (2.43), the mapping H W A ! B is Cauchy additive and R-linear. It follows from (2.45) that

34


kH.xy/ H.x/H.y/kB D lim .nq/2d f d!1

x y xy f f .nq/d .nq/d .nq/d .nq/d B

.nq/2d .kxkrA C kykrA / d!1 .nq/dr

lim D0 for all x; y 2 A and so

H.xy/ D H.x/H.y/ for all x; y 2 A. Now, let T W A ! B be another mapping satisfying (2.46). Then we have kH.x/ T.x/kB x x T D .nq/d H .nq/d .nq/d B x x x x f f .nq/d K H C T .nq/d .nq/d B .nq/d .nq/d B

2 .nq/d K ..nq/pr

1

.nq/p / p .nq/dr

kxkrA ;

which tends to zero as n ! 1 for all x 2 A. So we can conclude that H.x/ D T.x/ for all x 2 A. This proves the uniqueness of H. Thus the mapping H W A ! B is a unique homomorphism satisfying (2.46). This completes the proof. Theorem 2.16. Assume that 0 < r < 1 if nq > 1 and that r > 2 if nq < 1. Let be a positive real number, and let f W A ! B be an odd mapping satisfying (2.43) and (2.45). If f .tx/ is continuous in t 2 R for each fixed x 2 A, then there exists a unique homomorphism H W A ! B such that kf .x/ H.x/kB

..nq/p

1

.nq/pr / p

for all x 2 A. Proof. It follows from (2.46) that kf .x/

1 f .nqx/kB kxkrA nq nq

for all x 2 A. Since B is a p-Banach algebra, we have

kxkrA

(2.48)


35

1 p 1 l m f ..nq/ x/ f ..nq/ x/ B .nq/l .nq/m

jDl

p 1 1 j jC1 f ..nq/ x/ f ..nq/ x/ B .nq/j .nq/jC1

m1 X

(2.49)

m1 p X .nq/prj pr kxkA .nq/p jDl .nq/pj

for all m 1, l with m > l and x 2 A. It follows from (2.49) that the sequence 1 d f .nq/ d f ..nq/ x/g is a Cauchy sequence for all x 2 A. Since B is complete, the 1 d sequence f .nq/ d f ..nq/ x/g converges. So one can define a mapping H W A ! B by

H.x/ WD lim

d!1

1 f ..nq/d x/ .nq/d

for all x 2 A. Moreover, letting l D 0 and m ! 1 in (2.49), we get (2.48). The rest of the proof is similar to the proof of Theorem 2.15.

2.3.2 Isomorphisms in Quasi-Banach Algebras Assume that A is a quasi-Banach algebra with the quasi-norm k kA and the unit e and B is a p-Banach algebra with the p-norm k kB and the unit e0 . Let K be the modulus of concavity of k kB . Now, we consider isomorphisms in quasi-Banach algebras. Theorem 2.17. Assume that r > 2 if nq > 1 and that 0 < r < 1 if nq < 1. Let be a positive real number and f W A ! B be an odd bijective mapping satisfying (2.43) such that f .xy/ D f .x/f .y/

(2.50)

e 0 for all x; y 2 A. If limd!1 .nq/d f . .nq/ d / D e and f .tx/ is continuous in t 2 R for each fixed x 2 A, then the mapping f W A ! B is an isomorphism.

Proof. The condition (2.50) implies that f W A ! B satisfies (2.45). By the same reasoning as in the proof of Theorem 2.15, there exists a unique homomorphism H W A ! B, which is defined by H.x/ WD lim .nq/d f d!1

for all x 2 A. Thus we have

x .nq/d

36


H.x/ D H.ex/ D lim .nq/d f d!1

D lim .nq/d f d!1

ex e d D lim .nq/ f x d!1 .nq/d .nq/d

e f .x/ D e0 f .x/ D f .x/ .nq/d

for all x 2 A. So the bijective mapping f W A ! B is an isomorphism. This completes the proof. Remark 2.18. Assume that 0 < r < 1 if nq > 1 and that r > 2 if nq < 1. Let be a positive real number and f W A ! B be an odd bijective mapping satisfying (2.43) and (2.50). If f .tx/ is continuous in t 2 R for each fixed x 2 A and lim

d!1

1 f ..nq/d e/ D e0 ; .nq/d

then the mapping f W A ! B is an isomorphism.

2.3.3 Stability of Generalized Derivations in Quasi-Banach Algebras Assume that A is a p-Banach algebra with the p-norm k kA . Let K be the modulus of concavity of k kA . Definition 2.19 ([18]). A generalized derivation ı W A ! A is R-linear and fulfills the generalized Leibniz rule: ı.xyz/ D ı.xy/z xı.y/z C xı.yz/ for all x; y; z 2 A. Now, we prove the Hyers-Ulam stability of generalized derivations on quasiBanach algebras. Theorem 2.20. Assume that r > 3 if nq > 1 and that 0 < r < 1 if nq < 1. Let be a positive real number and f W A ! A be an odd mapping satisfying (2.43) such that kf .xyz/ f .xy/z C xf .y/z xf .yz/kA .kxkrA C kykrA C kzkrA /

(2.51)

for all x; y; z 2 A. If f .tx/ is continuous in t 2 R for each fixed x 2 A, then there exists a unique generalized derivation ı W A ! A such that


37

kf .x/ ı.x/kA

..nq/pr

1

.nq/p / p

kxkrA

(2.52)

for all x 2 A. Proof. By the same reasoning as in the proof of Theorem 2.15, there exists a unique R-linear mapping ı W A ! A satisfying (2.52). The mapping ı W A ! A is defined by ı.x/ WD lim .nq/d f d!1

x .nq/d

for all x 2 A. It follows from (2.51) that kı.xyz/ ı.xy/z C xı.y/z xı.yz/kA xyz xy z f D lim .nq/3d f d!1 .nq/3d .nq/2d .nq/d y x y x yz C f f .nq/d .nq/d .nq/d .nq/d .nq/2d A .nq/3d .kxkrA C kykrA C kzkrA / d!1 .nq/dr

lim D0 for all x; y; z 2 A and so

ı.xyz/ D ı.xy/z xı.y/z C xı.yz/ for all x; y; z 2 A. Thus the mapping ı W A ! A is a unique generalized derivation satisfying (2.52). This completes the proof. Theorem 2.21. Assume that 0 < r < 1 if nq > 1 and that r > 3 if nq < 1. Let be a positive real number and f W A ! A be an odd mapping satisfying (2.43) and (2.51). If f .tx/ is continuous in t 2 R for each fixed x 2 A, then there exists a unique generalized derivation ı W A ! A such that kf .x/ ı.x/kA

..nq/p

1

.nq/pr / p

kxkrA

(2.53)

for all x 2 A. Proof. By the same reasoning as in the proof of Theorem 2.16, there exists a unique R-linear mapping ı W A ! A satisfying (2.53). The mapping ı W A ! A is defined by 1 f ..nq/d x/ d!1 .nq/d

ı.x/ WD lim

for all x 2 A. The rest of the proof is similar to the proof of Theorem 2.20.

38


2.4 Stability of Cauchy–Jensen Functional Equations In this section, we prove the Hyers-Ulam stability of homomorphisms in real Banach algebras and generalized derivations on real Banach algebras for the following Cauchy-Jensen functional equations(see also [19, 147, 232]): f

x C y 2

x y Cz Cf C z D f .x/ C 2f .z/ 2

and 2f

x C y 2

C z D f .x/ C f .y/ C 2f .z/:

2.4.1 Stability of Homomorphisms in Real Banach Algebras Assume that A is a real Banach algebra with the norm k kA and B is a real Banach algebra with the norm k kB . For any mapping f W A ! B, we define Cf .x; y; z/ WD f

x C y 2

x y Cz Cf C z f .x/ 2f .z/ 2

for all x; y; z 2 A. Now, we prove the Hyers-Ulam stability of homomorphisms in real Banach algebras for the functional equation Cf .x; y; z/ D 0. Lemma 2.22. Let X and Y be vector spaces. If a mapping f W X ! Y satisfies f

x C y 2

f

x y Cz Cf C z D f .x/ C 2f .z/; 2

x C y 2

x y Cz f C z D f .y/ 2

(2.54)

(2.55)

or 2f

x C y 2

C z D f .x/ C f .y/ C 2f .z/

for all x; y; z 2 X, then the mapping f W X ! Y is Cauchy additive. Proof. Letting x D y in (2.54), we get f .x C z/ C f .z/ D f .x/ C 2f .z/

(2.56)

2.4 Stability of Cauchy–Jensen Functional Equations

39

for all x; z 2 X and so f .x C z/ D f .x/ C f .z/ for all x; z 2 X. Hence f W X ! Y is Cauchy additive. Letting x D y in (2.55), we get f .x C z/ f .z/ D f .x/ for all x; z 2 X and so f .x C z/ D f .x/ C f .z/ for all x; z 2 X. Hence f W X ! Y is Cauchy additive. Letting x D y in (2.56), we get 2f .x C z/ D 2f .x/ C 2f .z/ for all x; z 2 X and so f .x C z/ D f .x/ C f .z/ for all x; z 2 X. Hence f W X ! Y is Cauchy additive. This completes the proof. The mappings f W X ! Y given in the statement of Lemma 2.22 are called Cauchy–Jensen type additive mappings. Putting z D 0 in (2.56), we get the Jensen additive mapping 2f . xCy 2 / D f .x/ C f .y/ and, putting x D y in (2.56), we get the Cauchy additive mapping f .x C z/ D f .x/ C f .z/. Theorem 2.23. Let f W A ! B be a mapping for which there exists a function ' W A3 ! Œ0; 1/ such that 1 X 1 '.2j x; 2j y; 2j z/ < 1; j 2 jD0

(2.57)

kCf .x; y; z/kB '.x; y; z/;

(2.58)

kf .xy/ f .x/f .y/kB '.x; y; 0/

(2.59)

for all x; y; z 2 A. If there exists L < 1 such that x x x '.x; x; x/ 2L'. ; ; / 2 2 2 for all x 2 A and f .tx/ is continuous in t 2 R for each fixed x 2 A, then there exists a unique homomorphism H W A ! B such that kf .x/ H.x/kB

1 '.x; x; x/ 2 2L

for all x 2 A. Proof. Consider the set X WD fg W A ! Bg

(2.60)

40


and introduce the generalized metric on X defined by d.g; h/ D inffC 2 RC W kg.x/ h.x/kB C'.x; x; x/; 8x 2 Ag; which .X; d/ is complete. Now, we consider the linear mapping J W X ! X such that Jg.x/ WD

1 g.2x/ 2

for all x 2 A. Note that d.Jg; Jh/ Ld.g; h/ for all g; h 2 X. Letting y D z D x in (2.58), we get kf .2x/ 2f .x/kB '.x; x; x/

(2.61)

for all x 2 A and so 1 1 f .x/ f .2x/ '.x; x; x/ B 2 2 for all x 2 A. Hence d.f ; Jf / H W A ! B such that

1 2.


(1) H is a fixed point of J, i.e., H.2x/ D 2H.x/

(2.62)

for all x 2 A. The mapping H is a unique fixed point of J in the set Y D fg 2 X W d.f ; g/ < 1g: This implies that H is a unique mapping satisfying (2.62) such that there exists C 2 .0; 1/ satisfying kH.x/ f .x/kB C'.x; x; x/ for all x 2 A; (2) d.J n f ; H/ ! 0 as n ! 1. This implies the equality f .2n x/ D H.x/ n!1 2n lim

for all x 2 A;

(2.63)

2.4 Stability of Cauchy–Jensen Functional Equations

(3) d.f ; H/

1 1L d.f ; Jf /,

41

which implies the inequality d.f ; H/

1 : 2 2L

This implies that the inequality (2.60) holds. It follows from (2.57), (2.58) and (2.63) that x C y x y Cz CH C z H.x/ 2H.z/ H B 2 2 1 D lim n kf .2n1 .x C y/ C 2n z/ C f .2n1 .x y/ C 2n z/ n!1 2 f .2n x/ 2f .2n z/kB lim

n!1

1 '.2n x; 2n y; 2n z/ 2n

D0 for all x; y; z 2 A and so H

x C y 2

x y Cz CH C z D H.x/ C 2H.z/ 2

for all x; y; z 2 A. By Lemma 2.22, the mapping H W A ! B is Cauchy additive. By (2.58), the mapping H W A ! B is R-linear. It follows from (2.59) that 1 kf .4n xy/ f .2n x/f .2n y/kB 4n 1 lim n '.2n x; 2n y; 0/ n!1 4 1 lim n '.2n x; 2n y; 0/ n!1 2 D0


n!1

for all x; y 2 A and so H.xy/ D H.x/H.y/ for all x; y 2 A. Thus H W A ! B is a homomorphism satisfying (2.60). This completes the proof. Corollary 2.24. Let r < 1, be nonnegative real numbers and f W A ! B be a mapping such that kCf .x; y; z/kB .kxkrA C kykrA C kzkrA /

(2.64)

42


and kf .xy/ f .x/f .y/kB .kxkrA C kykrA /

(2.65)

for all x; y; z 2 A. If f .tx/ is continuous in t 2 R for each fixed x 2 A, then there exists a unique homomorphism H W A ! B such that kf .x/ H.x/kB

3 kxkrA 2 2r

for all x 2 A. Proof. The proof follows from Theorem 2.23 by taking '.x; y; z/ WD .kxkrA C kykrA C kzkrA / for all x; y; z 2 A and L D 2r1 . We get the desired result.

Theorem 2.25. Let f W A ! B be a mapping for which there exists a function ' W A3 ! Œ0; 1/ satisfying (2.58) and (2.59) such that 1 X jD0

4j '

x y z 1, be nonnegative real numbers and f W A ! B be a mapping such that 1 xCy / ; f .z x/ C f .z y/ C f .x C y/ 2f .z B B 2 4

(3.19)

kf .xy/ f .x/f .y/kB kxkrA kykrA ;

(3.20)

kf .x / f .x/ kB 2kxkrA

(3.21)

for all 2 T1 WD f 2 C W jj D 1g and x; y; z 2 A. Then the mapping f W A ! B is a C -algebra homomorphism. Proof. Let D 1 in (3.19). By Theorem 3.11, the mapping f W A ! B is additive. Letting y D x and z D 0 in (3.19), we get kf . x/ C f .x/kB k2f .0/kB D 0 for all x 2 A and 2 T1 and so f . x/ C f .x/ D f . x/ C f .x/ D 0 for all x 2 A and 2 T1 . Hence f . x/ D f .x/ for all x 2 A and 2 T1 . So, the mapping f W A ! B is C-linear. It follows from (3.20) that xy x y kf .xy/ f .x/f .y/kB D lim 4n f n n f n f n n!1 B 2 2 2 2 n 4 lim nr kxkrA kykrA n!1 4 D0 for all x; y 2 A and so f .xy/ D f .x/f .y/ for all x; y 2 A. It follows from (3.21) that x x kf .x / f .x/ kB D lim 2n f n f n n!1 B 2 2 2nC1 kxkrA n!1 2nr D0 lim

3.2 Apollonius Type Additive Functional Equations

67

for all x 2 A and so f .x / D f .x/ for all x 2 A. Therefore, the mapping f W A ! B is a C -algebra homomorphism. This completes the proof. Remark 3.14. Let r < 1, be positive real numbers and f W A ! B be a mapping satisfying (3.19), (3.20) and (3.21). Then the mapping f W A ! B is a C -algebra homomorphism. Theorem 3.15. Let r > 1, be nonnegative real numbers and f W A ! A be a mapping satisfying (3.21) such that kf .xy/ f .x/y xf .y/kA kxkrA kykrA

(3.22)

for all x; y 2 A. Then the mapping f W A ! A is a linear derivation. Proof. By applying Lemma 3.12, the mapping f W A ! A is C-linear. It follows from (3.22) that xy xy x y kf .xy/ f .x/y xf .y/kA D lim 4n f n f n n n f n n!1 A 4 2 2 2 2 4n lim nr kxkrA kykrA n!1 4 D0 for all x; y 2 A and so f .xy/ D f .x/y C xf .y/ for all x; y 2 A. Thus the mapping f W A ! A is a linear derivation. This completes the proof. Remark 3.16. Let r < 1, be positive real numbers and f W A ! A be a mapping satisfying (3.19) and (3.22). Then the mapping f W A ! A is a linear derivation.

3.2.2 Homomorphisms and Derivations in Lie C -Algebras Assume that A is a Lie C -algebra with the norm k kA and B is a Lie C -algebra with the norm k kB . We recall that a C-linear mapping H W A ! B is called a Lie C -algebra homomorphism if H W A ! B satisfies the following: H.Œx; y/ D ŒH.x/; H.y/

3 Stability of Functional Equations in C -Algebras

68

for all x; y 2 A. A C-linear mapping D W A ! A is called a Lie derivation if D W A ! A satisfies the following: D.Œx; y/ D ŒD.x/; y C Œx; D.y/ for all x; y 2 A. Now, we investigate Lie C -algebra homomorphisms in Lie C -algebras and Lie derivations on Lie C -algebras associated with the Apollonius type additive functional equation. Theorem 3.17. Let r > 1, be nonnegative real numbers and f W A ! B be a mapping satisfying (3.19) such that kf .Œx; y/ Œf .x/; f .y/kB kxkrA kykrA

(3.23)

for all x; y 2 A. Then the mapping f W A ! B is a Lie C -algebra homomorphism. Proof. It is straight forward to show that, the mapping f W A ! B is C-linear. It follows from (3.23) that Œx; y h x y i f .Œx; y/ Œf .x/; f .y/ D lim 4n f n n f n ; f n n!1 B B 2 2 2 2 4n lim nr kxkrA kykrA n!1 4 D0 for all x; y 2 A and so f .Œx; y/ D Œf .x/; f .y/ for all x; y 2 A. Hence the mapping f W A ! B is a Lie C -algebra homomorphism. This completes the proof. Remark 3.18. If r < 1, is positive real numbers and f W A ! B be a mapping satisfying (3.19) and (3.23). Then the mapping f W A ! B is a Lie C -algebra homomorphism. Theorem 3.19. Let r > 1, be nonnegative real numbers and f W A ! A be a mapping satisfying (3.19) such that kf .Œx; y/ Œf .x/; y Œx; f .y/kA kxkrA kykrA

(3.24)

for all x; y 2 A. Then the mapping f W A ! A is a Lie derivation. Proof. It is straight forward to show that, the mapping f W A ! A is C-linear. It follows from (3.24) that


69

kf .Œx; y/ Œf .x/; y Œx; f .y/kA Œx; y h x y i h x y i f n ; n n;f n D lim 4n f n!1 A 4n 2 2 2 2 n 4 lim nr kxkrA kykrA n!1 4 D0 for all x; y 2 A and so f .Œx; y/ D Œf .x/; y C Œx; f .y/ for all x; y 2 A. Thus the mapping f W A ! A is a Lie derivation. This completes the proof. Remark 3.20. If r < 1, is positive real numbers and f W A ! A be a mapping satisfying (3.19) and (3.24). Then the mapping f W A ! A is a Lie derivation.

3.2.3 Homomorphisms and Derivations in JC -Algebras Assume that A is a JC -algebra with the norm k kA and B is a JC -algebra with the norm k kB . A C-linear mapping H W A ! B is called a JC -algebra homomorphism if H W A ! B satisfies the following: H.x ı y/ D H.x/ ı H.y/ for all x; y 2 A. A C-linear mapping D W A ! A is called a Jordan derivation if D W A ! A satisfies the following: D.x ı y/ D D.x/ ı y C x ı D.y/ for all x; y 2 A. Now, we investigate JC -algebra homomorphisms between JC -algebras and Jordan derivations on JC -algebras associated with the Apollonius type additive functional equation. Remark 3.21. Let r > 1, be nonnegative real numbers and f W A ! B be a mapping satisfying (3.19) such that kf .x ı y/ f .x/ ı f .y/kB kxkrA kykrA

(3.25)

for all x; y 2 A. Then the mapping f W A ! B is a JC -algebra homomorphism.


70

Remark 3.22. Let r < 1, be positive real numbers and f W A ! B be a mapping satisfying (3.19) and (3.25). Then the mapping f W A ! B is a JC -algebra homomorphism. Remark 3.23. Let r > 1, be nonnegative real numbers and f W A ! A be a mapping satisfying (3.19) such that kf .x ı y/ f .x/ ı y x ı f .y/kA kxkrA kykrA

(3.26)

for all x; y 2 A. Then the mapping f W A ! A is a Jordan derivation. Remark 3.24. Let r < 1, be positive real numbers and f W A ! A be a mapping satisfying (3.19) and (3.26). Then the mapping f W A ! A is a Jordan derivation.

3.3 Stability of Jensen Type Functional Equations in C -Algebras Using the fixed point method, we consider [250] the Hyers-Ulam stability of homomorphisms in C -algebras and Lie C -algebras and derivations on C algebras and Lie C -algebras for the following Jensen type functional equation: f

x C y 2

Cf

x y 2

D f .x/:

3.3.1 Stability of Homomorphisms in C -Algebras Assume that A is a C -algebra with the norm k kA and B is a C -algebra with the norm k kB . For any mapping f W A ! B, we define D f .x; y/ WD f

x C y 2

C f

x y 2

f . x/

for all 2 T1 WD f 2 C W j j D 1g and x; y 2 A. Now, we prove the Hyers-Ulam stability of homomorphisms in C -algebras for the functional equation D f .x; y/ D 0. Theorem 3.25. Let f W A ! B be a mapping for which there exists a function ' W A2 ! Œ0; 1/ such that 1 X jD0

2j '.2j x; 2j y/ < 1;

(3.27)

3.3 Stability of Jensen Type Functional Equations in C -Algebras

71

kD f .x; y/kB '.x; y/;

(3.28)

kf .xy/ f .x/f .y/kB '.x; y/;

(3.29)

kf .x / f .x/ kB '.x; x/

(3.30)

for all 2 T1 and x; y 2 A. If there exists L < 1 such that x '.x; 0/ 2L' ; 0 2 for all x 2 A, then there exists a unique C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB

L '.x; 0/ 1L

(3.31)

for all x 2 A. Proof. Consider the set X WD fg W A ! Bg and introduce the generalized metric on X defined by d.g; h/ D inffC 2 RC W kg.x/ h.x/kB C'.x; 0/; 8x 2 Ag; which .X; d/ is complete. Now, we consider the linear mapping J W X ! X such that Jg.x/ WD

1 g.2x/ 2

for all x 2 A. Thus we have d.Jg; Jh/ Ld.g; h/ for all g; h 2 X. Letting D 1 and y D 0 in (3.28), we get x f .x/ '.x; 0/ 2f B 2 for all x 2 A and so 1 1 f .x/ f .2x/ '.2x; 0/ L'.x; 0/ B 2 2 for all x 2 A. Hence d.f ; Jf / L.

(3.32)


72

By Theorem 1.3, there exists a mapping H W A ! B such that (1) H is a fixed point of J, i.e., H.2x/ D 2H.x/

(3.33)

for all x 2 A. The mapping H is a unique fixed point of J in the set Y D fg 2 X W d.f ; g/ < 1g: This implies that H is a unique mapping satisfying (3.33) such that there exists C 2 .0; 1/ satisfying kH.x/ f .x/kB C'.x; 0/ for all x 2 A; (2) d.J n f ; H/ ! 0 as n ! 1. This implies the equality f .2n x/ D H.x/ n!1 2n lim

(3.34)

for all x 2 A; 1 d.f ; Jf /, which implies the inequality (3) d.f ; H/ 1L d.f ; H/

L : 1L

This implies that the inequality (3.31) holds. It follows from (3.27), (3.28) and (3.34) that x C y x y CH H.x/ H B 2 2 1 D lim n kf .2n1 .x C y// C f .2n1 .x y// f .2n x/kB n!1 2 1 lim n '.2n x; 2n y/ n!1 2 D0 for all x; y 2 A and so H

x C y

for all x; y 2 A. Letting z D

2 xCy 2

CH

x y

and w D

2 xy 2

D H.x/

in (3.35), we get

(3.35)


73

H.z/ C H.w/ D H.z C w/ for all z; w 2 A. So the mapping H W A ! B is Cauchy additive, i.e., H.z C w/ D H.z/ C H.w/ for all z; w 2 A. Letting y D x in (3.28), we get f .x/ D f . x/ for all 2 T1 and x 2 A. By a similar method to above, we get H.x/ D H. x/ for all 2 T1 and x 2 A. Thus one can show that the mapping H W A ! B is C-linear. It follows from (3.29) that 1 kf .4n xy/ f .2n x/f .2n y/kB 4n 1 lim n '.2n x; 2n y/ n!1 4 1 lim n '.2n x; 2n y/ n!1 2 D0


n!1

for all x; y 2 A. Then we have H.xy/ D H.x/H.y/ for all x; y 2 A. It follows from (3.30) that 1 kf .2n x / f .2n x/ kB n!1 2n 1 lim n '.2n x; 2n x/ n!1 2 D0

kH.x / H.x/ kB D lim

for all x 2 A. Then we have H.x / D H.x/ for all x 2 A. Thus H W A ! B is a C -algebra homomorphism satisfying (3.31). This completes the proof.


74

Corollary 3.26. Let r < 1, be nonnegative real numbers and f W A ! B be a mapping such that kD f .x; y/kB .kxkrA C kykrA /; kf .xy/ f .x/f .y/kB

.kxkrA

C

kykrA /;

kf .x / f .x/ kB 2kxkrA

(3.36) (3.37) (3.38)

for all 2 T1 and x; y 2 A. Then there exists a unique C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB

2r kxkrA 2 2r

(3.39)

for all x 2 A. Proof. The proof follows from Theorem 3.25 by taking '.x; y/ WD .kxkrA C kykrA /

for all x; y 2 A and L D 2r1 .

Theorem 3.27. Let f W A ! B be a mapping for which there exists a function ' W A2 ! Œ0; 1/ satisfying (3.28), (3.29) and (3.30) such that 1 X jD0

4j '

x y ; 2, be nonnegative real numbers and f W A ! B be a mapping satisfying (3.36), (3.37) and (3.38). Then there exists a unique C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB for all x 2 A.

kxkrA 2r 2

(3.43)


76

Proof. The proof follows from Theorem 3.27 by taking '.x; y/ WD .kxkrA C kykrA / for all x; y 2 A and L D 21r .

Theorem 3.29. Let f W A ! B be an odd mapping for which there exists a function ' W A2 ! Œ0; 1/ satisfying (3.27), (3.28), (3.29) and (3.30). If there exists L < 1 such that '.x; 3x/ 2L'

x 3x ; 2 2

for all x 2 A, then there exists a unique C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB

1 '.x; 3x/ 2 2L

(3.44)

for all x 2 A. Proof. Consider the set X WD fg W A ! Bg and introduce the generalized metric on X defined by d.g; h/ D inffC 2 RC W kg.x/ h.x/kB C'.x; 3x/; 8x 2 Ag; which .X; d/ is complete. Now, we consider the linear mapping J W X ! X such that Jg.x/ WD

1 g.2x/ 2

for all x 2 A. Now, we have d.Jg; Jh/ Ld.g; h/ for all g; h 2 X. Letting D 1 and replacing y by 3x in (3.28), we get kf .2x/ 2f .x/kB '.x; 3x/ for all x 2 A and so 1 1 kf .x/ f .2x/kB '.x; 3x/ 2 2

(3.45)


77

for all x 2 A. Hence d.f ; Jf / 12 . By Theorem 1.3, there exists a mapping H W A ! B such that (1) H is a fixed point of J, i.e., H.2x/ D 2H.x/

(3.46)

for all x 2 A. The mapping H is a unique fixed point of J in the set Y D fg 2 X W d.f ; g/ < 1g: This implies that H is a unique mapping satisfying (3.46) such that there exists C 2 .0; 1/ satisfying kH.x/ f .x/kB C'.x; 3x/ for all x 2 A; (2) d.J n f ; H/ ! 0 as n ! 1. This implies the equality f .2n x/ D H.x/ n!1 2n lim


1 : 2 2L

This implies that the inequality (3.44) holds. The rest of the proof is similar to the proof of Theorem 3.25. This completes the proof. Corollary 3.30. Let r < 12 , be nonnegative real numbers and f W A ! B be an odd mapping such that kD f .x; y/kB kxkrA kykrA ; kf .xy/ f .x/f .y/kB

kxkrA

kykrA ;

kf .x / f .x/ kB kxk2r A

(3.47) (3.48) (3.49)

for all 2 T1 and x; y 2 A. Then there exists a unique C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB for all x 2 A.

3r kxk2r A 2 22r

(3.50)


78

Proof. The proof follows from Theorem 3.29 by taking '.x; y/ WD kxkrA kykrA for all x; y 2 A and L D 22r1 .

Theorem 3.31. Let f W A ! B be an odd mapping for which there exists a function ' W A2 ! Œ0; 1/ satisfying (3.28), (3.29), (3.30) and (3.40). If there exists L < 1 such that '.x; 3x/

1 L'.2x; 6x/ 2

for all x 2 A, then there exists a unique C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB

L '.x; 3x/ 2 2L

(3.51)

for all x 2 A. Proof. We consider the linear mapping J W X ! X such that Jg.x/ WD 2g

x 2

for all x 2 A. It follows from (3.45) that x 3x L x '.x; 3x/ f .x/ 2f ' ; 2 B 2 2 2 for all x 2 A and hence d.f ; Jf / H W A ! B such that

L . 2



(3.52)

for all x 2 A. The mapping H is a unique fixed point of J in the set Y D fg 2 X W d.f ; g/ < 1g: This implies that H is a unique mapping satisfying (3.52) such that there exists C 2 .0; 1/ satisfying kH.x/ f .x/kB C'.x; 3x/ for all x 2 A;


79

(2) d.J n f ; H/ ! 0 as n ! 1. This implies the equality lim 2n f

n!1

x D H.x/ 2n


L ; 2 2L

which implies that the inequality (3.51) holds. The rest of the proof is similar to the proof of Theorem 3.25. This completes the proof. Corollary 3.32. Let r > 1, be nonnegative real numbers and f W A ! B be an odd mapping satisfying (3.47), (3.48) and (3.49). Then there exists a unique C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB

kxk2r A 22r 2

(3.53)

for all x 2 A. Proof. The proof follows from Theorem 3.31 by taking '.x; y/ WD kxkrA kykrA for all x; y 2 A and L D 212r .

3.3.2 Stability of Derivations in C -Algebras Now, we prove the Hyers-Ulam stability of derivations on C -algebras for the functional equation D f .x; y/ D 0. Theorem 3.33. Let f W A ! A be a mapping for which there exists a function ' W A2 ! Œ0; 1/ satisfying (3.27) such that kD f .x; y/kA '.x; y/

(3.54)

kf .xy/ f .x/y xf .y/kA '.x; y/

(3.55)

and

for all 2 T1 and x; y 2 A. If there exists L < 1 such that


80

'.x; 0/ 2L'

x 2

;0

for all x 2 A. Then there exists a unique derivation ı W A ! A such that kf .x/ ı.x/kA

L '.x; 0/ 1L

(3.56)

for all x 2 A. Proof. It is straight forward to show that, there exists a unique involutive C-linear mapping ı W A ! A satisfying (3.56). The mapping ı W A ! A is given by f .2n x/ n!1 2n

ı.x/ D lim for all x 2 A. It follows from (3.55) that kı.xy/ ı.x/y xı.y/kA

1 kf .4n xy/ f .2n x/ 2n y 2n xf .2n y/kA 4n 1 lim n '.2n x; 2n y/ n!1 4 1 lim n '.2n x; 2n y/ n!1 2 D0

D lim

n!1

for all x; y 2 A and so ı.xy/ D ı.x/y C xı.y/ for all x; y 2 A. Thus ı W A ! A is a derivation satisfying (3.56). This completes the proof. Corollary 3.34. Let r < 1, be nonnegative real numbers and f W A ! A be a mapping such that kD f .x; y/kA .kxkrA C kykrA /

(3.57)

kf .xy/ f .x/y xf .y/kA .kxkrA C kykrA /

(3.58)

and

for all 2 T1 and x; y 2 A. Then there exists a unique derivation ı W A ! A such that kf .x/ ı.x/kA for all x 2 A.

2r kxkrA 2 2r

(3.59)


81

Proof. The proof follows from Theorem 3.33 by taking '.x; y/ WD .kxkrA C kykrA /

for all x; y 2 A and L D 2r1 .

Remark 3.35. If f W A ! A is a mapping for which there exists a function ' W A2 ! Œ0; 1/ satisfying (3.40), (3.54) and (3.55). If there exists L < 1 such that '.x; 0/ 12 L'.2x; 0/ for all x 2 A, then there exists a unique derivation ı W A ! A such that kf .x/ ı.x/kA

L '.x; 0/ 2 2L

(3.60)

for all x 2 A. Corollary 3.36. Let r > 2, be nonnegative real numbers and f W A ! A be a mapping satisfying (3.57) and (3.58). Then there exists a unique derivation ı W A ! A such that kf .x/ ı.x/kA

kxkrA 2r 2

(3.61)

for all x 2 A. Proof. The proof follows from Theorem 3.35 by taking '.x; y/ WD .kxkrA C kykrA / for all x; y 2 A and L D 21r .

Remark 3.37. For the inequalities controlled by the product of powers of norms, one can obtain similar results to Theorems 3.29 and 3.31 and Corollaries 3.30 and 3.32.

3.3.3 Stability of Homomorphisms in Lie C -Algebras Assume that A is a Lie C -algebra with the norm k kA and B is a Lie C -algebra with the norm k kB . Now, we prove the Hyers-Ulam stability of homomorphisms in Lie C -algebras for the functional equation D f .x; y/ D 0. Theorem 3.38. Let f W A ! B be a mapping for which there exists a function ' W A2 ! Œ0; 1/ satisfying (3.27) and (3.28) such that kf .Œx; y/ Œf .x/; f .y/kB '.x; y/

(3.62)


82

for all x; y 2 A. If there exists L < 1 such that '.x; 0/ 2L'

x 2

;0

for all x 2 A, then there exists a unique Lie C -algebra homomorphism H W A ! B satisfying (3.31). Proof. It is straight forward to show that, there exists a unique C-linear mapping H W A ! B satisfying (3.31). The mapping H W A ! B is given by f .2n x/ n!1 2n

H.x/ D lim for all x 2 A. It follows from (3.62) that kH.Œx; y/ ŒH.x/; H.y/kB

1 kf .4n Œx; y/ Œf .2n x/; f .2n y/kB 4n 1 lim n '.2n x; 2n y/ n!1 4 1 lim n '.2n x; 2n y/ n!1 2 D0

D lim

n!1

for all x; y 2 A and so H.Œx; y/ D ŒH.x/; H.y/ for all x; y 2 A. Thus H W A ! B is a Lie C -algebra homomorphism satisfying (3.31). This completes the proof. Corollary 3.39. Let r < 1, be nonnegative real numbers and f W A ! B be a mapping satisfying (3.36) such that kf .Œx; y/ Œf .x/; f .y/kB .kxkrA C kykrA /

(3.63)

for all x; y 2 A. Then there exists a unique Lie C -algebra homomorphism H W A ! B satisfying (3.39). Proof. The proof follows from Theorem 3.38 by taking '.x; y/ WD .kxkrA C kykrA / for all x; y 2 A and L D 2r1 .


83

Remark 3.40. If f W A ! B is a mapping for which there exists a function ' W A2 ! Œ0; 1/ satisfying (3.28), (3.40) and (3.62). If there exists L < 1 such that '.x; 0/ 12 L'.2x; 0/ for all x 2 A, then there exists a unique Lie C -algebra homomorphism H W A ! B satisfying (3.41). Corollary 3.41. Let r > 2, be nonnegative real numbers and f W A ! B be a mapping satisfying (3.36) and (3.63). Then there exists a unique Lie C -algebra homomorphism H W A ! B satisfying (3.43). Proof. The proof follows from Theorem 3.40 by taking '.x; y/ WD .kxkrA C kykrA / for all x; y 2 A and L D 21r .

Remark 3.42. For the inequalities controlled by the product of powers of norms, one can obtain similar results to Theorems 3.29 and 3.31 and their corollaries.

3.3.4 Stability of Lie Derivations in C -Algebras Assume that A is a Lie C -algebra with the norm k kA . Now, we prove the Hyers-Ulam stability of derivations on Lie C -algebras for the functional equation D f .x; y/ D 0. Theorem 3.43. Let f W A ! A be a mapping for which there exists a function ' W A2 ! Œ0; 1/ satisfying (3.27) and (3.54) such that kf .Œx; y/ Œf .x/; y Œx; f .y/kA '.x; y/

(3.64)

for all x; y 2 A. If there exists L < 1 such that x '.x; 0/ 2L' ; 0 2 for all x 2 A. Then there exists a unique Lie derivation ı W A ! A satisfying (3.56). Proof. It is easy to show that, there exists a unique involution C-linear mapping ı W A ! A satisfying (3.56). The mapping ı W A ! A is given by f .2n x/ n!1 2n

ı.x/ D lim for all x 2 A. It follows from (3.62) that

kı.Œx; y/ Œı.x/; y Œx; ı.y/kA D lim

n!1

1 kf .4n Œx; y/ Œf .2n x/; 2n y Œ2n x; f .2n y/kA 4n


84

1 '.2n x; 2n y/ 4n 1 lim n '.2n x; 2n y/ n!1 2 D0

lim

n!1

for all x; y 2 A and so ı.Œx; y/ D Œı.x/; y C Œx; ı.y/ for all x; y 2 A. Thus ı W A ! A is a derivation satisfying (3.56). This completes the proof. Corollary 3.44. Let r < 1, be nonnegative real numbers and f W A ! A be a mapping satisfying (3.57) such that kf .Œx; y/ Œf .x/; y Œx; f .y/kA .kxkrA C kykrA /

(3.65)

for all x; y 2 A. Then there exists a unique Lie derivation ı W A ! A satisfying (3.59). Proof. The proof follows from Theorem 3.38 by taking '.x; y/ WD .kxkrA C kykrA / for all x; y 2 A and L D 2r1 .

Remark 3.45. If f W A ! A is a mapping for which there exists a function ' W A2 ! Œ0; 1/ satisfying (3.40), (3.54) and (3.64). Whenever there exists L < 1 such that '.x; 0/

1 L'.2x; 0/ 2

for all x 2 A, then there exists a unique Lie derivation ı W A ! A satisfying (3.60). Corollary 3.46. Let r > 2, be nonnegative real numbers and f W A ! A be a mapping satisfying (3.57) and (3.65). Then there exists a unique Lie derivation ı W A ! A satisfying (3.61). Proof. The proof follows from Theorem 3.45 by taking '.x; y/ WD .kxkrA C kykrA / for all x; y 2 A and L D 21r .

Remark 3.47. For the inequalities controlled by the product of powers of norms, one can obtain similar results to Theorems 3.29 and 3.31 and their corollaries.

3.4 Generalized Additive Mapping

85

3.4 Generalized Additive Mapping Recently, Park and Park [243] introduced and investigated the following generalized additive functional equation n X

0 1 ! ! n n n X X X @ A ri L rj .xi xj / C ri L ri xi

iD1

D

jD1 n X

! ri

iD1

iD1

n X

iD1

ri L.xi /

(3.66)

iD1

for all r1 ; ; rn 2 .0; 1/ whose solution is called a generalized additive mapping. In this section, we consider [210] the following additive functional equation which is somewhat different from (3.66): n 1 X f 2 jD1

X 1in;i¤j

1 X 1 X ri xi rj xj C ri f .xi / D nf ri xi 2 2 iD1 iD1 n

n

(3.67)

for all r1 ; ; rn 2 R: Every solution of the functional equation (3.67) is said to be a generalized additive mapping. Using the fixed point method, we investigate the Hyers–Ulam stability of the functional equation (3.67) in Banach modules over a C -algebra. These results are applied to investigate C -algebra homomorphisms in unital C -algebras. Throughout this section, assume that A is a unital C -algebra with the norm k kA and the unit e that B is a unital C -algebra with the norm k kB and X and Y are left Banach modules over a unital C -algebra A with the norms k kX and k kY ; respectively. Let U.A/ be the group of unitary elements in A and let r1 ; ; rn 2 R:

3.4.1 Hyers–Ulam Stability of Functional Equations in Banach Modules over a C -Algebra For any mapping f W X ! Y; u 2 U.A/ and 2 C; we define Du;r1 ; ;rn f and D ;r1 ; ;rn f W X n ! Y by Du;r1 ; ;rn f .x1 ; ; xn / WD

n 1 X f 2 jD1

X 1in;i¤j

X 1 X 1 ri uxi rj uxj C ri uf .xi / nf ri uxi 2 2 iD1 iD1 n

n


86

and D ;r1 ; ;rn f .x1 ; ; xn / WD

n 1 X f 2 jD1

X 1in;i¤j

1 X X 1 ri xi rj xj C ri f .xi / nf ri xi 2 2 iD1 iD1 n

n

for all x1 ; ; xn 2 X, respectively. Lemma 3.48. Let X and Y be linear spaces and r1 ; ; rn be real numbers with Pn kD1 rk ¤ 0 and ri ¤ 0, rj ¤ 0 for some 1 i < j n: Assume that a mapping L W X ! Y satisfies the functional equation (3.67) for all x1 ; ; xn 2 X: Then the mapping L is additive. Moreover, L.rk x/ D rk L.x/ for all x 2 X and 1 k n: P Proof. Since nkD1 rk ¤ 0, putting x1 D D xn D 0 in (3.67), we get L.0/ D 0. Without loss of generality, we assume that r1 ; r2 ¤ 0. Letting x3 D D xn D 0 in (3.67), we get r x r x r1 x1 C r2 x2 1 1 2 2 CL C r1 L.x1 / C r2 L.x2 / L 2 2 r1 x1 C r2 x2 (3.68) D 2L 2 for all x1 ; x2 2 X. Letting x2 D 0 in (3.68), we get r x r x 1 1 1 1 L r1 L.x1 / D L 2 2 for all x1 2 X. Similarly, by putting x1 D 0 in (3.68), we get r x r x 2 2 2 2 L r2 L.x2 / D L 2 2 for all x2 2 X. It follows from (3.68), (3.69) and (3.70) that r x r x r1 x1 C r2 x2 1 1 2 2 CL L 2 2 r x r x r x r x 1 1 2 2 1 1 2 2 CL CL L L 2 2 2 2 r1 x1 C r2 x2 D 2L 2 for all x1 ; x2 2 X. Replacing x1 and x2 by

2x r1

and

2x r2

(3.69)

(3.70)

(3.71)

in (3.71), we get

L.x C y/ C L.x y/ C L.x/ C L.y/ L.x/ L.y/ D 2L.x C y/

(3.72)


87

for all x; y 2 X. Letting y D x in (3.72), we get that L.2x/ C L.2x/ D 0 for all x 2 X. Then the mapping L is odd. Therefore, it follows from (3.72) that the mapping L is additive. Moreover, let x 2 X and 1 k n. Setting xk D x and xl D 0 for all 1 l n with l ¤ k in (3.67) and, using the oddness of L, it follows that L.rk x/ D rk L.x/. This completes the proof. Using the same method as Pin the proof of Lemma 3.48, we have an alternative result of Lemma 3.48 when nkD1 rk D 0. Lemma 3.49. Let X and Y be linear spaces and r1 ; ; rn be real numbers with ri ¤ 0; rj ¤ 0 for some 1 i < j n: Assume that a mapping L W X ! Y with L.0/ D 0 satisfies the functional equation (3.67) for all x1 ; ; xn 2 X: Then the mapping L is additive. Moreover, L.rk x/ D rk L.x/ for all x 2 X and all 1 k n: Now, we investigate the Hyers-Ulam stability of a generalized additive mapping in Banach modules over a unital C -algebra. Here r1 ; ; rn are real numbers such that ri ¤ 0 and rj ¤ 0 for fixed 1 i < j n: Theorem 3.50. Let f W X ! Y be a mapping satisfying f .0/ D 0 for which there exists a function ' W X n ! Œ0; 1/ such that kDe;r1 ; ;rn f .x1 ; ; xn /kY '.x1 ; ; xn /

(3.73)

for each x1 ; ; xn 2 X. Let 'ij .x; y/ WD ' 0; ; 0; „ƒ‚… x ; 0; ; 0; i th

y ; 0; ; 0 „ƒ‚…

j th

for all x; y 2 X and 1 i < j n: If there exists 0 < C < 1 such that '.2x1 ; ; 2xn / 2C'.x1 ; ; xn / for all x1 ; ; xn 2 X, then there exists a unique generalized additive mapping L W X ! Y such that kf .x/ L.x/kY

x 1 n 2x 2x x C 2'ij ; 'ij ; 4 4C ri rj ri rj 2x x C 'ij ; 0 C 2'ij ; 0 ri ri 2x x o C 2'ij 0; C 'ij 0; rj rj

for all x 2 X: Moreover, L.rk x/ D rk L.x/ for all x 2 X and 1 k n:

(3.74)

88


Proof. For each 1 k n with k ¤ i; j; let xk D 0 in (3.73). Then we get the following inequality: r x C r x r x r x r x C r x i i j j i i j j i i j j Cf 2f f 2 2 2 C ri f .xi / C rj f .xj / Y ' 0; ; 0; xi ; 0; ; 0; xj ; 0; ; 0 „ƒ‚… „ƒ‚… i th

(3.75)

j th

for all xi ; xj 2 X: Letting xi D 0 in (3.75), we get rx r x j j j j f C rj f .xj / 'ij .0; xj / f 2 2 Y

(3.76)

for all xj 2 X: Similarly, letting xj D 0 in (3.75), we get rx r x i i i i f C ri f .xi / 'ij .xi ; 0/ f 2 2 Y

(3.77)

for all xi 2 X: It follows from (3.75), (3.76) and (3.77) that r x C r x r x r x r x C r x i i j j i i j j i i j j Cf 2f f 2 2 2 r x rx r x r x i i j j i i j j Cf f f Cf Y 2 2 2 2 'ij .xi ; xj / C 'ij .xi ; 0/ C 'ij .0; xj / for all xi ; xj 2 X: Replacing xi and xj by

2x ri

and

2y rj

(3.78)

in (3.78), it follows that

kf .x C y/ C f .x y/ 2f .x C y/ C f .x/ C f .y/ f .x/ f .y/kY 2y 2x 2y 2x C 'ij 'ij ; ; 0 C 'ij 0; ri rj ri rj

(3.79)

for all x; y 2 X: Putting y D x in (3.79), we get k2f .x/ 2f .x/ 2f .2x/kY 2x 2x 2x 2x C 'ij 'ij ; ; 0 C 'ij 0; ri rj ri rj

(3.80)


89

for all x 2 X: Replacing x and y by

x 2

and 2x in (3.79), respectively, we get

kf .x/ C f .x/kY x x x x C 'ij ; 0 C 'ij 0; 'ij ; ri rj ri rj

(3.81)

for all x 2 X: It follows from (3.80) and (3.81) that 1 1 .x/ f .2x/ f .x/ Y 2 4

(3.82)

for all x 2 X; where 2x 2x x x C 2'ij ; ; ri rj ri rj 2x x 2x x C 2'ij 0; : ; 0 C 2'ij ; 0 C 'ij 0; C 'ij ri ri rj rj

.x/ WD 'ij

Consider the set W WD fg W X ! Yg and introduce the generalized metric on W defined by d.g; h/ D inffC 2 RC W kg.x/ h.x/kY C .x/; 8x 2 Xg: It is easy to show that .W; d/ is complete. Now, we consider the linear mapping J W W ! W such that Jg.x/ WD

1 g.2x/ 2

(3.83)

for all x 2 X. It follows that d.Jg; Jh/ Cd.g; h/ for all g; h 2 W and so d.f ; Jf / 14 . By Theorem 1.3, there exists a mapping L W X ! Y such that (1) L is a fixed point of J, i.e., L.2x/ D 2L.x/

(3.84)

for all x 2 X. The mapping L is a unique fixed point of J in the set Z D fg 2 W W d.f ; g/ < 1g: This implies that L is a unique mapping satisfying (3.84) such that there exists C 2 .0; 1/ satisfying kL.x/ f .x/kY C .x/ for all x 2 X;


90

(2) d.J n f ; L/ ! 0 as n ! 1. This implies the equality f .2n x/ D L.x/ n!1 2n lim

for all x 2 X; 1 d.f ; Jf /, which implies the inequality d.f ; L/ (3) d.f ; L/ 1C implies that the inequality (3.74) holds.

1 44C :

This

Since '.2x1 ; ; 2xn / 2C'.x1 ; ; xn /, we have 1 kDe;r1 ; ;rn f .2k x1 ; ; 2k xn /kY 2k 1 lim k '.2k x1 ; ; 2k xn / k!1 2

kDe;r1 ; ;rn L.x1 ; ; xn /kY D lim

k!1

lim Ck '.x1 ; ; xn / D 0 k!1

for all x1 ; ; xn 2 X: Therefore, the mapping L W X ! Y satisfies the Eq. (3.67) and L.0/ D 0: Hence, by Lemma 3.49, L is a generalized additive mapping and L.rk x/ D rk L.x/ for all x 2 X and all 1 k n: This completes the proof. Theorem 3.51. Let f W X ! Y be a mapping satisfying f .0/ D 0 for which there exists a function ' W X n ! Œ0; 1/ satisfying kDu;r1 ; ;rn f .x1 ; ; xn /k '.x1 ; ; xn /

(3.85)

for all x1 ; ; xn 2 X and u 2 U.A/: If there exists 0 < C < 1 such that '.2x1 ; ; 2xn / 2C'.x1 ; ; xn / for all x1 ; ; xn 2 X, then there exists a unique A-linear generalized additive mapping L W X ! Y satisfying (3.74) for all x 2 X: Moreover, L.rk x/ D rk L.x/ for all x 2 X and 1 k n: Proof. By Theorem 3.50, there exists a unique generalized additive mapping L W X ! Y satisfying (3.74) and, moreover, L.rk x/ D rk L.x/ for all x 2 X and 1 k n: By the assumption, for all u 2 U.A/, we get x ; 0 ; 0/ Du;r1 ; ;rn L.0; ; 0; „ƒ‚…

Y

i th

1 k f .0; ; 0; 2 x ; 0 ; 0/ D u;r ; ;r 1 n „ƒ‚… k!1 2k Y

D lim

i th


91

1 k ' 0; ; 0; 2 x ; 0 ; 0 „ƒ‚… k!1 2k

lim

i th

x ;0 ;0 lim Ck ' 0; ; 0; „ƒ‚… k!1

i th

D0 for all x 2 X and so ri uL.x/ D L.ri ux/ for all u 2 U.A/ and x 2 X: Since L.ri x/ D ri L.x/ for all x 2 X and ri ¤ 0; we have L.ux/ D uL.x/ for all u 2 U.A/ and x 2 X: Now, we have L.ax C by/ D L.ax/ C L.by/ D aL.x/ C bL.y/ for all a; b 2 A .a; b ¤ 0/ and x; y 2 X: Since L.0x/ D 0 D 0L.x/ for all x 2 X; the unique generalized additive mapping L W X ! Y is an A-linear mapping. This completes the proof. Theorem 3.52. Let f W X ! Y be a mapping satisfying f .0/ D 0 for which there exists a function ' W X n ! Œ0; 1/ such that kDe;r1 ; ;rn f .x1 ; ; xn /kY '.x1 ; ; xn /

(3.86)

for all x1 ; ; xn 2 X. If there exists 0 < C < 1 such that '.x1 ; ; 2n /

C '.2x1 ; ; 2xn / 2

for all x1 ; ; xn 2 X, then there exists a unique generalized additive mapping L W X ! Y such that kf .x/ L.x/kY

x C n 2x 2x x C 2'ij ; 'ij ; 4 4C ri rj ri rj 2x x 2x x o C 2'ij 0; C 'ij ; 0 C 2'ij ; 0 C 'ij 0; ri ri rj rj

(3.87)

for all x 2 X; where 'ij is defined in the statement of Theorem 3.50. Moreover, L.rk x/ D rk L.x/ for all x 2 X and 1 k n:


92

Proof. It follows from (3.82) that x 1 f .x/ f 2 Y 2

x 2

C 4

.x/

for all x 2 X; where is defined in the proof of Theorem 3.50. The rest of the proof is similar to the proof of Theorem 3.50. This completes the proof. Remark 3.53. Let f W X ! Y be a mapping with f .0/ D 0 for which there exists a function ' W X n ! Œ0; 1/ satisfying kDu;r1 ; ;rn f .x1 ; ; xn /k '.x1 ; ; xn /

(3.88)

for all x1 ; ; xn 2 X and u 2 U.A/: If there exists 0 < C < 1 such that '.x1 ; ; 2n /

C '.2x1 ; ; 2xn / 2

for all x1 ; ; xn 2 X, then there exists a unique A-linear generalized additive mapping L W X ! Y satisfying (3.87) for all x 2 X: Moreover, L.rk x/ D rk L.x/ for all x 2 X and 1 k n: P Remark 3.54. In Theorems 3.52 and 3.53, one can assume that nkD1 rk ¤ 0 instead of f .0/ D 0:

3.4.2 Homomorphisms in Unital C -Algebras Now, we investigate C -algebra homomorphisms in unital C -algebras. We use the following lemma in the proof of the following theorem. Lemma 3.55 ([229]). Let f W A ! B be an additive mapping such that f . x/ D f .x/ for all x 2 A and 2 S1 WD f 2 C W jj D 1g: Then the mapping f W A ! B is C-linear. Theorem 3.56. Let f W A ! B be a mapping with f .0/ D 0 for which there exists a function ' W An ! Œ0; 1/ satisfying D ;r ; ;r f .x1 ; ; xn / '.x1 ; ; xn /; 1 n B k f .2 u / f .2k u/ ' 2k u; ; 2k u ; B „ ƒ‚ …

(3.89) (3.90)

n times

k f .2 ux/ f .2k u/f .x/ ' 2k ux; ; 2k ux B ƒ‚ … „ n times

(3.91)


93

for all x; x1 ; ; xn 2 A; u 2 U.A/; k 2 N and 2 S1 . If there exists 0 < C < 1 such that '.2x1 ; ; 2xn / 2C'.x1 ; ; xn / for all x1 ; ; xn 2 A, then the mapping f W A ! B is a C -algebra homomorphism. Proof. Since jJj 3; letting D 1 and xk D 0 for all 1 k n with k ¤ i; j in (3.89), we get f

r x r x ri xi C rj xj i i j j Cf C ri f .xi / C rj f .xj / 2 2 ri xi C rj xj D 2f 2

for all xi ; xj 2 A: By Lemma 3.48, the mapping f is additive and f .rk x/ D rk f .x/ for all x 2 A and k D i; j: So, by letting xi D x and xk D 0 for all 1 k n with k ¤ i in (3.89), it follows that f . x/ D f .x/ for all x 2 A and 2 S1 . Therefore, by Lemma 3.55, the mapping f is C-linear. Hence it follows from (3.90) and (3.91) that 1 f .2k u / f .2k u/ B k 2 1 lim k ' 2k u; ; 2k u „ ƒ‚ … k!1 2

kf .u / f .u/ kB D lim

k!1

n times

lim C ' u; ; u „ ƒ‚ … k!1 k

n times

D0 and 1 f .2k ux/ f .2k u/f .x/ B k k!1 2 1 lim k ' 2k ux; ; 2k ux „ ƒ‚ … k!1 2

kf .ux/ f .u/f .x/kB D lim

n times

lim Ck ' ux; ; ux „ ƒ‚ … k!1 n times

D0


94

for all x 2 A and u 2 U.A/ and so f .u / D f .u/ ;

f .ux/ D f .u/f .x/

for all x 2 A and u 2 U.A/: Since f is C-linear and each P x 2 A is a finite linear combination of unitary elements (see [168]), i.e., x D m kD1 k uk ; where k 2 C and uk 2 U.A/ for all 1 k n; we have m X

f .x / D f

! k uk

kD1

D

m X

m X

D

kD1

!

k f .uk /

m X k f uk D k f .uk /

Df

kD1

kD1

m X

! k u k

D f .x/

kD1

and f .xy/ D f

m X kD1

D

m X

! k u k y D

m X

k f .uk y/

kD1

k f .uk /f .y/ D f

kD1

m X

! k uk f .y/ D f .x/f .y/

kD1

for all x; y 2 A: Therefore, the mapping f W A ! B is a C -algebra homomorphism. This completes the proof. Remark 3.57. Let f W A ! B be a mapping with f .0/ D 0 for which there exists a function ' W An ! Œ0; 1/ satisfying kD ;r1 ; ;rn f .x1 ; ; xn /kB ' x1 ; ; xn ; u u u u f k f k k ; ; k ; B 2 2 2 ƒ‚ 2… „

(3.92)

n times

ux u ux ux f k f k f .x/ k ; ; k B 2 2 „2 ƒ‚ 2…

(3.93)

n times

for all x; x1 ; ; xn 2 A; u 2 U.A/; k 2 N and 2 S1 . If there exists 0 < C < 1 such that '.x1 ; ; 2n /

C '.2x1 ; ; 2xn / 2

for all x1 ; ; xn 2 A, then the mapping f W A ! B is a C -algebra homomorphism.


95

Remark 3.58. In Theorem 3.56 and last remark, one can assume that instead of f .0/ D 0:

Pn

kD1 rk

¤0

Theorem 3.59. Let f W A ! B be a mapping with f .0/ D 0 for which there exists a function ' W An ! Œ0; 1/ satisfying (3.90), (3.91) and kD ;r1 ; ;rn f .x1 ; ; xn /kB '.x1 ; ; xn /; for all x1 ; ; xn 2 A and 2 S1 . Assume that limk!1 there exists 0 < C < 1 such that

1 f .2k e/ 2k

(3.94) is invertible. If

'.2x1 ; ; 2xn / 2C'.x1 ; ; xn / for all x1 ; ; xn 2 A, then the mapping f W A ! B is a C -algebra homomorphism. Proof. Consider the C -algebras A and B as left Banach modules over the unital C -algebra C: By Theorem 3.51, there exists a unique C-linear generalized additive mapping H W A ! B defined by 1 f .2k x/ k!1 2k

H.x/ D lim for all x 2 A: By (3.90) and (3.91), we get

H u H.u/ D lim 1 f 2k u f 2k u B B k k!1 2 1 lim k ' 2k u; ; 2k u „ ƒ‚ … k!1 2 n times

D0 and 1 f 2k ux f .2k u/f .x/ B k 2 1 lim k ' 2k ux; ; 2k ux ƒ‚ … „ k!1 2

kH.ux/ H.u/f .x/kB D lim

k!1

n times

D0 for all u 2 U.A/ and x 2 A and so H u D H.u/ ;

H.ux/ D H.u/f .x/


96

for all u 2 U.A/ and x 2 A: Therefore, by the additivity of H, we have H.ux/ D lim

k!1

1 k 1 H 2 ux D H.u/ lim k f 2k x D H.u/H.x/ k!1 2 2k

(3.95)

for all u 2 U.A/ and x 2 A: Since H is C-linear and each x 2 A is a finite linear P combination of unitary elements, i.e., x D m uk ; where k 2 C and uk 2 U.A/ k kD1 for all 1 k n; it follows from (3.95) that ! m m X X H.xy/ D H k u k y D k H.uk y/ kD1

D

m X

kD1

k H.uk /H.y/ D H

kD1

m X

! k uk H.y/

kD1

D H.x/H.y/ and m X H x D H k uk

! D

kD1

D

m X

k H.uk /

kD1

k H.uk / D

kD1

DH

m X

m X

! k H.uk /

kD1

X m

k u k

D H.x/

kD1

for all x; y 2 A: Since H.e/ D limk!1

1 f .2k e/ 2k

is invertible and

H.e/H.y/ D H.ey/ D H.e/f .y/;

H.y/ D f .y/

for all y 2 A: Therefore, the mapping f W A ! B is a C -algebra homomorphism. This completes the proof. Remark 3.60. Let f W A ! B be a mapping with f .0/ D 0 for which there exists a function ' W An ! Œ0; 1/ satisfying (3.92), (3.93) and kD ;r1 ; ;rn f .x1 ; ; xn /kB '.x1 ; ; xn /; for all x1 ; ; xn 2 A and 2 S1 . Assume that limk!1 2k f . 2ek / is invertible. If there exists 0 < C < 1 such that '.x1 ; ; 2n /

C '.2x1 ; ; 2xn / 2

for all x1 ; ; xn 2 A, then the mapping f W A ! B is a C -algebra homomorphism.


97

In the last Remark, one can assume that

Pn

kD1 rk

¤ 0 instead of f .0/ D 0:

Theorem 3.61. Let f W A ! B be a mapping with f .0/ D 0 for which there exists a function ' W An ! Œ0; 1/ satisfying (3.90), (3.91) and kD ;r1 ; ;rn f .x1 ; ; xn /kB '.x1 ; ; xn /

(3.96)

for D i; 1 and x1 ; ; xn 2 A. Assume that limk!1 21k f .2k e/ is invertible and, for each fixed x 2 A, the mapping t 7! f .tx/ is continuous in t 2 R. If there exists 0 < C < 1 such that '.2x1 ; ; 2xn / 2C'.x1 ; ; xn / for all x1 ; ; xn 2 A, then the mapping f W A ! B is a C -algebra homomorphism. Proof. Put D 1 in (3.96). By the same reasoning as in the proof of Theorem 3.50, there exists a unique generalized additive mapping H W A ! B defined by f .2k x/ k!1 2k

H.x/ D lim

for all x 2 A: It is straight forward to show that, the generalized additive mapping H W A ! B is R-linear. By the same method as in the proof of Theorem 3.51, we have x ; 0 ; 0/ D ;r1 ; ;rn H.0; ; 0; „ƒ‚… Y

j th

1 k f .0; ; 0; 2 x ; 0 ; 0/ D ;r ; ;r 1 n „ƒ‚… k!1 2k Y

D lim

j th

lim

k!1

1 '.0; ; 0; „ƒ‚… 2k x ; 0 ; 0/ 2k j th

D0 for all x 2 A and so rj H.x/ D H.rj x/ for all x 2 A: Since H.rj x/ D rj H.x/ for all x 2 X and rj ¤ 0; we have H. x/ D H.x/


98

for all x 2 A and D i; 1. For each element 2 C, we have D s C it; where s; t 2 R. Thus it follows that H.x/ D H.sx C itx/ D sH.x/ C tH.ix/ D sH.x/ C itH.x/ D .s C it/H.x/ D H.x/ for all 2 C and x 2 A: So, we have H. x C y/ D H. x/ C H. y/ D H.x/ C H.y/ for all ; 2 C and x; y 2 A: Hence the generalized additive mapping H W A ! B is C-linear. The rest of the proof is the same as in the proof of Theorem 3.59. This completes the proof. The following is an alternative result of Theorem 3.61. Remark 3.62. Let f W A ! B be a mapping with f .0/ D 0 for which there exists a function ' W An ! Œ0; 1/ satisfying (3.92), (3.93) and kD ;r1 ; ;rn f .x1 ; ; xn /kB '.x1 ; ; xn / for all x; x1 ; ; xn 2 A and D i; 1: Assume that limk!1 2k f . 2ek / is invertible and, for each fixed x 2 A, the mapping t 7! f .tx/ is continuous in t 2 R. If there exists 0 < C < 1 such that '.x1 ; ; 2n /

C '.2x1 ; ; 2xn / 2

for all x1 ; ; xn 2 A, then the mapping f W A ! B is a C -algebra homomorphism. P Remark 3.63. Also, one can assume that nkD1 rk ¤ 0 instead of f .0/ D 0:

3.5 Generalized Additive Mappings in Banach Modules Let X; Y be vector spaces. It is shown that, if an odd mapping f W X ! Y satisfies the functional equation ! ! Pd Pd .j/ X jD1 xj jD1 .1/ xj rf rf C r r .j/ D 0; 1 Pd jD1 .j/ D l D .d1 Cl d1 Cl1 C 1/

d X jD1

f .xj /;

(3.97)

3.5 Generalized Additive Mappings in Banach Modules

99

then the odd mapping f W X ! Y is additive. Also, we consider the Hyers-Ulam stability of the functional equation (3.97) in Banach modules over a unital C algebra. As an application, we show that every almost linear bijection h W A ! B of a unital C -algebra A onto a unital C -algebra B is a C -algebra isomorphism when n n h. 2rn uy/ D h. 2rn u/h.y/ for all unitaries u 2 A, y 2 A and n 0 [20]. Throughout this section, assume that r is a positive rational number and d, l are integers with 1 < l < d2 .

3.5.1 Odd Functional Equations in d–Variables Now, assume that X and Y are real linear spaces. Lemma 3.64. An odd mapping f W X ! Y satisfies (3.97) for all x1 ; x2 ; ; xd 2 X if and only if f is Cauchy additive. Proof. Assume that f W X ! Y satisfies (3.97) for all x1 ; x2 ; ; xd 2 X. Note that f .0/ D 0 and f .x/ D f .x/ for all x 2 X since f is an odd mapping. Putting x1 D x; x2 D y and x3 D D xd D 0 in (3.97), we get .d2 Cl d2 Cl2 C 1/rf

x C y

r D .d1 Cl d1 Cl1 C 1/.f .x/ C f .y//

(3.98)

for all x; y 2 X. Since d2 Cl d2 Cl2 C 1 Dd1 Cl d1 Cl1 C 1, we have rf

x C y r

D f .x/ C f .y/

for all x; y 2 X. Letting y D 0 in (3.98), we get rf . xr / D f .x/ for all x 2 X. Hence we have x C y D f .x/ C f .y/ f .x C y/ D rf r for all x; y 2 X. Thus f is Cauchy additive. The converse is obviously true. This completes the proof. In the proof of Lemma 3.64, we prove the following. Corollary 3.65. An odd mapping f W X ! Y satisfies rf

x C y r

D f .x/ C f .y/

for all x; y 2 X if and only if f is Cauchy additive.


100

3.5.2 Stability of Odd Functional Equations in Banach Modules over a C -Algebra Assume that A is a unital C -algebra with the norm j j and a unitary group U.A/ and X, Y are left Banach modules over a unital C -algebra A with norms k k and k k, respectively. For any mapping f W X ! Y, we set Du f .x1 ; ; xd / Pd WD rf

jD1

uxj

r

X

C

rf

Pd .1/.j/ uxj jD1

.j/ D 0; 1 Pd .j/ Dl jD1

.d1 Cl d1 Cl1 C 1/

d X

r

uf .xj /

jD1

for all u 2 U.A/ and x1 ; ; xd 2 X. Theorem 3.66. Let r ¤ 2. Let f W X ! Y be an odd mapping for which there exists a function ' W X d ! Œ0; 1/ such that '.x Q 1 ; ; xd / WD

1 j j X r 2 2j ' x ; ; xd < 1 1 2j r j rj jD0

(3.99)

and kDu f .x1 ; ; xd /k '.x1 ; ; xd /

(3.100)

for all u 2 U.A/ and x1 ; ; xd 2 X. Then there exists a unique A-linear generalized additive mapping L W X ! Y such that kf .x/ L.x/k

'Q x; x; 0; ; 0 „ ƒ‚ … 2.d2 Cl d2 Cl2 C 1/ 1

(3.101)

d 2 times

for all x 2 X. Proof. Note that f .0/ D 0 and f .x/ D f .x/ for all x 2 X since f is an odd mapping. Let u D 1 2 U.A/. Putting x1 D x2 D x and x3 D D xd D 0 in (3.100), we have 2 rf x 2f .x/ r

' x; x; 0; ; 0 „ ƒ‚ … d2 Cl d2 Cl2 C 1 1

d 2 times


101

for all x 2 X. Letting t WDd2 Cl d2 Cl2 C 1, we get r 2 1 f .x/ f x ' x; x; 0; ; 0 „ ƒ‚ … 2 r 2t d 2 times

for all x 2 X. Hence we have rn 2n rnC1 2nC1 n f n x nC1 f nC1 x 2 r 2 r n r 2 2n rn 2 x D n f n x f 2 r 2 r rn r n 2n 2n nC1 ' n x; n x; 0; ; 0 2 t r r „ ƒ‚ …

(3.102)

d 2 times

for all x 2 X and n 1. By (3.102), we have rm 2m rn 2n mf mx nf nx 2 r 2 r n1 2k 2k X rk ' x; x; 0; ; 0 2kC1 t rk rk „ ƒ‚ … kDm

(3.103)

d 2 times

for all x 2 X and m; n 1 with m < n. This shows that the sequence f 2r n f . 2rn x/g n n is a Cauchy sequence for all x 2 X. Since Y is complete, the sequence f 2r n f . 2rn x/g converges for all x 2 X. So we can define a mapping L W X ! Y by n

n

r n 2n f nx n!1 2n r

L.x/ WD lim

for all x 2 X. Since f .x/ D f .x/ for all x 2 X, we have L.x/ D L.x/ for all x 2 X. Also, we get 2n rn 2n D1 f n x1 ; ; n xd n n!1 2 r r r n 2n 2n lim n ' n x1 ; ; n xd n!1 2 r r D0

kD1 L.x1 ; ; xd /k D lim

for all x1 ; ; xd 2 X. So L is a generalized additive mapping. Putting m D 0 and letting n ! 1 in (3.103), we get (3.101). Now, let L0 W X ! Y be another additive mapping satisfying (3.101). By Lemma 3.64, L and L0 are additive. So we have


102

n n rn 2 0 2 x L x L 2n rn rn n 2n 2n 2n rn 2 n L n x f n x C L0 n x f n x / 2 r r r r 2rn 2n 2n nC1 'Q n x; n x; 0; ; 0 ; 2 t r r „ ƒ‚ …

kL.x/ L0 .x/k D

d 2 times

which tends to zero as n ! 1 for all x 2 X. So we can conclude that L.x/ D L0 .x/ for all x 2 X. This proves the uniqueness of L. By the assumption, for each u 2 U.A/, we get 2n rn kDu L.x; 0; ; 0/k D lim n Du f n x; 0; ; 0 „ ƒ‚ … „ ƒ‚ … n!1 2 r d 1 times

d 1 times

r ' x; 0; ; 0 n!1 2n rn „ ƒ‚ …

lim

n

2n

d 1 times

D0 for all x 2 X and so

ux .d1 Cl d1 Cl1 C 1/rL D .d1 Cl d1 Cl1 C 1/uL.x/ r

for all u 2 U.A/ and x 2 X. Since L is additive, ux L.ux/ D rL D uL.x/ r

(3.104)

for all u 2 U.A/ and x 2 X. Now, let a 2 A .a ¤ 0/ and M be an integer greater than 4jaj. Then we have ˇaˇ 1 2 1 ˇ ˇ ˇ ˇ< 2 and , p > 1 be positive real numbers or let r < 2 and , p < 1 be positive real numbers. Let f W X ! Y be an odd mapping such that kDu f .x1 ; ; xd /k

d X

jjxj jjp

jD1


rp1 kxkp .rp1 2p1 /.d2 Cl d2 Cl2 C 1/

for all x 2 X. Proof. Defining '.x1 ; ; xd / D the desired result.

Pd jD1

jjxj jjp and applying Theorem 3.66, we get

Theorem 3.68. Let r ¤ 2. Let f W X ! Y be an odd mapping for which there exists a function ' W X d ! Œ0; 1/ satisfying (3.100) such that '.x Q 1 ; ; xd / WD

1 X 2j r j rj ' x ; ; xd < 1 1 r j 2j 2j jD1


'Q x; x; 0; ; 0 „ ƒ‚ … 2.d2 Cl d2 Cl2 C 1/ 1

d 2 times

for all x 2 X. Proof. Note that f .0/ D 0 and f .x/ D f .x/ for all x 2 X since f is an odd mapping. Let u D 1 2 U.A/. Putting x1 D x2 D x and x3 D D xd D 0 in (3.100), we have 2 rf x 2f .x/ r

' x; x; 0; ; 0 „ ƒ‚ … d2 Cl d2 Cl2 C 1 1

d 2 times


104

for all x 2 X. Letting t WDd2 Cl d2 Cl2 C 1, we get 2 r 1 r r f .x/ f x ' x; x; 0; ; 0 r 2 rt 2 2 „ ƒ‚ … d 2 times

for all x 2 X. The rest of the proof is similar to the proof of Theorem 3.66. This completes the proof. Corollary 3.69. Let r < 2 and , p > 1 be positive real numbers or let r > 2 and , p < 1 be positive real numbers. Let f W X ! Y be an odd mapping such that kDu f .x1 ; ; xd /k

d X

kxj kp

jD1


rp1 kxkp .2p1 rp1 /.d2 Cl d2 Cl2 C 1/

for all x 2 X. Proof. Defining '.x1 ; ; xd / D

d X

jjxj jjp

jD1

and applying Theorem 3.68, we get the desired result.

Now, we investigate the Hyers–Ulam stability of linear mappings for the case d D 2. Theorem 3.70. Let r ¤ 2. Let f W X ! Y be an odd mapping for which there exists a function ' W X 2 ! Œ0; 1/ such that '.x; Q y/ WD

1 j j X r 2 2j ' x; y < 1 2j r j r j jD0

and ux C uy uf .x/ uf .y/ '.x; y/ rf r

(3.105)


105

for all u 2 U.A/ and x; y 2 X. Then there exists a unique A-linear mapping L W X ! Y such that kf .x/ L.x/k

1 '.x; Q x/ 2

for all x 2 X. Proof. Let u D 1 2 U.A/. Putting x D y in (3.105), we have 2 rf x 2f .x/ '.x; x/ r for all x 2 X and so r 2 1 f .x/ f x '.x; x/ 2 r 2 for all x 2 X. The rest of the proof is the same as in the proof of Theorem 3.66. This completes the proof. Corollary 3.71. Let r > 2 and , p > 1 be positive real numbers or let r < 2 and , p < 1 be positive real numbers. Let f W X ! Y be an odd mapping such that ux C uy uf .x/ uf .y/ .kxkp C kykp / rf r for all u 2 U.A/ and x; y 2 X. Then there exists a unique A-linear mapping L W X ! Y such that kf .x/ L.x/k

rp1 jjxjjp 2p1

rp1

for all x 2 X. Proof. Defining '.x; y/ D .jjxjjp C jjyjjp / and applying Theorem 3.70, we get the desired results. Theorem 3.72. Let r ¤ 2. Let f W X ! Y be an odd mapping for which there exists a function ' W X 2 ! Œ0; 1/ satisfying (3.105) such that '.x; Q y/ WD

1 X 2j r j r j ' x; y < 1 r j 2j 2j jD1

for all u 2 U.A/ and all x; y 2 X. Then there exists a unique A-linear mapping L W X ! Y such that


106

kf .x/ L.x/k

1 '.x; Q x/ 2

for all x 2 X. Proof. Let u D 1 2 U.A/. Putting x D y in (3.105), we have 2 rf x 2f .x/ '.x; x/ r for all x 2 X and so 2 r 1 r r f .x/ f x ' x; x r 2 r 2 2 for all x 2 X. The rest of the proof is similar to the proof of Theorem 3.66. This completes the proof. Corollary 3.73. Let r < 2 and , p > 1 be positive real numbers or let r > 2 and , p < 1 be positive real numbers. Let f W X ! Y be an odd mapping such that ux C uy uf .x/ uf .y/ .kxkp C kykp / rf r for all u 2 U.A/ and x; y 2 X. Then there exists a unique A-linear mapping L W X ! Y such that kf .x/ L.x/k

rp1 kxkp rp1

2p1

for all x 2 X. Proof. Defining '.x; y/ D .kxkp C kykp / and applying Theorem 3.72, we get the desired results.

3.5.3 Isomorphisms in Unital C -Algebras Assume that A is a unital C -algebra with the norm jj jj and the unit e and B is a unital C -algebra with the norm k k. Let U.A/ be the set of unitary elements in A. Now, we investigate C -algebra isomorphisms in unital C -algebras. Theorem 3.74. Let r ¤ 2. Let h W A ! B be an odd bijective mapping satisfying n n h. 2rn uy/ D h. 2rn u/h.y/ for all u 2 U.A/, y 2 A and n 0 for which there exists a function ' W Ad ! Œ0; 1/ such that


107

1 X 2j r j 2j ' j x1 ; ; j xd < 1; j 2 r r jD0

(3.106)

kD h.x1 ; ; xd /k '.x1 ; ; xd /; 2n 2n 2n 2n h. n u / h. n u/ ' n u; ; n u r r ƒ‚ r … „r

(3.107)

d times

for all 2 S1 WD f 2 C W jj D 1g, u 2 U.A/, n 0 and x1 ; ; xd 2 A. Assume n n that limn!1 2r n h. 2rn e/ is invertible. Then the odd bijective mapping h W A ! B is a C -algebra isomorphism. Proof. Consider the C -algebras A and B as left Banach modules over the unital C -algebra C. By Theorem 3.66, there exists a unique C-linear generalized additive mapping H W A ! B such that kh.x/ H.x/k

'Q x; x; 0; ; 0 „ ƒ‚ … 2.d2 Cl d2 Cl2 C 1/ 1

(3.108)

d 2 times

for all x 2 A. The generalized additive mapping H W A ! B is given by r n 2n h nx n!1 2n r

H.x/ D lim

for all x 2 A. By (3.106) and (3.107), we get r n 2n rn 2n h n u D lim n h n u n n!1 2 n!1 2 r r n n r 2 D lim n h n u D H.u/ n!1 2 r

H.u / D lim

for all u 2 U.A/. Since H is C-linear and P each x 2 A is a finite linear combination of unitary elements (see [168]), i.e., x D m jD1 j uj for all j 2 C and uj 2 U.A/, we have

H.x / D H

m X

j uj

D

jD1

D

m X

m X

j H.uj /

D

jD1

j H.uj /

jD1

DH

m X

j H.uj /

jD1 m X

j u j

D H.x/

jD1

for all x 2 A. Since h. 2rn uy/ D h. 2rn u/h.y/ for all u 2 U.A/, y 2 A and n 0, we have n

n


108

r n 2n r n 2n h uy D lim h n u h.y/ n!1 2n n!1 2n rn r D H.u/h.y/

H.uy/ D lim

(3.109)

for all u 2 U.A/ and y 2 A. By the additivity of H and (3.109), we have 2n 2n 2n 2n H.uy/ D H uy D H u y D H.u/h y rn rn rn rn for all u 2 U.A/ and y 2 A. Hence we have 2n r n 2n rn H.u/h y D H.u/ h y 2n rn 2n r n

H.uy/ D

(3.110)

for all u 2 U.A/ and y 2 A. Taking n ! 1 in (3.110), we obtain H.uy/ D H.u/H.y/

(3.111)

for all u 2 U.A/ and y 2 A. Since H is C-linear and each x 2 A is a finite linear P combination of unitary elements, i.e., x D m jD1 j uj for all j 2 C and uj 2 U.A/, it follows from (3.111) that H.xy/ D H

m X

m m X X j u j y D j H.uj y/ D j H.uj /H.y/

jD1

D H.

m X

jD1

jD1

j uj /H.y/ D H.x/H.y/

jD1

for all x; y 2 A. By (3.109) and (3.111), we have H.e/H.y/ D H.ey/ D H.e/h.y/ for all y 2 A. Since limn!1

n rn h. 2rn e/ 2n

D H.e/ is invertible,

H.y/ D h.y/ for all y 2 A. Therefore, the odd bijective mapping h W A ! B is a C -algebra isomorphism. This completes the proof. Corollary 3.75. Let r > 2 and , p > 1 be positive real numbers or let r < 2 and , p < 1 be positive real numbers. Let h W A ! B be an odd bijective mapping n n satisfying h. 2rn uy/ D h. 2rn u/h.y/ for all u 2 U.A/, y 2 A and n 0 such that


kD h.x1 ; ; xd /k

109

d X

jjxj jjp

jD1

and 2n 2n 2pn h n u h n u d pn r r r for all 2 S1 , u 2 U.A/, n 0 and x1 ; ; xd 2 A. Assume that limn!1 2r n h. 2rn e/ is invertible. Then the odd bijective mapping h W A ! B is a C -algebra isomorphism. n

n

Proof. Defining '.x1 ; ; xd / D

d X

jjxj jjp

jD1

and applying Theorem 3.74, we get the desired results.

Theorem 3.76. Let r ¤ 2. Let h W A ! B be an odd bijective mapping satisfying n n h. 2rn uy/ D h. 2rn u/h.y/ for all u 2 U.A/, y 2 A and n 0 for which there exists a n n function ' W Ad ! Œ0; 1/ satisfying (3.106), (3.107). Assume that limn!1 2r n h. 2rn e/ is invertible such that kD h.x1 ; ; xd /k '.x1 ; ; xd /

(3.112)

for all x1 ; ; xd 2 A and D 1; i. If h.tx/ is continuous in t 2 R for each fixed x 2 A, then the odd bijective mapping h W A ! B is a C -algebra isomorphism. Proof. Put D 1 in (3.112). By the same reasoning as in the proof of Theorem 3.74, there exists a unique generalized additive mapping H W A ! B satisfying (3.108). By (3.112), the generalized additive mapping H W A ! B is R-linear. Put D i in (3.112). By the same method as in the proof of Theorem 3.66, one can obtain that r n 2n i r n 2n h ix D lim h n x D iH.x/ n!1 2n n!1 2n rn r

H.ix/ D lim

for all x 2 A. For each element 2 C, let D s C it, where s; t 2 R. Then we have H.x/ D H.sx C itx/ D sH.x/ C tH.ix/ D sH.x/ C itH.x/ D .s C it/H.x/ D H.x/ for all 2 C and x 2 A. Thus we have

110


H. x C y/ D H. x/ C H. y/ D H.x/ C H.y/ for all ; 2 C and x; y 2 A. Hence the generalized additive mapping H W A ! B is C-linear. The rest of the proof is the same as in the proof of Theorem 3.66. This completes the proof.

3.6 Jordan -Derivations and Quadratic Jordan -Derivations Jordan -derivations were introduced in [307, 308] for the first time and the structure of such derivations has been investigated in [52]. The reason for introducing these mappings was the fact that the problem of representing quadratic forms by sesquilinear ones is closely connected with the structure of Jordan -derivations. In [13], An et al. investigated Jordan -derivations on C -algebras and Jordan -derivations on JC -algebras associated with a special functional inequality. In this section, we consider the Hyers-Ulam stability of Jordan -derivations and quadratic Jordan -derivations on real C -algebras and real JC -algebras. We also prove the superstability of Jordan -derivations and quadratic Jordan -derivations on real C -algebras and real JC -algebras under some conditions [50].

3.6.1 Stability of Jordan -Derivations Here we prove the Hyers–Ulam stability of Jordan -derivations on real C -algebras and real JC -algebras. Definition 3.77. Let A be a real C -algebra. An R-linear mapping D W A ! A is called a Jordan -derivation if D.a2 / D a D.a/ C D.a/a for all a 2 A. The mapping Dx W A ! A; a 7! a x xa , where x is a fixed element in A, is a Jordan -derivation. A real C -algebra A endowed with the Jordan product a ı b WD abCba on A is called a real JC -algebra (see [13, 225]). 2 Definition 3.78. Let A be a real JC -algebra. An R-linear mapping ı W A ! A is called a Jordan -derivation if ı.a2 / D a ı D.a/ C D.a/ ı a for all a 2 A.

3.6 Jordan -Derivations and Quadratic Jordan -Derivations

111

Theorem 3.79. Let A be a real C -algebra. Suppose that f W A ! A is a mapping with f .0/ D 0 for which there exists a function ' W A3 ! Œ0; 1/ such that '.a; Q b; c/ WD

1 X 1 '.2n a2n b; 2n c/ < 1 nC1 2 nD0

(3.113)

and kf .a C b C c2 / f .a/ f .b/ f .c/c c f .c/k '.a; b; c/

(3.114)

for all 2 R and a; b; c 2 A. Then there exists a unique Jordan -derivation ı on A satisfying kf .a/ ı.a/k '.a; Q a; 0/

(3.115)

for all a 2 A. Proof. Setting a D b; c D 0 and D 1 in (3.114), we have kf .2a/ 2f .a/k '.a; a; 0/ for all a 2 A. One can use induction to show that n1 n f .2 a/ f .2m a/ X 1 '.2k a; 2k a; 0/ 2n 2m D kC1 2 kDm

(3.116)

for all n > nm 0 and a 2 A. It follows from (3.113) and (3.116) that the sequence f f .22na/ g is a Cauchy sequence. Due to the completeness of A, this sequence is convergent. Define f .2n a/ n!1 2n

d.a/ WD lim

(3.117)

for all a 2 A. Then we have ı

1 1 f .2nk a/ 1 a D lim k D k d.a/ k nk n!1 2 2 2 2

for each k 2 N. Putting c D 0 and replacing a and b by 2n a and 2n b, respectively, in (3.114), we get 1 1 1 1 n f .2n .a C b// n f .2n a/ n f .2n b/ n '.2n a; 2n b; 0/: 2 2 2 2


112

Taking n ! 1, we obtain ı.a C b/ D ı.a/ C ı.b/ for all a; b 2 A and 2 R. So ı is R-linear. Putting a D b D 0 and substituting c by 2n c in (3.114), we get 1 1 1 2n f .22n c2 / 2n f .2n c/.2n c / 2n .2n c /f .2n c/ 2 2 2 1 2n '.0; 0; 2n c/ 2 1 n '.0; 0; 2n c/: 2 Taking n ! 1, we obtain ı.c2 / D ı.c/c C c ı.c/ for all c 2 A. Moreover, it follows from (3.116) with m D 0 and (3.117) that kı.a/ f .a/k '.a; Q a; 0/ for all a 2 A. For the uniqueness of ı, let ıQ W A ! B be another Jordan -derivation satisfying (3.115). Then we have 1 Q Q n a/k kı.a/ ı.a/k D n kı.2n a/ ı.2 2 1 Q n a/k/ n .kı.2n a/ f .2n a/k C kf .2n a/ ı.2 2 1 X 1 '.2nCj a; 2nCj a; 0/ 2 nCj 2 jD1 D2

1 X 1 '.2j a; 2j a; 0/; j 4 jDn

which tends to zero as n ! 1 for all a 2 A. So ı is unique. Therefore, ı is a Jordan -derivation on A. This completes the proof. Remark 3.80. Let A be a real C -algebra. Suppose that f W A ! A is a mapping with f .0/ D 0 for which there exists a function ' W A3 ! Œ0; 1/ satisfying (3.114) and

3.6 Jordan -Derivations and Quadratic Jordan -Derivations

'.a; Q b; c/ WD

1 X

2n1 '

nD1

113

a b c be a mapping such that

1 , , 3 1 2

C

be non-negative real numbers and f W A ! BC

1 1 1 1 1 1 f x C y C 3x 3 y 3 x 3 C 3y 3 x 3 y 3 f .x/ f .y/ p

p

1 .kxkp C kykp / C 2 kxk 2 kyk 2

(3.168)

for all x; y 2 AC . Then there exists a unique 3rd root mapping T W AC ! BC satisfying (3.155) and kf .x/ T.x/k

21 C 2 kxkp 8p 2

for all x 2 AC . Proof. Define p

p

'.x; y/ D 1 .kxkp C kykp / C 2 kxk 2 kyk 2 and apply Theorem 3.112. Then we get the desired result. C

C

Theorem 3.114. Let f W A ! B be a mapping for which there exists a function ' W AC AC ! Œ0; 1/ satisfying (3.166) such that '.x; Q y/ WD

1 X

2j '.8j x; 8j y/ < 1

jD0


1 '.x; Q x/ 2

for all x 2 AC . Proof. It follows from (3.167) that 1 1 f .x/ f .8x/ '.x; x/ 2 2 for all x 2 AC . The rest of the proof is similar to the proof of Theorem 3.107.

3.9 Square Root and 3rd Root Functional Equations: The Fixed Point Method

137

Corollary 3.115. Let 0 < p < 13 , 1 , 2 be non-negative real numbers and f W AC ! BC be a mapping satisfying (3.168). Then there exists a unique 3rd root mapping T W AC ! BC satisfying (3.155) and kf .x/ T.x/k

21 C 2 kxkp 2 8p

for all x 2 AC . p

p

Proof. Define '.x; y/ D 1 .kxkp Ckykp /C2 kxk 2 kyk 2 and apply Theorem 3.114. Then we get the desired result.

3.9 Square Root and 3rd Root Functional Equations: The Fixed Point Method In this section, we prove the Hyers-Ulam stability of the square root functional equation and the 3rd root functional equation in C -algebras via fixed point method [235].

3.9.1 Stability of the Square Root Functional Equation In this section, we investigate the square root functional equation in C -algebras. Now, we prove the Hyers-Ulam stability of the square root functional equation in C -algebras. Theorem 3.116. Let ' W AC AC ! Œ0; 1/ be a function such that there exists L < 1 with '.x; y/

L '.4x; 4y/ 2

(3.169)

for all x; y 2 AC . Let f W AC ! BC be a mapping satisfying 1 1 1 1 1 1 f x C y C x 4 y 2 x 4 C y 4 x 2 y 4 f .x/ f .y/ '.x; y/

(3.170)

for all x; y 2 AC . Then there exists a unique square root mapping S W AC ! AC satisfying (3.154) and kf .x/ S.x/k for all x 2 AC .

L '.x; x/ 2 2L

(3.171)


138

Proof. Letting y D x in (3.170), we get kf .4x/ 2f .x/k '.x; x/

(3.172)

for all x 2 AC . Consider the set X WD fg W AC ! BC g and introduce the generalized metric on X defined by d.g; h/ D inff 2 RC W kg.x/ h.x/k '.x; x/; 8x 2 AC g; where, as usual, inf D C1 which .X; d/ is complete. Now, we consider the linear mapping J W X ! X such that Jg.x/ WD 2g

x 4

for all x 2 AC . Let g; h 2 X be given such that d.g; h/ D ". Then kg.x/ h.x/k '.x; x/ for all x 2 AC . Hence we have x x 2h kJg.x/ Jh.x/k D 2g L'.x; x/ 4 4 for all x 2 AC . So, d.g; h/ D " implies that d.Jg; Jh/ L". This means that d.Jg; Jh/ Ld.g; h/ for all g; h 2 X. It follows from (3.172) that x L '.x; x/ f .x/ 2f 4 2 for all x 2 AC and so d.f ; Jf / L2 . By Theorem 1.3, there exists a mapping S W AC ! BC satisfying the following: (1) S is a fixed point of J, i.e., S

x 4

D

1 S.x/ 2

for all x 2 AC . The mapping S is a unique fixed point of J in the set M D fg 2 X W d.f ; g/ < 1g:

(3.173)


139

This implies that S is a unique mapping satisfying (3.173) such that there exists a 2 .0; 1/ satisfying kf .x/ S.x/k '.x; x/ for all x 2 AC ; (2) d.J n f ; S/ ! 0 as n ! 1. This implies the equality lim 2n f

n!1

x D S.x/ 4n

for all x 2 AC ; 1 (3) d.f ; S/ 1L d.f ; Jf /, which implies the inequality d.f ; S/

L : 2 2L

This implies that the inequality (3.171) holds. By (3.169) and (3.170), we have x C y C x 14 y 12 x 14 C y 14 x 12 y 41 x y f n f n 2n f n 4 4 4 x y 2n ' n ; n 4 4 n L '.x; y/ for all x; y 2 AC and n 2 N. So, we have 1 1 1 1 1 1 S x C y C x 4 y 2 x 4 C y 4 x 2 y 4 S.x/ S.y/ D 0 for all x; y 2 AC . Thus the mapping S W AC ! BC is a square root mapping. This completes the proof. Corollary 3.117. Let p > 12 , 1 , 2 be non-negative real numbers and f W AC ! BC be a mapping such that 1 1 1 1 1 1 f x C y C x 4 y 2 x 4 C y 4 x 2 y 4 f .x/ f .y/ p

p


(3.174)

for all x; y 2 AC . Then there exists a unique square root mapping S W AC ! BC satisfying (3.154) and

140


21 C 2 jjxjjp 4p 2

kf .x/ S.x/k for all x 2 AC .

Proof. The proof follows from Theorem 3.116 by taking p p '.x; y/ D 1 .kxkp C kykp / C 2 kxk 2 kyk 2 for all x; y 2 AC and L D 212p .

Theorem 3.118. Let ' W AC AC ! Œ0; 1/ be a function such that there exists L < 1 with x y ; '.x; y/ 2L' 4 4 for all x; y 2 AC . Let f W AC ! BC be a mapping satisfying (3.170). Then there exists a unique square root mapping S W AC ! AC satisfying (3.154) and kf .x/ S.x/k

1 '.x; x/ 2 2L

for all x 2 AC . Proof. Let .X; d/ be the generalized metric space defined in the proof of Theorem 3.116. Consider the linear mapping J W X ! X such that Jg.x/ WD

1 gt.4x/ 2

for all x 2 AC . It follows from (3.172) that 1 1 f .x/ f .4x/ '.x; x/ 2 2 for all x 2 AC . So d.f ; Jf / 12 . The rest of the proof is similar to the proof of Theorem 3.116. This completes the proof. Corollary 3.119. Let 0 < p < 12 , 1 , 2 be non-negative real numbers and f W AC ! BC be a mapping satisfying (3.174). Then there exists a unique square root mapping S W AC ! BC satisfying (3.154) and kf .x/ S.x/k for all x 2 AC .

21 C 2 kxkp 2 4p


141

Proof. The proof follows from Theorem 3.118 by taking p

p

'.x; y/ D 1 .kxkp C kykp / C 2 kxk 2 kyk 2 for all x; y 2 AC and L D 22p1 .

3.9.2 Stability of the 3rd Root Functional Equation Now, we investigate the 3rd root functional equation in C -algebras. Now, we prove the Hyers-Ulam stability of the 3rd root functional equation in C -algebras. Theorem 3.120. Let ' W AC AC ! Œ0; 1/ be a function such that there exists an L < 1 with '.x; y/

L '.8x; 8y/ 2

for all x; y 2 AC . Let f W AC ! BC be a mapping satisfying 1 1 1 1 1 1 f x C y C 3x 3 y 3 x 3 C 3y 3 x 3 y 3 f .x/ f .y/ '.x; y/

(3.175)

for all x; y 2 AC . Then there exists a unique 3rd root mapping T W AC ! AC satisfying (3.155) and kf .x/ T.x/k

L '.x; x/ 2 2L

for all x 2 AC . Proof. Letting y D x in (3.175), we get kf .8x/ 2f .x/k '.x; x/

(3.176)

for all x 2 AC . Let .X; d/ be the generalized metric space defined in the proof of Theorem 3.116. Consider the linear mapping J W X ! X such that Jg.x/ WD 2g for all x 2 AC .

x 8


142

Now, we consider the linear mapping J W X ! X such that Jg.x/ WD 2g

x 8

for all x 2 AC . It follows from (3.176) that x L '.x; x/ f .x/ 2f 8 2 for all x 2 X and so d.f ; Jf / L2 . The rest of the proof is similar to the proof of Theorem 3.116. This completes the proof. Corollary 3.121. Let p > 13 , 1 , 2 be non-negative real numbers and f W AC ! BC be a mapping such that 1 1 1 1 1 1 f x C y C 3x 3 y 3 x 3 C 3y 3 x 3 y 3 f .x/ f .y/ p

p


(3.177)


21 C 2 kxkp 8p 2

for all x 2 AC . Proof. The proof follows from Theorem 3.120 by taking p

p

'.x; y/ D 1 .kxkp C kykp / C 2 kxk 2 kyk 2 for all x; y 2 AC and L D 213p .

Theorem 3.122. Let ' W AC AC ! Œ0; 1/ be a function such that there exists L < 1 with x y ; '.x; y/ 2L' 8 8 for all x; y 2 AC . Let f W AC ! BC be a mapping satisfying (3.175). Then there exists a unique 3rd root mapping T W AC ! AC satisfying (3.155) and kf .x/ T.x/k for all x 2 AC .

1 '.x; x/ 2 2L

3.10 Positive-Additive Functional Equation

143

Proof. Let .X; d/ be the generalized metric space defined in the proof of Theorem 3.116. Consider the linear mapping J W X ! X such that Jg.x/ WD

1 g.8x/ 2

for all x 2 AC . It follows from (3.176) that 1 1 f .x/ f .8x/ '.x; x/ 2 2 for all x 2 AC and so d.f ; Jf / 12 . The rest of the proof is similar to the proof of Theorem 3.116. This completes the proof. Corollary 3.123. Let 0 < p < 13 , 1 , 2 be non-negative real numbers and f W AC ! BC be a mapping satisfying (3.177). Then there exists a unique 3rd root mapping T W AC ! BC satisfying (3.155) and kf .x/ T.x/k

21 C 2 kxkp 2 8p


p

'.x; y/ D 1 .kxkp C kykp / C 2 kxk 2 kyk 2 for all x; y 2 AC and L D 23p1 .

3.10 Positive-Additive Functional Equation In this section, we consider a positive-additive functional equation in C -algebras [258]. Using fixed point and direct methods, we prove the stability of the positiveadditive functional equation in C -algebras. Definition 3.124 ([96]). Let A be a C -algebra and x 2 A be a self-adjoint element, i.e., x D x. Then x is said to be positive if it is of the form yy for some y 2 A. The set of positive elements of A is denoted by AC . Note that AC is a closed convex cone (see [96]). It is well known that for a positive element x and a positive integer n there exists a unique positive element 1 y 2 AC such that x D yn . We denote y by x n (see [128]).


144

In this section, we introduce the following functional equation: T

1 1 m 1 1 m D T.x/ m C T.y/ m xm C ym

(3.178)

for all x; y 2 AC and a fixed integer m greater than 1, which is called a positiveadditive functional equation. Each solution of the positive-additive functional equation is called a positive-additive mapping. Note that the function f .x/ D cx for any c 0 in the set of non-negative real numbers is a solution of the functional equation (3.178). Throughout this section, let AC and BC be the sets of positive elements in C -algebras A and B, respectively. Assume that m is a fixed integer greater than 1.

3.10.1 Stability of the Positive-Additive Functional Equations: The Fixed Point Method Here we investigate the positive-additive functional equation (3.178) in C -algebras. Lemma 3.125. Let T W AC ! BC be a positive-additive mapping satisfying (3.178). Then T satisfies T.2mn x/ D 2mn T.x/ for all x 2 AC and n 2 Z. Proof. Putting x D y in (3.178), we obtain T.2m x/ D 2m T.x/ for all x 2 AC . So, one can show that T.2mn x/ D 2mn T.x/ for all x 2 AC and n 2 Z.

Using the fixed point method, we prove the Hyers-Ulam stability of the positiveadditive functional equation (3.178) in C -algebras. Note that the fundamental ideas in the proofs of the main results in this section are contained in [62–64]. Theorem 3.126. Let ' W AC AC ! Œ0; 1/ be a function such that there exists L < 1 with '.x; y/

L ' .2m x; 2m y/ 2m

for all x; y 2 AC . Let f W AC ! BC be a mapping satisfying

(3.179)


145

1 1 m 1 1 m f .x/ m C f .y/ m '.x; y/ f x m C y m

(3.180)

for all x; y 2 AC . Then there exists a unique positive-additive mapping T W AC ! AC satisfying (3.178) and kf .x/ T.x/k

2m

L '.x; x/ 2m L

(3.181)

for all x 2 AC . Proof. Letting y D x in (3.180), we get kf .2m x/ 2m f .x/k '.x; x/ for all x 2 AC . Consider the set X WD fg W AC ! BC g and introduce the generalized metric on X defined by d.g; h/ D inff 2 RC W kg.x/ h.x/k '.x; x/; 8x 2 AC g; where, as usual, inf D C1 which .X; d/ is complete. Now, we consider the linear mapping J W X ! X such that Jg.x/ WD 2m g

x 2m

for all x 2 AC . Let g; h 2 X be given such that d.g; h/ D ". Then we have kg.x/ h.x/k '.x; x/ for all x 2 AC and so x x kJg.x/ Jh.x/k D 2m g m 2m h m L'.x; x/ 2 2 for all x 2 AC . So, d.g; h/ D " implies that d.Jg; Jh/ L". This means that d.Jg; Jh/ Ld.g; h/ for all g; h 2 X. It follows from (3.182) that x L f .x/ 2m f m k m '.x; x/ 2 2

(3.182)


146

for all x 2 AC and so d.f ; Jf / 2Lm . By Theorem 1.3, there exists a mapping T W AC ! BC satisfying the following: (1) T is a fixed point of J, i.e., T

x 1 D m T.x/ m 2 2

(3.183)

for all x 2 AC . The mapping T is a unique fixed point of J in the set M D fg 2 X W d.f ; g/ < 1g: This implies that T is a unique mapping satisfying (3.183) such that there exists a 2 .0; 1/ satisfying kf .x/ T.x/k '.x; x/ for all x 2 AC ; (2) d.J n f ; T/ ! 0 as n ! 1. This implies the equality lim 2mn f

n!1

x D T.x/ 2mn

for all x 2 AC ; 1 (3) d.f ; T/ 1L d.f ; Jf /, which implies the inequality d.f ; T/

L : 2m 2m L

This implies that the inequality (3.181) holds. By (3.179) and (3.180), we have x m1 C y m1 m x m1 y m1 m mn mn 2 f 2 f C 2 f 2mn 2mn 2mn x y 2mn ' mn ; mn 2 2 Lmn '.x; y/ mn

for all x; y 2 AC and n 2 N and so 1 1 m 1 1 m T.x/ m C T.y/ m D 0 T x m C y m for all x; y 2 AC . Thus the mapping T W AC ! BC is positive-additive. This completes the proof.


147

Corollary 3.127. Let p > 1, 1 , 2 be non-negative real numbers and f W AC ! BC be a mapping such that 1 1 m 1 1 m f .x/ m C f .y/ m f x m C y m p

p


(3.184)

for all x; y 2 AC . Then there exists a unique positive-additive mapping T W AC ! BC satisfying (3.178) and kf .x/ T.x/k

21 C 2 kxkp 2mp 2m


p

'.x; y/ D 1 .kxkp C kykp / C 2 kxk 2 kyk 2 for all x; y 2 AC and L D 2mmp .

Theorem 3.128. Let ' W AC AC ! Œ0; 1/ be a function such that there exists an L < 1 with x y '.x; y/ 2m L' m ; m 2 2 for all x; y 2 AC . Let f W AC ! BC be a mapping satisfying (3.180). Then there exists a unique positive-additive mapping T W AC ! AC satisfying (3.178) and kf .x/ T.x/k

1 '.x; x/ 2m 2m L

for all x 2 AC . Proof. Let .X; d/ be the generalized metric space defined in the proof of Theorem 3.126. Consider the linear mapping J W X ! X such that Jg.x/ WD

1 g .2m x/ 2m

for all x 2 AC . It follows from (3.182) that f .x/ 1 f .2m x/ 1 '.x; x/ 2m m 2 for all x 2 AC and so d.f ; Jf / 21m . The rest of the proof is similar to the proof of Theorem 3.126. This completes the proof.


148

Corollary 3.129. Let 0 < p < 1, 1 , 2 be non-negative real numbers and f W AC ! BC be a mapping satisfying (3.184). Then there exists a unique positiveadditive mapping T W AC ! BC satisfying (3.178) and kf .x/ T.x/k

21 C 2 kxkp 2m 2mp


p

'.x; y/ D 1 .kxkp C kykp / C 2 kxk 2 kyk 2 for all x; y 2 AC and L D 2mpm .

3.10.2 Stability of the Positive-Additive Functional Equations: The Direct Method Now, using the direct method of Hyers, we prove the Hyers-Ulam stability of the positive-additive functional equation (3.178) in C -algebras. Theorem 3.130. Let f W AC ! BC be a mapping for which there exists a function ' W AC AC ! Œ0; 1/ satisfying (3.180) and '.x; Q y/ WD

1 X jD1

2mj '

x y 1 and f W A ! B be a mapping satisfying kf .2x/ C f .2y/ C 2f .z/kB p

p

p

k2f .x C y C z/kB C .kxkA C kykA C kzkA / and p

p

kf .xy/ f .x/f .y/kB kxkA kykA for all 2 T1 and x; y; z 2 A whenever the multiplication is defined. Then there exists a unique proper CQ -algebra homomorphism h W A ! B such that kf .x/ h.x/kB

2 p kxkA 2p 2

for all x 2 A. Proof. Let ' W A3 ! Œ0; 1/ be a mapping defined by p

p

p

'.x; y; z/ D .kxkA C kykA C kzkA / for all x; y; z 2 A and W A2 ! Œ0; 1/ be a mapping defined by p

p

.x; y/ D kxkA kykA for all x; y 2 A. For any p > 1, we have '.x; Q y; z/ < 1 and lim 4 n

n!1

x y ; .2/n .2/n

4n p p kxkA kykA D 0 n!1 22pn

D lim

178

4 Stability of Functional Inequalities in Banach Algebras

for all x; y; z 2 A. By applying Theorem 4.10, there exists a unique proper CQ algebra homomorphism h W A ! B such that kf .x/ h.x/kB

2 1 p '.0; Q x; x/ D p kxkA 2 2 2

for all x 2 A. This completes the proof.

4.1.3 Stability of Derivations in Proper CQ -Algebras Now, we consider the Hyers-Ulam stability of derivations on proper CQ -algebras associated with the additive functional inequality. Theorem 4.13. Let f W A ! A be a mapping with f .0/ D 0. Suppose that there exists a function ' W A3 ! Œ0; 1/ such that kf .2x/ C f .2y/ C 2f .z/kA k2f .x C y C z/kA C '.x; y; z/

(4.18)

1 X 1 '.x; Q y; z/ WD ' .2/j x; .2/j y; .2/j z < 1 j 2 jD0

(4.19)

and

for all 2 T1 and x; y; z 2 A. If, in addition, there exists a function such that kf .xy/ f .x/y xf .y/kA

.x; y/

W A2 ! Œ0; 1/

(4.20)

and lim

n!1

1 ..2/n x; .2/n y/ D 0 4n

(4.21)

for all x; y 2 A whenever the multiplication is defined, then there exists a unique derivation ı on A such that kf .x/ ı.x/kA

1 '.0; Q x; x/ 2

(4.22)

for all x 2 A. Proof. By Theorem 4.5, we have a unique C-linear mapping ı W A ! A defined by f ..2/n x/ n!1 .2/n

ı.x/ WD lim for all x 2 A which satisfies (4.22).

4.1 Stability of Additive Functional Inequalities in Banach Algebras

179

Now, we show that ı.xy/ D ı.x/ı.y/ for all x; y 2 A whenever the multiplication is defined. Replacing x; y by .2/n x; .2/n y, respectively, and dividing by 4n in (4.3), we have 1 Œf ..2/n x.2/n y/ f ..2/n x/.2/n y .2/n xf ..2/n y/ 4n A

1 ..2/n x; .2/n y/ 4n

for all x; y 2 A whenever the multiplication is defined. Also, we have 1 f ..2/2n xy/ n n f ..2/ x.2/ y/ D lim D ı.xy/; n!1 4n n!1 .2/2n lim

1 f ..2/n x/ .2/n y n n f ..2/ x/ .2/ y D lim D ı.x/y n!1 4n n!1 .2/n .2/n lim

and 1 .2/n x .2/n y n n .2/ x f ..2/ y/ D lim D xı.y/ n!1 4n n!1 .2/n .2/n lim

for all x; y 2 A whenever the multiplication is defined. If we let n ! 1 in the above inequality, then (4.21) gives ı.xy/ D ı.x/y xı.y/ for all x; y 2 A whenever the multiplication is defined. This completes the proof. Corollary 4.14. Let ; p be nonnegative real numbers with p < 1 and f W A ! A be a mapping satisfying kf .2x/ C f .2y/ C 2f .z/kA p

p

p

k2f .x C y C z/kA C .kxkA C kykA C kzkA / and 2p

2p

kf .xy/ f .x/y xf .y/kA .kxkA C kykA / for all 2 T1 and x; y; z 2 A whenever the multiplication is defined. Then there exists a unique derivation ı on A such that kf .x/ ı.x/kA for all x 2 A.

2 p kxkA 2 2p

180


Proof. Let ' W A3 ! Œ0; 1/ be a mapping defined by p

p

p

'.x; y; z/ D .kxkA C kykA C kzkA / for all x; y; z 2 A. For any p < 1, we have '.x; Q y; z/ W D

D

1 X 1 ' .2/j x; .2/j y; .2/j z j 2 jD0 1 X 2pj jD0

D In addition, let

2j

p

p

p

.kxkA C kykA C kzkA /

2 p p p .kxkA C kykA C kzkA /: 2 2p

W A2 ! Œ0; 1/ be a mapping defined by 2p

2p

.x; y/ D .kxkA C kykA / for all x; y 2 A. For any p < 1, we have 1 22pn 2p 2p n n ..2/ x; .2/ y/ D lim .kxkA C kykA / D 0 n!1 4n n!1 4n lim

for all x; y 2 A. By applying Theorem 4.13, there exists a unique proper CQ -algebra homomorphism h W A ! B such that kf .x/ ı.x/kA

2 1 p '.0; Q x; x/ D kxkA 2 2 2p


Corollary 4.15. Let ; p be nonnegative real numbers with p < 1 and f W A ! B be a mapping satisfying p

p

p

kf .2x/ C f .2y/ C 2f .z/kA k2f .x C y C z/kA C .kxkA C kykA C kzkA / and p

p

kf .xy/ f .x/y xf .y/kA kxkA kykA for all x; y; z 2 A whenever the multiplication is defined. Then there exists a unique derivation ı on A such that kf .x/ ı.x/kA for all x 2 A.

2 p kxkA 2 2p

4.1 Stability of Additive Functional Inequalities in Banach Algebras

181

Proof. Let ' W A3 ! Œ0; 1/ be a mapping defined by p

p

p

'.x; y; z/ D .kxkA C kykA C kzkA / for all x; y; z 2 A and

W A2 ! Œ0; 1/ be a mapping by p

p

.x; y/ D kxkA kykA for all x; y 2 A. For any p < 1, we have '.x; Q y; z/ < 1 and 1 22pn p p n n ..2/ x; .2/ y/ D lim kxkA kykA D 0 n!1 4n n!1 4n lim

for all x; y; z 2 A. By applying Theorem 4.13, there exists a unique proper CQ algebra homomorphism ı W A ! A such that kf .x/ ı.x/kA

2 1 p '.0; Q x; x/ D kxkA 2 2 2p


Theorem 4.16. Let f W A ! A be a mapping. Suppose that there exists a function ' W A3 ! Œ0; 1/ satisfying (4.18) and '.x; Q y; z/ WD

1 X

2i '

iD1

x y z ; ; j j .2/ .2/ .2/j

1 and f W A ! B be a mapping satisfying kf .2x/ C f .2y/ C 2f .z/kA p

p

p

k2f .x C y C z/kA C .kxkA C kykA C kzkA /

184


and p

p

kf .xy/ f .x/y xf .y/kA kxkA kykA for all 2 T1 and x; y; z 2 A whenever the multiplication is defined. Then there exists a unique derivation ı on A such that kf .x/ ı.x/kA

2 p kxkA 2

2p

for all x 2 A. Proof. Let ' W A3 ! Œ0; 1/ be a mapping defined by p

p

p

'.x; y; z/ D .kxkA C kykA C kzkA / for all x; y; z 2 A and

W A2 ! Œ0; 1/ be a mapping defined by p

p

.x; y/ D kxkA kykA for all x; y 2 A. When p > 1, we have '.x; Q y; z/ < 1 and lim 4 n

n!1

x y ; .2/n .2/n

4n p p kxkA kykA D 0 n!1 22pn

D lim

for all x; y; z 2 A. By applying Theorem 4.10, there exists a unique proper CQ algebra homomorphism ı W A ! A such that kf .x/ ı.x/kA

2 1 p '.0; Q x; x/ D p kxkA 2 2 2


4.2 Stability of Functional Inequalities over C -Algebras In this section, we investigate a module linear mapping associated with the following functional inequality. kf .x/ C f .y/ C f .z/ C f .w/k kf .x/ C f .y C z C w/k:

(4.26)

This is applied to understand homomorphisms in C -algebras. Moreover, we prove the Hyers-Ulam stability of the functional following inequality:

4.2 Stability of Functional Inequalities over C -Algebras

185

kf .x/ C f .y/ C f .z/ C f .w/k kf .x/ C f .y C z C w/k p

p

p

p

C.kxkp C kykp C kzkp C kwkp C kxk 4 kyk 4 kzk 4 kwk 4 /

(4.27)

in real Banach spaces. Using the fixed point method, we prove the Hyers-Ulam stability of the functional inequality (4.27) in real Banach spaces.

4.2.1 Functional Inequalities in Normed Modules over C -Algebras Throughout this section, let A be a unital C -algebra with the unitary group U.A/ and the unit e and let B be a C -algebra. Assume that X is a normed A-module with the norm k k and Y is a normed A-module with the norm k k. Now, we investigate an A-linear mapping associated with the functional inequality (4.26). Theorem 4.19. Let f W X ! Y be a mapping such that kf .x/ C f .y/ C f .z/ C uf .w/k kf .x/ C f .y C z C uw/k

(4.28)

for all x; y; z; w 2 X and u 2 U.A/. Then the mapping f W X ! Y is A-linear. Proof. Letting x D y D z D w D 0 and u D e 2 U.A/ in (4.28), we have k4f .0/k k2f .0/k: So f .0/ D 0. Letting x D w D 0 in (4.28), we have kf .y/ C f .z/k kf .y C z/k

(4.29)

for all y; z 2 X. Replacing y and z by x and y C z C w in (4.29), respectively, we get kf .x/ C f .y C z C w/k kf .x C y C z C w/k for all x; y; z; w 2 X and so kf .x/ C f .y/ C f .z/ C f .w/k kf .x C y C z C w/k for all x; y; z; w 2 X. Letting z D w D 0 and y D x in (4.30), we have kf .x/ C f .x/k kf .0/k D 0

(4.30)

186


for all x 2 X. Hence f .x/ D f .x/ for all x 2 X. Letting z D x y and w D 0 in (4.30), we have kf .x/ C f .y/ f .x C y/k D kf .x/ C f .y/ C f .x y/k kf .0/k D 0 for all x; y 2 X. Thus f .x C y/ D f .x/ C f .y/ for all x; y 2 X. Letting z D uw and x D y D 0 in (4.28), we have k f .uw/ C uf .w/k D kf .uw/ C uf .w/k k2f .0/k D 0 for all w 2 X and all u 2 U.A/. Thus we have f .uw/ D uf .w/

(4.31)

for all u 2 U.A/ and all w 2 X. Now, let a 2 A .a ¤ 0/ and M be an integer greater than 4jaj. Then we have j

a 1 2 1 j< l and x 2 X. Assume that p > 1. It follows from (4.36) that the sequence f2k f . 2xk /g is a Cauchy sequence for all x 2 X. Since Y is complete, the sequence f2k f . 2xk /g converges and so one can define the mapping A W X ! Y by x A.x/ WD lim 2k f k k!1 2 for all x 2 X. Letting l D 0 and m ! 1 in (4.36), we get kf .x/ A.x/k

2p C 2 jjxjjp 2p 2

188


for all x 2 X. It follows from (4.33) that x y z w k 2 f k C 2k f k C 2k f k C 2k f k 2 2 2 2 k x yCzCw k 2 f k C 2 f 2 2k C

p p p p 2k .kxkp C kykp C kzkp C kwkp C kxk 4 kyk 4 kzk 4 kwk 4 / (4.37) pk 2

for all x; y; z; w 2 X. Letting k ! 1 in (4.37), we get kA.x/ C A.y/ C A.z/ C A.w/k kA.x/ C A.y C z C w/k

(4.38)

for all x; y; z; w 2 X. It is easy to show that A W X ! Y is odd. Letting w D y z and x D 0 in (4.38), we get A.y C z/ D A.y/ C A.z/ for all y; z 2 X. So, there exists a Cauchy additive mapping A W X ! Y satisfying (4.34) for the case p > 1. Now, let T W X ! Y be another Cauchy additive mapping satisfying (4.33). Then we have x x kA.x/ T.x/k D 2q A q T q 2 2 x x x x 2q L q f q C T q f q 2 2 2 2 2p C 2 2 2q p kxkp ; 2 2 2pq which tends to zero as q ! 1 for all x 2 X. So, we can conclude that A.x/ D T.x/ for all x 2 X. This proves the uniqueness of A. Assume that p < 1. It follows from (4.35) that p f .x/ 1 f .2x/ 2 C 2 jjxjjp 2 2 for all x 2 X. Hence we have m1 p X 2pj 1 f .2l x/ 1 f .2m x/ 2 C 2 kxkp 2l j 2m 2 2 jDl

(4.39)

for all m; l 1 with m > l and x 2 X. It follows from (4.39) that the sequence f 21k f .2k x/g is a Cauchy sequence for all x 2 X. Since Y is complete, the sequence f 21k f .2k x/g converges and so one can define the mapping A W X ! Y by A.x/ WD lim

k!1

1 k f 2x 2k

for all x 2 X. Letting l D 0 and m ! 1 in (4.39), we get


kf .x/ A.x/k

189

2p C 2 jjxjjp 2 2p

for all x 2 X. The rest of the proof is similar to the above proof. So, there exists a unique Cauchy additive mapping A W X ! Y such that kf .x/ A.x/k

2p C 2 kxkp j2p 2j

(4.40)

for all x 2 X. Assume that f W X ! Y is a mapping such that the transformation t ! f .tx/ is continuous in t 2 R for each fixed x 2 X. By the same reasoning as in the proof of Theorem 4.19, one can prove that A is an R-linear mapping. This completes the proof. Using the fixed point method, we prove the Hyers-Ulam stability of the functional inequality (4.27) in Banach spaces. Theorem 4.22. Let f W X ! Y be an odd mapping for which there exists a function ' W X 4 ! Œ0; 1/ such that there exists L < 1 such that '.x; y; z; w/

1 L'.2x; 2y; 2z; 2w/ 2

for all x; y; z; w 2 X and kf .x/ C f .y/ C f .z/ C f .w/k kf .x/ C f .y C z C w/k C '.x; y; z; w/

(4.41)

for all x; y; z; w 2 X. Then there exists a unique Cauchy additive mapping A W X ! Y such that kf .x/ A.x/k

L '.0; x; x; 2x/ 2 2L

(4.42)

for all x 2 X. Proof. Consider the set S WD fg W X ! Yg and introduce the generalized metric on S as follows: d.g; h/ D inffK 2 RC W kg.x/ h.x/k K'.0; x; x; 2x/; 8x 2 Xg: which .S; d/ is complete. Now, we consider the linear mapping J W S ! S defined by Jg.x/ WD 2g

x 2

190


for all x 2 X. Now, we have d.Jg; Jh/ Ld.g; h/ for all g; h 2 S. Since f W X ! Y is odd, f .0/ D 0 and f .x/ D f .x/ for all x 2 X. Letting z D y D x and w D 2x in (4.41), we have k2f .x/ f .2x/k D k2f .x/ C f .2x/k '.0; x; x; 2x/

(4.43)

for all x 2 X. It follows from (4.43) that x x x L ' 0; ; ; x '.0; x; x; 2x/ f .x/ 2f 2 2 2 2 for all x 2 X. Hence d.f ; Jf / A W X ! Y satisfying the following:

L 2.


(1) A is a fixed point of J, i.e., A

x 2

D

1 A.x/ 2

(4.44)

for all x 2 X. Then A W X ! Y is an odd mapping. The mapping A is a unique fixed point of J in the set M D fg 2 S W d.f ; g/ < 1g: This implies that A is a unique mapping satisfying (4.44) such that there exists a K 2 .0; 1/ satisfying kf .x/ A.x/k K'.0; x; x; 2x/ for all x 2 X; (2) d.J n f ; A/ ! 0 as n ! 1. This implies the equality lim 2n f

n!1

x D A.x/ 2n

for all x 2 X; 1 (3) d.f ; A/ 1L d.f ; Jf /, which implies the inequality d.f ; A/

L : 2 2L

This implies that the inequality (4.42) holds.

(4.45)


191

It follows from (4.41) and (4.45) that kA.x/ C A.y/ C A.z/ C A.w/k kA.x/ C A.y C z C w/k for all x; y; z; w 2 X. By Theorem 4.19, the mapping A W X ! Y is a Cauchy additive mapping. Therefore, there exists a unique Cauchy additive mapping A W X ! Y satisfying (4.43). This completes the proof. Corollary 4.23. Let r > 1 and be nonnegative real numbers and f W X ! Y be an odd mapping such that kf .x/ C f .y/ C f .z/ C f .w/k kf .x/ C f .y C z C w/k r

r

r

r

C.kxkr C kykr C kzkr C kwkr C kxk 4 kyk 4 kzk 4 kwk 4 /

(4.46)

for all x; y; z; w 2 X. Then there exists a unique Cauchy additive mapping A W X ! Y such that kf .x/ A.x/k

2r C 2 kxkr 2r 2

for all x 2 X. Proof. The proof follows from Theorem 4.22 by taking '.x; y; z; w/ r

r

r

r

WD .kxkr C kykr C kzkr C kwkr C kxk 4 kyk 4 kzk 4 kwk 4 / for all x; y; z; w 2 X and L D 21r and so we get the desired result.

Remark 4.24. Let f W X ! Y be an odd mapping for which there exists a function ' W X 4 ! Œ0; 1/ satisfying (4.41). By the similar method to the proof of Theorem 4.22, one can show that, if there exists L < 1 such that x y z w ; ; ; '.x; y; z; w/ 2L' 2 2 2 2 for all x; y; z; w 2 X, then there exists a unique Cauchy additive mapping A W X ! Y such that kf .x/ A.x/k

1 '.0; x; x; 2x/ 2 2L

for all x 2 X. For the case 0 < r < 1, one can obtain the similar result to Corollary 4.23: Let 0 < r < 1, 0 be real numbers and f W X ! Y be an odd mapping satisfying (4.46). Then there exists a unique Cauchy additive mapping A W X ! Y such that

192


kf .x/ A.x/k

2 C 2r kxkr 2 2r

for all x 2 X.

4.2.2 On Additive Functional Inequalities in Normed Modules over C -Algebras In [12], An investigated the following additive functional inequality: kf .x/ C f .y/ C f .z/ C f .w/k kf .x C y/ C f .z C w/k

(4.47)

in normed modules over a C -algebra. This is applied to understand homomorphisms in C -algebras. Moreover, he proved the Hyers-Ulam stability of the functional following inequality: kf .x/ C f .y/ C f .z/ C f .w/k kf .x C y C z C w/k C kxkp kykp kzkp kwkp

(4.48)

in real Banach spaces, where and p are positive real numbers with 4p ¤ 1. Gilányi [126] showed that, if f satisfies the functional following inequality: k2f .x/ C 2f .y/ f .x y/k kf .x C y/k;

(4.49)

then f satisfies the Jordan-von Neumann functional equation: 2f .x/ C 2f .y/ D f .x C y/ C f .x y/: Fechner [113] and Gilányi [127] proved the Hyers-Ulam stability of the functional inequality (4.49). Park et al. [253] investigated the functional following inequality: kf .x/ C f .y/ C f .z/k kf .x C y C z/k

(4.50)

in Banach spaces and proved the Hyers-Ulam stability of the functional inequality (4.50) in Banach spaces. Now, let A be a unital C -algebra with the unitary group U.A/ and the unit e and let B be a C -algebra. Assume that X is a normed A-module with the norm k kX and Y is a normed A-module with the norm k kY . Theorem 4.25. Let f W X ! Y be a mapping such that kuf .x/ C f .y/ C f .z/ C f .w/kY kf .ux C y/ C f .z C w/kY

(4.51)


193

for all x; y; z; w 2 X and u 2 U.A/. Then the mapping f W X ! Y is A-linear. Proof. Letting x D y D z D w D 0 and u D e 2 U.A/ in (4.51), we have k4f .0/kY k2f .0/kY and so f .0/ D 0. Letting z D w D 0 in (4.51), we have kf .x/ C f .y/kY kf .x C y/kY

(4.52)

for all x; y 2 X. Replacing x and y by x C y and z C w in (4.52), respectively, we have kf .x C y/ C f .z C w/kY kf .x C y C z C w/kY for all x; y; z; w 2 X and so kf .x/ C f .y/ C f .z/ C f .w/kY kf .x C y C z C w/kY

(4.53)

for all x; y; z; w 2 X. Letting z D w D 0 and y D x in (4.53), we have kf .x/ C f .x/kY kf .0/kY D 0 for all x 2 X. Hence f .x/ D f .x/ for all x 2 X. Letting z D x y and w D 0 in (4.53), we have kf .x/ C f .y/ f .x C y/kY D kf .x/ C f .y/ C f .x y/kY kf .0/kY D 0 for all x; y 2 X. Thus we have f .x C y/ D f .x/ C f .y/ for all x; y 2 X. Letting y D ux and y D w D 0 in (4.51), we have kf .ux/ f .ux/kY D kf .ux/ C f .ux/kY k2f .0/kY D 0 for all x 2 X and u 2 U.A/. Thus we have f .ux/ D uf .x/

(4.54)

for all u 2 U.A/ and x 2 X. Now, let a 2 A with a ¤ 0 and M be an integer greater than 4jaj. Then we have ˇaˇ 1 2 1 ˇ ˇ ˇ ˇ< l and x 2 X. Assume that p > 14 . It follows from (4.59) that the sequence f3k f . 3xk /g is a Cauchy sequence for all x 2 X. Since Y is complete, the sequence f3k f . 3xk /g converges and so one can define the mapping L W X ! Y by L.x/ WD lim 3k f k!1

x 3k

for all x 2 X. Letting l D 0 and m ! 1 in (4.59), we have kf .x/ L.x/k

3p jjxjj4p 81p 3

for all x 2 X. It follows from (4.56) that x y z w k 3 f k C 3k f k C 3k f k C 3k f k 3 3 3 3 k k xCyCzCw C 3 kxkp kykp kzkp kwkp 3 f 81pk 3k

(4.60)

for all x; y; z; w 2 X. Letting k ! 1 in (4.60), we get kL.x/ C L.y/ C L.z/ C L.w/k kL.x C y C z C w/k

(4.61)

for all x; y; z; w 2 X. It is easy to show that L W X ! Y is odd. Letting z D x y and w D 0 in (4.61), we get L.x C y/ D L.x/ C L.y/ for all x; y 2 X. So, there exists an additive mapping L W X ! Y satisfying (4.57) for the case p > 14 . Now, let T W X ! Y be another additive mapping satisfying (4.57). Then we have x x kL.x/ T.x/k D 3q L q T q 3 3 x x x x 3q L q f q C T q f q 3 3 3 3 3p 2 3q p jjxjj4p ; 81 3 81q which tends to zero as q ! 1 for all x 2 X. So we can conclude that L.x/ D T.x/ for all x 2 X. This proves the uniqueness of L.

196


Assume that p < 14 . It follows from (4.58) that f .x/ 1 f .3x/ 3p1 kxk4p 3 for all x 2 X. Hence we have m1 X 81pj 1 f .3l x/ 1 f .3m x/ 3p1 jjxjj4p 3l j 3m 3 jDl

(4.62)

for all m; l 1 with m > l and x 2 X. It follows from (4.62) that the sequence f 31k f .3k x/g is a Cauchy sequence for all x 2 X. Since Y is complete, the sequence f 31k f .3k x/g converges and so one can define the mapping L W X ! Y by 1 k f 3x k!1 3k

L.x/ WD lim

for all x 2 X. Letting l D 0 and m ! 1 in (4.62), we have kf .x/ L.x/k

3p jjxjj4p 3 81p

for all x 2 X. The rest of the proof is similar to the above proof. So, there exists a unique additive mapping L W X ! Y satisfying kf .x/ L.x/k

3p kxk4p j81p 3j

(4.63)

for all x 2 X. Assume that f W X ! Y is a mapping such that the transformation t ! f .tx/ is continuous in t 2 R for each fixed x 2 X. By the same reasoning as in the proof of Theorem 4.25, one can prove that L is an R-linear mapping This completes the proof.

4.2.3 Generalization of Cauchy-Rassias Inequalities via the Fixed Point Method Now, we improve the results of An’s results [12], which presented at last pages. In fact, we get a better error estimation of main result of An by applying the fixed point alternative theorem.


197

Theorem 4.28. Let X be a real normed linear space and Y be a real Banach space. Assume that f W X ! Y is an approximately additive odd mapping satisfying the general Cauchy-Rassias inequality: kf .x/ C f .y/ C f .z/ C f .w/k kf .x C y C z C w/k C '.x; y; z; w/

(4.64)

for all x; y; z; w 2 X, where ' W X 4 ! Œ0; 1/ is a given function. If there exists 0 < L < 1 such that '.x; y; z; w/

1 L'.3x; 3y; 3z; 3w/ 3

(4.65)

for all x; y; z; w 2 X. Then there exists a unique additive mapping A W X ! Y such that kf .x/ A.x/k

L '.x; x; x; 3x/ 3 3L

(4.66)

for all x 2 X. If, in addition, f W X ! Y is a mapping such that the transformation t ! f .tx/ is continuous in t 2 R for each fixed x 2 X; then A is an R-linear mapping. Proof. Since f is odd, f .0/ D 0 and f .x/ D f .x/ for all x 2 X. Consider the set S WD fg W X ! Yg and introduce the generalized metric on S as follows: d.g; h/ D inffC 2 RC W kg.x/ h.x/k C'.x; x; x; 3x/; 8x 2 Xg: Now, we show that .S; d/ is complete. Let fhn g be a Cauchy sequence in .S; d/. Then, for any " > 0, there exists an integer N" > 0 such that d.hm ; hn / < " for all m; n N" . Then there exists C 2 .0; "/ such that khm .x/ hn .x/k C'.x; x; x; 3x/ "'.x; x; x; 3x/

(4.67)

for all m; n N" and x 2 X. Since Y is complete, fhn .x/g converges for each x 2 X. Thus a mapping h W X ! Y can be defined by h.x/ WD lim hn .x/ n!1

(4.68)

for all x 2 X. Letting n ! 1 in (4.67), we have m N" H) khm .x/ hn .x/k "'.x; x; x; 3x/ H) " 2 fC 2 RC W kg.x/ h.x/k C'.x; x; x; 3x/; 8x 2 Xg H) d.hm ; h/ " for all x 2 X. This means that the Cauchy sequence fhn g converges to h in .S; d/. Hence .S; d/ is complete.

198


Now, we consider the linear mapping W S ! S defined by g.x/ WD 3g

x

(4.69)

3

for all x 2 X. We show that is a strictly contractive on S. For any g; h 2 S, let Cg;h 0 be an arbitrary constant with d.g; h/ Cg;h . Then we have d.g; h/ Cg;h H) kg.x/ h.x/k Cg;h '.x; x; x; 3x/ x x x x x 3h k 3Cg;h ' ; ; ; x H) k3g 3 3 3 3 3 x x 3h k LCg;h '.x; x; x; 3x/ H) k3g 3 3 for all x 2 X. This means d.g; h/ LCg;h . Hence we see that d.g; h/ Ld.g; h/ for any g; h 2 S. Therefore, is a strictly contractive on S with the Lipschitz constant 0 < L < 1. Letting y D z D x and w D 3x in (4.64), we have k3f .x/ f .3x/k '.x; x; x; 3x/

(4.70)

for all x 2 X and so x x x x 1 ; ; ; x L'.x; x; x; 3x/ f .x/ 3f ' 3 3 3 3 3 for all x 2 X. Thus we have d.f ; f /

L : 3

Therefore, it follows of Theorem 1.3 that the sequence fn f g converges to the unique fixed point A of , i.e., A.x/ D .f /.x/ DWD lim 3k f k!1

x 3k

and A.3x/ D 3A.x/ for all x 2 X. Also, we have d.A; f /

L 1 d.f ; f / ; 1L 3 3L

which means that (4.66) holds. Assume that f W X ! Y is a mapping such that the transformation t ! f .tx/ is continuous in t 2 R for each fixed x 2 X. By the same reasoning as in the proof of Theorem 4.25, one can prove that A is an R-linear mapping. This completes the proof.


199

In the following, we get a better error estimation of the main result of An [12]. Corollary 4.29. Let X be a real normed linear space and Y be a real Banach space. Assume that f W X ! Y is an approximately additive odd mapping for which there exist the constants 0 and p 2 R such that 4p ¤ 1 and f satisfies the general Cauchy-Rassias inequality: kf .x/ C f .y/ C f .z/ C f .w/k kf .x C y C z C w/k C kxkp kykp kzkp kwkp for all x; y; z; w 2 X. Then there exists a unique additive mapping A W X ! Y such that kf .x/ A.x/k

3p1 kxk4p j81p 3j

for all x 2 X. If, in addition, f W X ! Y is a mapping such that the transformation t ! f .tx/ is continuous in t 2 R for each fixed x 2 X; then A is an R-linear mapping. Proof. In Theorem 4.28, take '.x; y; z; w/ WD kxkp kykp kzkp kwkp for all x; y; z; w 2 X and L D 811p . Then we have desired result.

Corollary 4.30. Let X be a real normed linear space and Y be a real Banach space. Assume that f W X ! Y is an approximately additive odd mapping for which there exist the constants 0 and p 2 R such that 4p ¤ 1 and f satisfies the general Cauchy-Rassias inequality: kf .x/ C f .y/ C f .z/ C f .w/k kf .x C y C z C w/k C .kxkp C kykp C kzkp C kwkp / for all x; y; z; w 2 X. Then there exists a unique additive mapping A W X ! Y such that kf .x/ A.x/k

.3p C 3/ kxkp j3p 3j

for all x 2 X. If, in addition, f W X ! Y is a mapping such that the transformation t ! f .tx/ is continuous in t 2 R for each fixed x 2 X; then A is an R-linear mapping. Proof. In Theorem 4.28, take '.x; y; z; w/ WD kxkp C kykp C kzkp C kwkp for all x; y; z; w 2 X and L D 31p . Then we have desired result.

Chapter 5

Stability of Functional Equations in C -Ternary Algebras

Ternary algebraic operations were considered in the nineteenth century by several mathematicians such as Cayley [68] who introduced the notion of cubic matrix which, in turn, was generalized by Kapranov et al. [169]. The simplest example of such non-trivial ternary operation is given by the following composition rule: fa; b; cgijk D

X

anil bljm cmkn

1l;m;nN

for each i; j; k D 1; 2; ; N. Ternary structures and their generalization, the so-called n-ary structures, raise certain hopes in view of their applications in physics. Some significant applications are as follows (see [171] and [172]): (1) The algebra of nonions generated by two matrices 0

1 010 @0 0 1A; 100

1 0 1 0 @ 0 0 !A ! 2 0 0; 0

2i

where ! D e 3 , was introduced by Sylvester as a ternary analog of Hamiltons quaternions (see [2]). (2) The quark model inspired a particular brand of ternary algebraic systems. The so-called Nambu mechanics is based on such structures (see [93]). There are also some applications, although still hypothetical, in the fractional quantum Hall effect, the nonstandard statistics, supersymmetric theory and the Yang-Baxter equation (see [2, 171] and [325]). In Sect. 5.1, we prove the Hyers-Ulam stability of C -ternary 3-derivations and of C -ternary 3-homomorphisms for the following functional equation:


201

5 Stability of Functional Equations in C -Ternary Algebras

202

f .x1 C x2 ; y1 C y2 ; z1 C z2 / D

X

f .xi ; yj ; zk /

1i;j;k2

in C -ternary algebras. In Sect. 5.2, we consider the following Apollonius type additive functional equation: xCy 1 f .z x/ C f .z y/ D f .x C y/ C 2f z 2 4 and prove the superstability of C -ternary homomorphisms, C -ternary derivations, JB -triple homomorphisms and JB -triple derivations by using the fixed point method. In Sect. 5.3, we prove the Hyers-Ulam stability of bi--derivations on JB -triples.

5.1 C -Ternary 3-Homomorphism and C -Ternary 3-Derivations In this section, we prove the Hyers-Ulam stability of C -ternary 3-derivations and of C -ternary 3-homomorphisms for the following functional equation: f .x1 C x2 ; y1 C y2 ; z1 C z2 / D

X

f .xi ; yj ; zk /

1i;j;k2

in C -ternary algebras (see [94]). Let X and Y be complex vector spaces. A mapping f W XXX ! Y is called a 3additive mapping if f is additive for each variable and a mapping f W XXX ! Y is called a 3-C-linear mapping if f is C-linear for each variable. A 3-C-linear mapping H W A A A ! B is called a C -ternary 3-homomorphism if it satisfies H.Œx1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / D ŒH.x1 ; x2 ; x3 /; H.y1 ; y2 ; y3 /; H.z1 ; z2 ; z3 / for all x1 ; y1 ; z1 ; x2 ; y2 ; z2 ; x3 ; y3 ; z3 2 A. For a given mapping f W A3 ! B, we define D; ; f .x1 ; x2 ; y1 ; y2 ; z1 ; z2 / WD f .x1 C x2 ; y1 C y2 ; z1 C z2 /

X

f .xi ; yj ; zk /

1i;j;k2

for all ; ; 2 S1 WD f 2 C W jj D 1g and x1 ; x2 ; y1 ; y2 ; z1 ; z2 2 A.

5.1 C -Ternary 3-Homomorphism and C -Ternary 3-Derivations

203

Bae and Park [21] proved the Hyers–Ulam stability of 3-homomorphisms in C -ternary algebras for the following functional equation: D; ; f .x1 ; x2 ; y1 ; y2 ; z1 ; z2 / D 0: Lemma 5.1 ([21]). Let X and Y be complex vector spaces and f W X X X ! Y be a 3-additive mapping such that f .x; y; z/ D f .x; y; z/ for all ; ; 2 S1 and x; y; z 2 X. Then f is 3-C-linear. Theorem 5.2 ([21]). Let p; q; r 2 .0; 1/ with p C q C r < 3 and 2 .0; 1/ and f W A3 ! B be a mapping such that kD; ; f .x1 ; x2 ; y1 ; y2 ; z1 ; z2 /k

(5.1)

maxfkx1 k; kx2 kg maxfky1 k; ky2 kg maxfkz1 k; kz2 kg p

q

r

and kf .Œx1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / Œf .x1 ; x2 ; x3 /; f .y1 ; y2 ; y3 /; f .z1 ; z2 ; z3 /k

3 X

(5.2)

kxi kp kyi kq kzi kr

iD1

for all ; ; 2 S1 and x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 2 A. Then there exists a unique C -ternary 3-homomorphism H W A3 ! B such that kf .x; y; z/ H.x; y; z/k

23

kxkp kykq kzkr 2pCqCr

(5.3)

for all x; y; z 2 A.

5.1.1 C -Ternary 3-Homomorphisms in C -Ternary Algebras Now, we investigate C -ternary 3-homomorphisms in C -ternary algebras. Theorem 5.3. Let p; q; r 2 .0; 1/ with p C q C r < 3 and 2 .0; 1/ and f W A3 ! B be a mapping satisfying (5.1) and (5.2). If there exists .x0 ; y0 ; z0 / 2 A3 such that lim

n!1

1 f .2n x0 ; 2n y0 ; 2n z0 / D e0 ; 8n

then the mapping f is a C -ternary 3-homomorphism.


204

Proof. By Theorem 5.2, there exists a unique C -ternary 3-homomorphism H W A3 ! B satisfying (5.3). Note that 1 f .2n x; 2n y; 2n z/ n!1 8n

H.x; y; z/ WD lim

for all x; y; z 2 A. By the assumption, we have H.x0 ; y0 ; z0 / D lim

n!1

1 f .2n x0 ; 2n y0 ; 2n z0 / D e0 : 8n

It follows from (5.2) that kŒH.x1 ; x2 ; x3 /; H.y1 ; y2 ; y3 /; H.z1 ; z2 ; z3 / ŒH.x1 ; x2 ; x3 /; H.y1 ; y2 ; y3 /; f .z1 ; z2 ; z3 /k D kH.Œx1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / ŒH.x1 ; x2 ; x3 /; H.y1 ; y2 ; y3 /; f .z1 ; z2 ; z3 /k D lim

n!1

1 kf .Œ2n x1 ; 2n y1 ; z1 ; Œ2n x2 ; 2n y2 ; z2 ; Œ2n x3 ; 2n y3 ; z3 / 82n Œf .2n x1 ; 2n x2 ; 2n x3 /; f .2n y1 ; 2n y2 ; 2n y3 /; f .z1 ; z2 ; z3 /k

3 2n.pCq/ X kxi kp kyi kq kzi kr n!1 82n iD1

lim D0

for all x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 2 A and so ŒH.x1 ; x2 ; x3 /; H.y1 ; y2 ; y3 /; H.z1 ; z2 ; z3 / D ŒH.x1 ; x2 ; x3 /; H.y1 ; y2 ; y3 /; f .z1 ; z2 ; z3 / for all x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 2 A. Letting x1 D y1 D x0 ; x2 D y2 D y0 and x3 D y3 D z0 in the last equality, we get f .z1 ; z2 ; z3 / D H.z1 ; z2 ; z3 / for all z1 ; z2 ; z3 2 A. Therefore, the mapping f is a C -ternary 3-homomorphism. This completes the proof. Theorem 5.4. Let pi ; qi ; ri 2 .0; 1/ .i D 1; 2; 3/ such that pi ¤ 1 or qi ¤ 1 or ri ¤ 1 for some 1 i 3, ; 2 .0; 1/ and f W A3 ! B be a mapping such that kD; ; f .x1 ; x2 ; y1 ; y2 ; z1 ; z2 /k .kx1 kp1 kx2 kp2 ky1 kq1 ky2 kq2 Cky1 kq1 ky2 kq2 kz1 kr1 kz2 kr2 C kx1 kp1 kx2 kp2 kz1 kr1 kz2 kr2 /

(5.4)


205

and kf .Œx1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / Œf .x1 ; x2 ; x3 /; f .y1 ; y2 ; y3 /; f .z1 ; z2 ; z3 /k kx1 kp1 kx2 kp2 kx3 kp3

(5.5)

ky1 k ky2 k ky3 k kz1 k kz2 k kz3 k q1

q2

q3

r1

r2

r3

for all ; ; 2 S1 and x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 2 A. Then the mapping f W A3 ! B is a C -ternary 3-homomorphism. Proof. Letting xi D yj D zk D 0 .i; j; k D 1; 2/ in (5.4), we get f .0; 0; 0/ D 0. Putting D D D 1; x2 D 0 and yj D zk D 0 .j; k D 1; 2/ in (5.4), we have f .x1 ; 0; 0/ D 0 for all x1 2 A. Similarly, we get f .0; y1 ; 0/ D f .0; 0; z1 / D 0 for all y1 ; z1 2 A. Setting D D D 1; x2 D 0; y2 D 0 and z1 D z2 D 0, we have f .x1 ; y1 ; 0/ D 0 for all x1 ; y1 2 A. Similarly, we get f .x1 ; 0; z1 / D f .0; y1 ; z1 / D 0 for all x1 ; y1 ; z1 2 A. Now, letting D D D 1 and y2 D z2 D 0 in (5.4), we have f .x1 C x2 ; y1 ; z1 / D f .x1 ; y1 ; z1 / C f .x2 ; y1 ; z1 / for all x1 ; x2 ; y1 ; z1 2 A. Similarly, one can show that the other equations hold and so f is 3-additive. Letting x2 D y2 D z2 D 0 in (5.4), we get f .x1 ; y1 ; z1 / D f .x1 ; y1 ; z1 / for all ; ; 2 S1 and x1 ; y1 ; z1 2 A. So, by Lemma 5.1, the mapping f is 3-C-linear. Without any loss of generality, we may suppose that p1 ¤ 1. Let p1 < 1. It follows from (5.5) that kf .Œx1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / Œf .x1 ; x2 ; x3 /; f .y1 ; y2 ; y3 /; f .z1 ; z2 ; z3 /k D lim

n!1

1 kf .Œ3n x1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / 3n Œf .3n x1 ; x2 ; x3 /; f .y1 ; y2 ; y3 /; f .z1 ; z2 ; z3 /k

3np1 .kx1 kp1 kx2 kp2 kx3 kp3 ky1 kq1 ky2 kq2 n!1 3n ky3 kq3 kz1 kr1 kz2 kr2 kz3 kr3 /

lim

D0


206

for all x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 2 A. Let p1 > 1. It follows from (5.5) that kf .Œx1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / Œf .x1 ; x2 ; x3 /; f .y1 ; y2 ; y3 /; f .z1 ; z2 ; z3 /k h 1 i D lim 3n f x ; y ; z 1 1 1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 n n!1 3 i h 1 f n x1 ; x2 ; x3 ; f .y1 ; y2 ; y3 /; f .z1 ; z2 ; z3 / 3 3n lim np .kx1 kp1 kx2 kp2 kx3 kp3 ky1 kq1 ky2 kq2 n!1 3 1 ky3 kq3 kz1 kr1 kz2 kr2 kz3 kr3 / D0 for all x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 2 A. Therefore, we have f .Œx1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / D Œf .x1 ; x2 ; x3 /; f .y1 ; y2 ; y3 /; f .z1 ; z2 ; z3 / for all x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 2 A. Therefore, the mapping f W A3 ! B is a C -ternary 3-homomorphism. This completes the proof. Remark 5.5. Let ' W A6 ! Œ0; 1/ and

W A9 ! Œ0; 1/ be the functions such that

'.x1 ; ; x6 / D 0 if xi D 0 for some 1 i 6 and 1 .x1 ; ; 3n xi ; ; x9 / D 0 3n or 3n

x1 ; ;

1 D 0: x ; ; x i 9 3n

Suppose that f W A3 ! B is a mapping satisfying kD; ; f .x1 ; x2 ; y1 ; y2 ; z1 ; z2 /k '.x1 ; x2 ; y1 ; y2 ; z1 ; z2 / and kf .Œx1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / Œf .x1 ; x2 ; x3 /; f .y1 ; y2 ; y3 /; f .z1 ; z2 ; z3 /k

.x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 /


207

for all ; ; 2 S1 and x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 2 A. Then the mapping f is a C -ternary 3-homomorphism. Corollary 5.6. Let pi ; qi ; ri 2 .0; 1/ .i D 1; 2; 3/ such that pi ¤ 1 or qi ¤ 1 or ri ¤ 1 for some 1 i 3, ; 2 .0; 1/ and f W A3 ! B be a mapping such that kD; ; f .x1 ; x2 ; y1 ; y2 ; z1 ; z2 /k kx1 kp1 kx2 kp2 ky1 kq1 ky2 kq2 kz1 kr1 kz2 kr2 and kf .Œx1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / Œf .x1 ; x2 ; x3 /; f .y1 ; y2 ; y3 /; f .z1 ; z2 ; z3 /k kx1 kp1 kx2 kp2 kx3 kp3 ky1 kq1 ky2 kq2 ky3 kq3 kz1 kr1 kz2 kr2 kz3 kr3 for all ; ; 2 S1 and x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 2 A. Then the mapping f W A3 ! B is a C -ternary 3-homomorphism.

5.1.2 C -Ternary 3-Derivations in C -Ternary Algebras Now, we investigate C -ternary 3-derivations in C -ternary algebras. Definition 5.7. A 3-C-linear mapping D W A3 ! A is called a C -ternary 3-derivation if it satisfies the following: D.Œx1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / D ŒD.x1 ; x2 ; x3 /; Œy1 ; y2 ; y3 ; Œz1 ; z2 ; z3 CŒŒx1 ; x2 ; x3 ; D.y1 ; y2 ; y3 /; Œz1 ; z2 ; z3 CŒŒx1 ; x2 ; x3 ; Œy1 ; y2 ; y3 ; D.z1 ; z2 ; z3 / for all x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 2 A. Theorem 5.8. Let p; q; r 2 .0; 1/ with p C q C r < 3 and 2 .0; 1/, and let f W A3 ! A be a mapping such that kD; ; f .x1 ; x2 ; y1 ; y2 ; z1 ; z2 /k

(5.6)

maxfkx1 k; kx2 kg maxfky1 k; ky2 kg maxfkz1 k; kz2 kg p

q

and kf .Œx1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / Œf .x1 ; x2 ; x3 /; Œy1 ; y2 ; y3 ; Œz1 ; z2 ; z3

r


208

ŒŒx1 ; x2 ; x3 ; f .y1 ; y2 ; y3 /; Œz1 ; z2 ; z3 ŒŒx1 ; x2 ; x3 ; Œy1 ; y2 ; y3 ; f .z1 ; z2 ; z3 /k

3 X

kxi kp kyi kq kzi kr

(5.7)

iD1

for all ; ; 2 S1 and x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 2 A. Then there exists a unique C -ternary 3-derivation ı W A3 ! A such that kf .x; y; z/ ı.x; y; z/k

kxkp kykq kzkr 23 2pCqCr

(5.8)

for all x; y; z 2 A. Proof. By the same method as in the proof of [21, Theorem 1.2], we obtain a 3-Clinear mapping ı W A3 ! A satisfying (5.8) and the mapping ı.x; y; z/ WD lim

j!1

1 f .2j x; 2j y; 2j z/ 8j

for all x; y; z 2 A. It follows from (5.7) that kı.Œx1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / Œı.x1 ; x2 ; x3 /; Œy1 ; y2 ; y3 ; Œz1 ; z2 ; z3 ŒŒx1 ; x2 ; x3 ; ı.y1 ; y2 ; y3 /; Œz1 ; z2 ; z3 ŒŒx1 ; x2 ; x3 ; Œy1 ; y2 ; y3 ; ı.z1 ; z2 ; z3 /k D lim

n!1

1 kf .23n Œx1 ; y1 ; z1 ; 23n Œx2 ; y2 ; z2 ; 23n Œx3 ; y3 ; z3 / 83n Œf .2n x1 ; 2n x2 ; 2n x3 /; Œ2n y1 ; 2n y2 ; 2n y3 ; Œ2n z1 ; 2n z2 ; 2n z3 ŒŒ2n x1 ; 2n x2 ; 2n x3 ; f .2n y1 ; 2n y2 ; 2n y3 /; Œ2n z1 ; 2n z2 ; 2n z3 ŒŒ2n x1 ; 2n x2 ; 2n x3 ; Œ2n y1 ; 2n y2 ; 2n y3 ; f .2n z1 ; 2n z2 ; 2n z3 /k

3 2n.pCqCr/ X kxi kp kyi kq kzi kr D 0 n!1 83n iD1

lim

for all x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 2 A. Now, let T W A3 ! A be another 3-derivation satisfying (5.8). Then we have 1 kı.2n x; 2n y; 2n z/ T.2n x; 2n y; 2n z/k 8n 1 n kı.2n x; 2n y; 2n z/ f .2n x; 2n y; 2n z/k 8

kı.x; y; z/ T.x; y; z/k D


C

209

1 kf .2n x; 2n y; 2n z/ T.2n x; 2n y; 2n z/k 8n

2.pCqCr3/nC1 kxkp kykq kzkr ; 23 2pCqCr

which tends to zero as n ! 1 for all x; y; z 2 A. So we can conclude that ı.x; y; z/ D T.x; y; z/ for all x; y; z 2 A. This proves the uniqueness of ı. Therefore, the mapping ı W A3 ! A is a unique C -ternary 3-derivation satisfying (5.8). This completes the proof. Corollary 5.9. Let 2 .0; 1/ and f W A3 ! A be a mapping satisfying kD; ; f .x1 ; x2 ; y1 ; y2 ; z1 ; z2 /k and kf .Œx1 ; y1 ; z1 ; Œx2 ; y2 ; z2 ; Œx3 ; y3 ; z3 / Œf .x1 ; x2 ; x3 /; Œy1 ; y2 ; y3 ; Œz1 ; z2 ; z3 ŒŒx1 ; x2 ; x3 ; f .y1 ; y2 ; y3 /; Œz1 ; z2 ; z3 ŒŒx1 ; x2 ; x3 ; Œy1 ; y2 ; y3 ; f .z1 ; z2 ; z3 /k 3 for all ; ; 2 S1 and x1 ; x2 ; x3 ; y1 ; y2 ; y3 ; z1 ; z2 ; z3 2 A. Then there exists a unique C -ternary 3-derivation ı W A3 ! A such that kf .x; y; z/ ı.x; y; z/k

7

for all x; y; z 2 A.

5.2 Apollonius Type Additive Functional Equations C. Park and Th. M. Rassias proved the superstability of C -ternary homomorphisms, C -ternary derivations, JB -triple homomorphisms and JB -triple derivations associated with the following Apollonius type additive functional equation: xCy 1 f .z x/ C f .z y/ D f .x C y/ C 2f z 2 4 by using the direct method (see [244–248]).

210


In this section, under the conditions of the theorems, we can show that the mappings f must be zero and we correct some conditions. Furthermore, we prove the superstability of C -ternary homomorphisms, C -ternary derivations, JB -triple homomorphisms and JB -triple derivations by using fixed point method. In an inner product space, the following equality holds: kz xk2 C kz yk2 D

x C y 1 2 kx yk2 C 2z ; 2 2

which is called the Apollonius’ identity. The following functional equation, which was motivated by this equation: Q.z x/ C Q.z y/ D

1 x C y Q.x y/ C 2Q z ; 2 2

(5.9)

is quadratic. For this reason, the function equation (5.9) is called a quadratic functional equation of Apollonius type and each solution of the functional equation (5.9) is said to be a quadratic mapping of Apollonius type. Jun and Kim [144] investigated the quadratic functional equation of Apollonius type. Now, employing the above equality (5.9), we introduce a new functional equation, which is called the Apollonius type additive functional equation and whose solution of the functional equation is said to be the Apollonius type additive mapping: 1 x C y L.z x/ C L.z y/ D L.x C y/ C 2L z : 2 4

5.2.1 Homomorphisms in C -Ternary Algebras Assume that A is a C -ternary algebra with the norm k kA and that B is a C -ternary algebra with the norm k kB . Now, we investigate homomorphisms in C -ternary algebras. Lemma 5.10. Let f W A ! B be a mapping such that 1 xCy / kf .z x/ C f .z y/ C f .x C y/kB 2f .z B 2 4 for all x; y; z 2 A. Then f is additive.

(5.10)

Proof. Letting x D y D z D 0 in (5.10), we get 52 f .0/ k2f .0/kB and so B f .0/ D 0. Letting z D 0 and y D x in (5.10), we have kf .x/ C f .x/kB k2f .0/kB D 0


211

for all x 2 A. Hence f .x/ D f .x/ for all x 2 A. Letting x D y D 2z in (5.10), we have 1 2f .z/ C f .4z/ k2f .0/kB D 0 B 2 for all z 2 A and so f .4z/ D 4f .z/ D 4f .z/ for all z 2 A. Letting z D

xCy 4

in (5.10), we have

3x C y x 3y 1 Cf C f .x C y/ k2f .0/kB D 0 f B 4 4 2 for all x; y 2 A and so f

3x C y 4

for all x; y 2 A. Let w1 D

3xCy 4

Cf

x 3y 4

and w2 D

1 C f .x C y/ D 0 2

x3y 4

(5.11)

in (5.11). Then we have

w C w 1 1 1 2 f .w1 / C f .w2 / D f .2w1 2w2 / D f .2w1 C 2w2 / D 2f 2 2 2 for all w1 ; w2 2 A and so f is additive. This completes the proof.

Theorem 5.11. Let r ¤ 1, be a nonnegative real number and f W A ! B be a mapping such that f . z x/ C f .z y/ C f .x C y/ 2 B xCy (5.12) 2f z 4 B and kf .Œx; y; z/ Œf .x/; f .y/; f .z/kB 3r 3r .kxk3r A C kykA C kzkA /

(5.13)

for all 2 T1 WD f 2 C W jj D 1g and all x; y; z 2 A. Then the mapping f W A ! B is a C -ternary algebra homomorphism. Proof. Assume r > 1. Let D 1 in (5.12). By Lemma 5.10, the mapping f W A ! B is additive. Letting y D x and z D 0, we get kf . x/ C f .x/kB k2f .0/kB D 0


212

for all x 2 A and 2 T1 and so f . x/ C f .x/ D f . x/ C f .x/ D 0 for all x 2 A and 2 T1 . Hence f . x/ D f .x/ for all x 2 A and 2 T1 . By the same reasoning as in the proof of Theorem 2.1 of Park [227], the mapping f W A ! B is C-linear. It follows from (5.13) that kf .Œx; y; z/ Œf .x/; f .y/; f .z/kB Œx; y; z h x y z i D lim 8n f n n n f n ; f n ; f n n!1 B 2 2 2 2 2 2 8n 3r 3r lim nr .kxk3r A C kykA C kzkA / n!1 8 D0 for all x; y; z 2 A and so f .Œx; y; z/ D Œf .x/; f .y/; f .z/ for all x; y; z 2 A. Hence the mapping f W A ! B is a C -ternary algebra homomorphism. Similarly, one obtains the result for the case r < 1. This completes the proof.

5.2.2 Derivations in C -Ternary Algebras Assume that A is a C -ternary algebra with the norm k kA . Now, we investigate derivations on C -ternary algebras. Theorem 5.12. Let r ¤ 1 and be nonnegative real numbers, and let f W A ! A be a mapping such that f . z x/ C f .z y/ C f .x C y/ 2 A x C y 2f z 4 A and kf .Œx; y; z/ Œf .x/; y; z Œx; f .y/; z Œx; y; f .z/kA 3r 3r .kxk3r A C kykA C kzkA /

for all x; y; z 2 A. Then the mapping f W A ! A is a C -ternary derivation.

(5.14)


213

Proof. Assume r > 1. By the same reasoning as in the proof of Theorem 5.11, the mapping f W A ! A is C-linear. It follows from (5.15) that kf .Œx; y; z/ Œf .x/; y; z Œx; f .y/; z Œx; y; f .z/kA Œx; y; z h x y z i f n ; n; n D lim 8n f n!1 8n 2 2 2 h x y z i h x y z i n;f n ; n n; n;f n A 2 2 2 2 2 2 8n 3r 3r lim nr .kxk3r A C kykA C kzkA / n!1 8 D0 for all x; y; z 2 A and so f .Œx; y; z/ D Œf .x/; y; z C Œx; f .y/; z C Œx; y; f .z/ for all x; y; z 2 A. Thus the mapping f W A ! A is a C -ternary derivation. Similarly, one obtains the result for the case r < 1. This completes the proof.

5.2.3 Homomorphisms in JB -Triples Assume that J is a JB -triple with the norm k kJ and that L is a JB -triple with the norm k kL . Now, we investigate homomorphisms in JB -triples. Theorem 5.13. Let r ¤ 1, be a nonnegative real number and f W J ! L be a mapping such that f . z x/ C f .z y/ C f .x C y/ 2 L xCy 2f z 4 L

(5.15)

and kf .fxyzg/ ff .x/f .y/f .z/gkL 3r 3r .kxk3r J C kykJ C kzkJ /

(5.16)

for all 2 T1 and all x; y; z 2 J . Then the mapping f W J ! L is a JB -triple homomorphism.


214

Proof. Assume r > 1. By the same reasoning as in the proof of Theorem 5.11, the mapping f W J ! L is C-linear. It follows from (5.16) that kf .fxyg/ ff .x/f .y/f .z/gkL fxyzg n x y z o D lim 8n f n n n f n f n f n kL n!1 2 2 2 2 2 2 n 8 3r 3r lim nr .kxk3r J C kykJ C kzkJ / n!1 8 D0 for all x; y; z 2 J and so f .fxyzg/ D ff .x/f .y/f .z/g for all x; y; z 2 J . Hence the mapping f W J ! L is a JB -triple homomorphism. Similarly, one obtains the result for the case r < 1. This completes the proof.

5.2.4 Derivations in JB -Triples Assume that J is a JB -triple with the norm k kJ . Now, we investigate derivations on JB -triples. Theorem 5.14. Let r ¤ 1, be a nonnegative real number and f W J ! J be a mapping such that f . z x/ C f .z y/ C f .x C y/ 2 J xCy 2f z 4 J

(5.17)

and kf .fxyzg/ ff .x/yzg fxf .y/zg fxyf .z/gkJ 3r 3r .kxk3r J C kykJ C kzkJ /

(5.18)

for all x; y; z 2 J . Then the mapping f W J ! J is a JB -triple derivation. Proof. Assume r > 1. By the same reasoning as in the proof of Theorem 5.11, the mapping f W J ! J is C-linear. It follows from (5.18) that


215

kf .fxyzg/ ff .x/yzg fxf .y/zg fxyf .z/gkJ fxyzg n x y z o f n n n D lim 8n f n!1 8n 2 2 2 n x y z o n x y z o nf n n n nf n J 2 2 2 2 2 2 n 8 3r 3r lim nr .kxk3r J C kykJ C kzkJ / n!1 8 D0 for all x; y; z 2 J and so f .fxyzg/ D ff .x/yzg C fxf .y/zg C fxyf .z/g for all x; y; z 2 J . Thus the mapping f W J ! J is a JB -triple derivation. Similarly, one obtains the result for the case r < 1. This completes the proof.

5.2.5 C -Ternary Homomorphisms: Fixed Point Method Now, we prove the superstability of C -ternary homomorphisms by the using fixed point method. Theorem 5.15. Let ' W A3 ! Œ0; 1/ be a function such that there exists ˛ < 1 with x y z ; ; (5.19) '.x; y; z/ 8˛' 2 2 2 for all x; y; z 2 A. Let f W A ! B be a mapping satisfying (5.12) and kf .Œx; y; z/ Œf .x/; f .y/; f .z/kB '.x; y; z/

(5.20)

for all x; y; z 2 A: Then the mapping f W A ! B is a C -ternary homomorphism. Proof. By the same reasoning as in the proof of Theorem 5.11, one can show that the mapping f W A ! B is C-linear. It follows from (5.19) that lim

n!1

1 '.2n x; 2n y; 2n z/ D 0 8n

(5.21)

for all x; y; z 2 A. Since f W A ! B is additive, it follows from (5.20) and (5.21) that f .Œx; y; z/ D Œf .x/; f .y/; f .z/ for all x; y; z 2 A. Thus the mapping f W A ! B is a C -ternary homomorphism. This completes the proof.


216

Theorem 5.16. Let ' W A3 ! Œ0; 1/ be a function such that there exists ˛ < 1 with '.x; y; z/

˛ ' .2x; 2y; 2z/ 2

(5.22)

for all x; y; z 2 A. Let f W A ! B be a mapping satisfying (5.12) and (5.20). Then the mapping f W A ! B is a C -ternary homomorphism. Proof. By the same reasoning as in the proof of Theorem 5.11, one can show that the mapping f W A ! B is C-linear. It follows from (5.22) that lim 2n '

n!1

x y z D0 ; ; 2n 2n 2n

(5.23)

for all x; y; z 2 A. Since f W A ! B is additive, it follows from (5.20) and (5.23) that f .Œx; y; z/ D Œf .x/; f .y/; f .z/ for all x; y; z 2 A. Thus the mapping f W A ! B is a C -ternary homomorphism. This completes the proof. Remark 5.17. Theorem 5.11 follows from Theorems 5.15 and 5.16 by taking '.x; y; z/ D .kxk3r C kyk3r C kzk3r / for all x; y; z 2 A.

5.2.6 C -Ternary Derivations: The Fixed Point Method Now, we prove the superstability of C -ternary derivations by using the fixed point method. Theorem 5.18. Let ' W A3 ! Œ0; 1/ be a function satisfying (5.19). Let f W A ! A be a mapping satisfying (5.14) and kf .Œx; y; z/ Œf .x/; y; z Œx; f .y/; z Œx; y; f .z/kA '.x; y; z/

(5.24)

for all x; y; z 2 A: Then the mapping f W A ! A is a C -ternary derivation. Proof. The proof is similar to the proof of Theorem 5.15. 3

Theorem 5.19. Let ' W A ! Œ0; 1/ be a function satisfying (5.22). Let f W A ! A be a mapping satisfying (5.14) and (5.24). Then the mapping f W A ! A is a C ternary derivation.


217

Proof. The proof is similar to the proof of Theorem 5.16. Remark 5.20. Theorem 5.12 follows from Theorems 5.18 and 5.19 by taking '.x; y; z/ D .kxk3r C kyk3r C kzk3r / for all x; y; z 2 A.

5.2.7 JB -Triple Homomorphisms: The Fixed Point Method Now, we prove the superstability of JB -triple homomorphisms by using the fixed point method. Theorem 5.21. Let ' W J 3 ! Œ0; 1/ be a function such that there exists ˛ < 1 with x y z ; ; (5.25) '.x; y; z/ 8˛' 2 2 2 for all x; y; z 2 J . Let f W J ! L be a mapping satisfying (5.15) and kf .fxyzg/ ff .x/f .y/f .z/gkL '.x; y; z/

(5.26)

for all x; y; z 2 J : Then the mapping f W J ! L is a JB -triple homomorphism. Proof. By the same reasoning as in the proof of Theorem 5.13, one can show that the mapping f W J ! L is C-linear. It follows from (5.25) that 1 '.2n x; 2n y; 2n z/ D 0 n!1 8n lim

(5.27)

for all x; y; z 2 J . Since f W J ! L is additive, it follows from (5.26) and (5.27) that f .Œx; y; z/ D Œf .x/; f .y/; f .z/ for all x; y; z 2 J . Thus the mapping f W J ! L is a JB -triple homomorphism. This completes the proof. Theorem 5.22. Let ' W J 3 ! Œ0; 1/ be a function such that there exists an ˛ < 1 with '.x; y; z/

˛ ' .2x; 2y; 2z/ 2

(5.28)

for all x; y; z 2 J . Let f W J ! L be a mapping satisfying (5.15) and (5.26). Then the mapping f W J ! L is a JB -triple homomorphism.


218

Proof. By the same reasoning as in the proof of Theorem 5.13, one can show that the mapping f W J ! L is C-linear. It follows from (5.28) that lim 2n '

n!1

x y z D0 ; ; 2n 2n 2n

(5.29)

for all x; y; z 2 J . Since f W J ! L is additive, it follows from (5.26) and (5.29) that f .Œx; y; z/ D Œf .x/; f .y/; f .z/ for all x; y; z 2 J . Thus the mapping f W J ! L is a C -ternary homomorphism. This completes the proof. Remark 5.23. Theorem 5.13 follows from Theorems 5.21 and 5.22 by taking '.x; y; z/ D .kxk3r C kyk3r C kzk3r / for all x; y; z 2 J .

5.2.8 JB -Triple Derivations: Fixed Point Method Now, we prove the superstability of JB -triple derivations by using the fixed point method. Theorem 5.24. Let ' W J 3 ! Œ0; 1/ be a function satisfying (5.25). Let f W J ! J be a mapping satisfying (5.17) and kf .fxyzg/ ff .x/yzg fxf .y/zg fxyf .z/gkJ '.x; y; z/

(5.30)

for all x; y; z 2 J : Then the mapping f W J ! J is a JB -triple derivation. Proof. The proof is similar to the proof of Theorem 5.21.

Theorem 5.25. Let ' W J 3 ! Œ0; 1/ be a function satisfying (5.28). Let f W J ! J be a mapping satisfying (5.17) and (5.30). Then the mapping f W J ! J is a JB -triple derivation. Proof. The proof is similar to the proof of Theorem 5.22. Remark 5.26. Theorem 5.14 follows from Theorems 5.24 and 5.25 by taking '.x; y; z/ D .kxk3r C kyk3r C kzk3r / for all x; y; z 2 J .

5.3 Bi- -Derivations in JB -Triples

219

5.3 Bi--Derivations in JB -Triples In this section, we prove the Hyers-Ulam stability of bi--derivations in JB triples (see [237]). Definition 5.27 ([97]). Let J be a complex JB -triple and W J ! J be a C-linear mapping. A C-bilinear mapping D W J J ! J is called a bi--derivation on J if D.fx; y; zg; w/ D fD.x; w/; .y/; .z/g C f.x/; D.y; w/; .z/g Cf.x/; .y/; D.z; w/g and D.x; fy; z; wg/ D fD.x; y/; .z/; .w/g C f.y/; D.x; z/; .w/g Cf.y/; .z/; D.x; w/g for all x; y; z; w 2 J. The w-variable of the left side in the first equality is C-linear and the x-variable of the left side in the second equality is C-linear. But the w-variable of the right side in the first equality is not C-linear and the x-variable of the right side in the second equality is not C-linear. Thus we correct the definition of bi--derivation as follows: Definition 5.28. Let J be a complex JB -triple and W J ! J be a C-linear mapping. A C-bilinear mapping D W J J ! J is called a bi--derivation on J if D.fx; y; zg; w/ D fD.x; w/; .y/; .z/g C f.x/; D.y; w /; .z/g Cf.x/; .y/; D.z; w/g and D.x; fy; z; wg/ D fD.x; y/; .z/; .w/g C f.y/; D.x ; z/; .w/g Cf.y/; .z/; D.x; w/g for all x; y; z; w 2 J. Under the conditions of [97, Theorem 2.5], we can easily show that the mapping D W J J ! J must be zero. In particular, if f is bi-additive, then D must be zero. In this section, we correct the statements of the results, and prove the corrected theorems and corollaries. Throughout this section, assume that J is a JB -triple with the norm k k: For any mapping f W J J ! J, we define


220

E; f .x; y; z; w/ D f .x C y; z w/ Cf .x y; z C w/ 2 f .x; z/ C 2 f .y; w/ for all x; y; z; w 2 J and ; 2 T1 WD f 2 C W jj D 1g. Lemma 5.29 ([21]). Let f W J J ! J be a mapping satisfying E; f .x; y; z; w/ D 0 for all x; y; z; w 2 J and ; 2 T1 . Then the mapping f W J J ! J is C-bilinear. Now, we prove the Hyers-Ulam stability of bi--derivations on JB -triples. Theorem 5.30. Let p; " be positive real numbers with p < 1 and f W J J ! J with f .0; 0/ D 0, h W J ! J with h.0/ D 0 be the mappings such that kE; f .x; y; z; w/ C h. a C b/ h.a/ h.b/k ".kxkp C kykp C kzkp C kwkp C kakp C kbkp /

(5.31)

and kf .fx; y; zg; w/ ff .x; w/; h.y/; h.z/g fh.x/; f .y; w /; h.z/g fh.x/; h.y/; f .z; w/gk Ckf .x; fy; z; wg/ ff .x; y/; h.z/; h.w/g fh.y/; f .x ; z/; h.w/g fh.y/; h.z/; f .x; w/gk ".kxkp C kykp C kzkp C kwkp /

(5.32)

for all ; 2 T1 and x; y; z; w 2 J: If the mapping f W J J ! J satisfies the following: lim

n!1

1 1 f .2n x; 2n y/ D lim f .2n x; 8n y/ n n!1 4 16n 1 D lim f .8n x; 2n y/; n!1 16n

(5.33)

then there exist a unique C-linear mapping W J ! J and a unique bi--derivation D W J J ! J such that kh.a/ .a/k

2" kakp 2 2p

(5.34)


221

and kf .x; z/ D.x; z/k

5" .kxkp C kzkp / 4 2p

(5.35)

for all a; x; z 2 J: Proof. Letting x D y D z D w D 0 in (5.31), we have kh. a C b/ h.a/ h.b/k ".kakp C kbkp / for all a; b 2 J. By the same reasoning as in the proof of Park [227, Theorem 2.1], one can show that there exists a unique C-linear mapping W J ! J satisfying (5.34) and the mapping W J ! J is given by .a/ WD lim

n!1

1 h.2n a/ 2n

for all a 2 J. Letting D D 1, a D b D 0, y D x and w D z in (5.31), we have kf .2x; 2z/ 2f .x; z/ C 2f .x; z/k 2".kxkp C kzkp /

(5.36)

for all x; z 2 J. Letting D D 1, a D b D 0, y D x and w D z in (5.31), we have kf .2x; 2z/ 2f .x; z/ C 2f .x; z/k 2".kxkp C kzkp /

(5.37)

for all x; z 2 J. Letting D D 1, a D b D 0, x D z D 0 in (5.31), we have kf .y; w/ C f .y; w/ C 2f .y; w/k ".kykp C kwkp /

(5.38)

for all y; w 2 J. Replacing y; w by x; z in (5.38), respectively, we have kf .x; z/ C f .x; z/ C 2f .x; z/k ".kxkp C kzkp /

(5.39)

for all x; z 2 J. By (5.36) and (5.39), we obtain kf .2x; 2z/ 4f .x; z/ C f .x; z/ f .x; z/k 3".kxkp C kzkp /

(5.40)

for all x; z 2 J. By (5.36) and (5.37), we obtain kf .x; z/ f .x; z/k 2".kxkp C kzkp /

(5.41)

222


for all x; z 2 J. By (5.40) and (5.41), we obtain kf .2x; 2z/ 4f .x; z/k 5".kxkp C kzkp /

(5.42)

for all x; z 2 J. It follows from (5.42) that m1 1 X 5" 2pj f .2l x; 2l z/ 1 f .2m ; 2m z/ .kxkp C kzkp / 4l j 4m 4 4 jDl

(5.43)

˚for1 alln x; z n 2 J and m; l 1 with m > l. This implies that the sequence f .2 x; 2 z/ is a Cauchy sequence for all x; z 2 J: Since J is complete, the 4n ˚ sequence 41n f .2n x; 2n z/ converges and so one can define the mapping D W J J ! J by 1 f .2n x; 2n z/ n!1 4n

D.x; z/ WD lim

for all x; z 2 J: Moreover, letting l D 0 and passing the limit m ! 1 in (5.43), we have (5.35). Let a D b D 0 in (5.31). Then, by the definition of the mapping D, we have 1 kE; f .2n x; 2n y; 2n z; 2n w/k 4n 2pn lim n ".kxkp C kykp C kzkp C kwkp / n!1 4 D0

kE; D.x; y; z; w/k D lim

n!1

for all ; 2 T1 and all x; y; z; w 2 J: By Lemma 5.29, the mapping D W J J ! J is C-bilinear. It follows from (5.32) and (5.33) that kD.fx; y; zg; w/ fD.x; w/; .y/; .z/g f.x/; D.y; w /; .z/g f.x/; .y/; D.z; w/gk C kD.x; fy; z; wg/ fD.x; y/; .z/; .w/g f.y/; D.x ; z/; .w/g f.y/; .z/; D.x; w/gk 1 n1 o 1 1 D lim 4n f .8n fx; y; zg; 2n w/ n f .2n x; 2n w/; n h.2n y/; n h.2n z/ n!1 2 4 2 2 o n1 1 1 n h.2n x/; n f .2n y; 2n w /; n h.2n z/ 2 4 2 n1 o 1 1 n h.2n x/; n h.2n y/; n f .2n z; 2n w/ 2 2 4 1 n1 o 1 1 C 4n f .2n x; 8n fy; z; wg/ n f .2n x; 2n y/; n h.2n z/; n h.2n w/ 2 4 2 2


223

n1 o 1 1 n n n n h.2 y/; f .2 x ; 2 z/; h.2 w/ 2n 4n 2n n1 o 1 1 n h.2n y/; n h.2n z/; n f .2n x; 2n w/ 2 2 4 2pn " lim 4n .kxkp C kykp C kzkp C kwkp / n!1 2

D0 for all x; y; z; w 2 J and so D.fx; y; zg; w/ D fD.x; w/; .y/; .z/g Cf.x/; D.y; w /; .z/g C f.x/; .y/; D.z; w/g and D.x; fy; z; wg/ D fD.x; y/; .z/; .w/g Cf.y/; D.x ; z/; .w/g C f.y/; .z/; D.x; w/g for all x; y; z; w 2 J: Let T W J J ! J be another C-bilinear mapping satisfying (5.35). Then we have kD.x; z/ T.x; z/k 1 kD.2n x; 2n z/ T.2n x; 2n z/k 4n 1 1 n kD.2n x; 2n z/ f .2n x; 2n z/k C n kf .2n x; 2n z/ T.2n x; 2n z/k 4 4 2pn 5" 2 n .kxkp C kzkp /; 4 4 2p

D

which tends to zero as n ! 1 for all x; z 2 J. This proves the uniqueness of D. Therefore, the mapping D W J J ! J is a unique bi--derivation satisfying (5.35). This completes the proof. Similarly, one can obtain the following theorem. Remark 5.31. Let p; " be positive real numbers with p > 4 and f W J J ! J with f .0; 0/ D 0, h W J ! J with h.0/ D 0 be the mappings satisfying (5.31), (5.32) and (5.33). Then there exist a unique C-linear mapping W J ! J and a unique bi--derivation D W J J ! J such that kh.a/ .a/k

2" kakp 2

2p


224

and kf .x; z/ D.x; z/k

5" .kxkp C kzkp / 2p 4

for all a; x; z 2 J: Theorem 5.32. Let p; "; ı be nonnegative real numbers with 0 < p < 1 and f W J J ! J with f .0; 0/ D 0, h W J ! J with h.0/ D 0 be the mappings satisfying (5.33) and kE; f .x; y; z; w/ C h. a C b/ h.a/ h.b/k "kxkp kykp kzkp kwkp kakp kbkp C ı

(5.44)

and kf .fx; y; zg; w/ ff .x; w/; h.y/; h.z/g fh.x/; f .y; w /; h.z/g fh.x/; h.y/; f .z; w/gk Ckf .x; fy; z; wg/ ff .x; y/; h.z/; h.w/g fh.y/; f .x ; z/; h.w/g fh.y/; h.z/; f .x; w/gk "kxkp kykp kzkp kwkp C ı

(5.45)

for all ; 2 T1 and x; y; z; w 2 J: Then there exist a unique C-linear mapping W J ! J and a unique bi--derivation D W J J ! J such that kh.a/ .a/k ı

(5.46)

kf .x; z/ D.x; z/k ı

(5.47)

and

for all a; x; z 2 J: Proof. Letting x D y D z D w D 0 in (5.44), we have kh. a C b/ h.a/ h.b/k ı for all a; b 2 J. By the same reasoning as in the proof of Park [227, Theorem 2.1], one can show that there exists a unique C-linear mapping W J ! J satisfying (5.46) and the mapping W J ! J is given by .a/ WD lim

n!1

1 h.2n a/ 2n


225

for all a 2 J. Letting D D 1, a D b D 0, y D x and w D z in (5.44), we have kf .2x; 2z/ 2f .x; z/ C 2f .x; z/k ı

(5.48)

for all x; z 2 J. Letting D D 1, a D b D 0, y D x and w D z in (5.44), we have kf .2x; 2z/ 2f .x; z/ C 2f .x; z/k ı

(5.49)

for all x; z 2 J. Letting D D 1, a D b D 0, x D z D 0 in (5.44), we have kf .y; w/ C f .y; w/ C 2f .y; w/k ı

(5.50)

for all y; w 2 J. Replacing y; w by x; z in (5.50), respectively, we have kf .x; z/ C f .x; z/ C 2f .x; z/k ı

(5.51)

for all x; z 2 J. By (5.48) and (5.51), we obtain kf .2x; 2z/ 4f .x; z/ C f .x; z/ f .x; z/k 2ı

(5.52)

for all x; z 2 J. By (5.48) and (5.49), we obtain kf .x; z/ f .x; z/k ı

(5.53)

for all x; z 2 J. By (5.52) and (5.53), we obtain kf .2x; 2z/ 4f .x; z/k 3ı

(5.54)

for all x; z 2 J. It follows from (5.54) that m1 X 3ı 1 1 f .2l x; 2l z/ 1 f .2m ; 2m z/ 4l m 4 4 4j jDl

(5.55)

˚for1 alln x; z n 2 J and m; l 1 with m > l. This implies that the sequence f .2 x; 2 z/ is a Cauchy sequence for all x; z 2 J: Since J is complete, the 4n ˚ sequence 41n f .2n x; 2n z/ converges and so one can define the mapping D W JJ ! J by 1 f .2n x; 2n z/ n!1 4n

D.x; z/ WD lim

for all x; z 2 J: Moreover, letting l D 0 and passing the limit m ! 1 in (5.55), we have (5.47).


226

Let a D b D 0 in (5.44). Then by the definition of the mapping D, we have 1 kE; f .2n x; 2n y; 2n z; 2n w/k n!1 4n ı lim n n!1 4 D0

kE; D.x; y; z; w/k D lim

for all ; 2 T1 and x; y; z; w 2 J: By Lemma 5.29, the mapping D W J J ! J is C-bilinear. It follows from (5.33) and (5.45) that kD.fx; y; zg; w/ fD.x; w/; .y/; .z/g f.x/; D.y; w /; .z/g f.x/; .y/; D.z; w/gk C kD.x; fy; z; wg/ fD.x; y/; .z/; .w/g f.y/; D.x ; z/; .w/g f.y/; .z/; D.x; w/gk 1 n1 o 1 1 D lim 4n f .8n fx; y; zg; 2n w/ n f .2n x; 2n w/; n h.2n y/; n h.2n z/ n!1 2 4 2 2 o n1 1 1 n h.2n x/; n f .2n y; 2n w /; n h.2n z/ 2 4 2 n1 o 1 1 n h.2n x/; n h.2n y/; n f .2n z; 2n w/ 2 2 4 1 n1 o 1 1 C 4n f .2n x; 8n fy; z; wg/ n f .2n x; 2n y/; n h.2n z/; n h.2n w/ 2 4 2 2 n1 o 1 1 n h.2n y/; n f .2n x ; 2n z/; n h.2n w/ 2 4 2 n1 o 1 1 n h.2n y/; n h.2n z/; n f .2n x; 2n w/ 2 2 4 24pn " ı kxkp kykp kzkp kwkp C 4n lim 4n n!1 2 2 D0 for all x; y; z; w 2 J and so D.fx; y; zg; w/ D fD.x; w/; .y/; .z/g C f.x/; D.y; w /; .z/g C f.x/; .y/; D.z; w/g and D.x; fy; z; wg/ D fD.x; y/; .z/; .w/g C f.y/; D.x ; z/; .w/g C f.y/; .z/; D.x; w/g for all x; y; z; w 2 J:


227

Let T W J J ! J be another C-bilinear mapping satisfying (5.47). Then we have kD.x; z/ T.x; z/k 1 kD.2n x; 2n z/ T.2n x; 2n z/k 4n 1 1 n kD.2n x; 2n z/ f .2n x; 2n z/k C n kf .2n x; 2n z/ T.2n x; 2n z/k 4 4 2ı n; 4

D

which tends to zero as n ! 1 for all x; z 2 J. This proves the uniqueness of D. Therefore, the mapping D W J J ! J is a unique bi--derivation satisfying (5.47). This completes the proof. Theorem 5.33. Let p; " be positive real numbers with p ¤ 1 and f W J J ! J with f .0; 0/ D 0, h W J ! J with h.0/ D 0 be the mappings such that kE; f .x; y; z; w/ C h. a C b/ h.a/ h.b/k "kxkp kykp kzkp kwkp kakp kbkp

(5.56)

and kf .fx; y; zg; w/ ff .x; w/; h.y/; h.z/g fh.x/; f .y; w /; h.z/g fh.x/; h.y/; f .z; w/gk Ckf .x; fy; z; wg/ ff .x; y/; h.z/; h.w/g fh.y/; f .x ; z/; h.w/g fh.y/; h.z/; f .x; w/gk "kxkp kykp kzkp kwkp

(5.57)

for all ; 2 T1 and x; y; z; w 2 J: Then the mapping h W J ! J is a C-linear mapping and the mapping f W J J ! J is a bi-h-derivation. Proof. Letting x D y D z D w D 0 in (5.56), we have kh. a C b/ h.a/ h.b/k 0 for all a; b 2 J. By the same reasoning as in the proof of [227, Theorem 2.1], one can show that the mapping h W J ! J is a C-linear mapping. Letting a D b D 0 in (5.56), we have kE; f .x; y; z; w/k D 0


228

for all ; 2 T1 and x; y; z; w 2 J: By Lemma 5.29, the mapping f W J J ! J is C-bilinear. For the case p < 1, it follows from (5.57) that kf .fx; y; zg; w/ ff .x; w/; h.y/; h.z/g fh.x/; f .y; w /; h.z/g fh.x/; h.y/; f .z; w/gk C kf .x; fy; z; wg/ ff .x; y/; h.z/; h.w/g fh.y/; f .x ; z/; h.w/g fh.y/; h.z/; f .x; w/gk 1 n1 o 1 1 D lim 4n f .8n fx; y; zg; 2n w/ n f .2n x; 2n w/; n h.2n y/; n h.2n z/ n!1 2 4 2 2 o n1 1 1 n h.2n x/; n f .2n y; 2n w /; n h.2n z/ 2 4 2 n1 o 1 1 n h.2n x/; n h.2n y/; n f .2n z; 2n w/ 2 2 4 1 n1 o 1 1 C 4n f .2n x; 8n fy; z; wg/ n f .2n x; 2n y/; n h.2n z/; n h.2n w/ 2 4 2 2 n1 o 1 1 n h.2n y/; n f .2n x ; 2n z/; n h.2n w/ 2 4 2 n1 o 1 1 n h.2n y/; n h.2n z/; n f .2n x; 2n w/ 2 2 4 24pn " kxkp kykp kzkp kwkp n!1 24n

lim D0

for all x; y; z; w 2 J and so the mapping f W J J ! J is a bi-h-derivation. Similarly, for the case p > 1, one can show that the mapping f W J J ! J is a bi-h-derivation. Therefore, the mapping h W J ! J is a C-linear mapping and the mapping f W J J ! J is a bi-h-derivation. This completes the proof.

Chapter 6

Stability of Functional Equations in Multi-Banach Algebras

In this chapter, we extend some results from last chapters in multi-Banach algebras (see [7, 91, 218, 252]). In Sect. 6.1, we consider the stability of the m-variable additive functional equation: m m m m X X X X f mxi C xj C f xi D 2f mxi iD1

jD1; j¤i

iD1

iD1

for each m 2, which was presented at Sect. 2.2 of Chap. 2 and, by the fixed point method, we approximate homomorphisms and derivations in multi-Banach algebras. In Sect. 6.2, by using the fixed point method, we prove the Hyers-Ulam stability of homomorphisms in multi-C -ternary algebras and derivations on multi-C ternary algebras for the additive functional equation: m m m m X X X X f mxi C xj C f xi D 2f mxi iD1

jD1; j¤i

iD1

iD1

for each m 2. In Sect. 6.3, we consider the functional equation (3.97) presented at Sect. 3.5 of Chap. 3 and we use a fixed point method to prove the Hyers-Ulam stability of the functional equation (3.97) in multi-Banach modules over a unital multi-C -algebra.


229

230

6 Stability of Functional Equations in Multi-Banach Algebras

As an application, we show that every almost linear bijection h W A ! B of a unital multi-C -algebra A onto a unital multi-C -algebra B is a C -algebra isomorphism when 2n 2n h n uy D h n u h.y/ r r for all unitaries u 2 U.A/, y 2 A and n 0. In Sect. 6.4, we approximate the following additive functional inequality: dC1 dC1 X X f .x1i /; ; f .xki // . iD1

k

iD1

PdC1 x PdC1 x iD1 1i iD1 ki ; ; mf mf k m m for all x11 ; ; xk dC1 2 X where d 2 is a fixed integer. Also, we investigate homomorphisms in proper multi-CQ -algebras and derivations on proper multiCQ -algebras associated with the above additive functional inequality. In Sect. 6.5, by using the fixed point method, we prove the Hyers-Ulam stability of homomorphisms and derivations on multi-C –ternary algebras for the additive functional equation: 2f

Pp

jD1 xj

2

C

p d d X X X yj D f .xj / C 2 f .yj /: jD1

jD1

jD1

6.1 Stability of m-Variable Additive Mappings For any mapping f W A ! B, we define D f .x1 ; ; xm / WD

m X iD1

m m m X X X f mxi C xj C f xi 2f mxi jD1; j¤i

iD1

iD1

for all 2 T1 WD f 2 C W j j D 1g and x1 ; ; xm 2 A.

6.1.1 Stability of Homomorphisms in Multi-Banach Algebras Now, we prove the Hyers-Ulam stability of homomorphisms in multi-Banach algebras for the functional equation D f .x1 ; ; xm / D 0.

6.1 Stability of m-Variable Additive Mappings

231

Theorem 6.1. Let ..Bk ; k kk / W k 1/ be a multi-Banach algebra. Let f W A ! B be a mapping for which there exists the functions ' W Amk ! Œ0; 1/ and W A2k ! Œ0; 1/ such that lim mj '.mj x11 ; ; mj x1m ; ; mj xk1 ; ; mj xkm / D 0;

j!1

(6.1)

k.D f .x11 ; ; x1m /; ; D f .xk1 ; ; xkm //kk '.x11 ; ; x1m ; ; xk1 ; ; xkm /;

(6.2)

k.f .x1 y1 / f .x1 /f .y1 /; ; f .xk yk / f .xk /f .yk //kk

.x1 ; y1 ; ; xk ; yk /

(6.3)

and lim m2j .mj x1 ; mj y1 ; ; mj xk ; mj yk / D 0

(6.4)

j!1

for all 2 T1 and x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xm ; x; y1 ; ; yk 2 A. If there exists L < 1 such that m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ ' mx11 ; 0; ; 0; mx21 ; 0; ; 0; ; mxk1 ; 0; ; 0 m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ mL' x11 ; 0; ; 0; x21 ; 0; ; 0; ; xk1 ; 0; ; 0

for all x11 ; x21 ; ; xk1 2 A, then there exists a unique homomorphism H W A ! B such that k.f .x1 / H.x1 /; ; f .xk / H.xk //kk m

m

m

‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ 1 ' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0 m mL

(6.5)

for all x1 ; ; xk 2 A. Proof. Consider the set X WD fg W A ! Bg and introduce the generalized metric on X as follows: d.g; h/ D inffC 2 RC W k.g.x1 / h.x1 /; ; g.xk / h.xk //kk m

m

m

‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ C'.x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0/; 8x1 ; ; xk 2 Ag; which .X; d/ is complete.

232


Now, we consider the linear mapping J W X ! X such that Jg.x/ WD

1 g.mx/ m

for all x 2 A. Now, we have d.Jg; Jh/ Ld.g; h/ for all g; h 2 X. Letting D 1, xi1 D x1 and xi2 D D xim D 0, 1 i k, in (6.2), we have k.f .mx1 / mf .x1 /; ; f .mxk / mf .xk //kk m

m

m

‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ ' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0

(6.6)

for all x1 ; ; xk 2 A and so k.f .x1 /

1 1 f .mx1 /; ; f .xk / f .mxk //kk m m m

m

m

‚ …„ ƒ 1 ‚ …„ ƒ ‚ …„ ƒ ' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0 m for all x1 ; ; xk 2 A. Hence d.f ; Jf / H W A ! B such that

1 m.



(6.7)

for all x 2 A. The mapping H is a unique fixed point of J in the set Y D fg 2 X W d.f ; g/ < 1g: This implies that H is a unique mapping satisfying (6.7) such that there exists C 2 .0; 1/ satisfying k.H.x1 / f .x1 /; ; H.xk / f .xk //kk m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ C' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0

for all x1 ; ; xk 2 A;


233

(2) d.J n f ; H/ ! 0 as n ! 1. This implies the equality f .mn x/ D H.x/ n!1 mn

(6.8)

lim

for all x 2 A; 1 d.f ; Jf /, which implies the inequality (3) d.f ; H/ 1L 1 : m mL

d.f ; H/

This implies that the inequality (6.4) holds. It follows from (6.1), (6.2) and (6.8) that m m m m X X X X H mx1i C x1j C H x1i 2H mx1i ; iD1

iD1

jD1; j¤i

;

iD1

m m m m X X X X H mxki C xkj C H xki 2H mxki iD1

iD1

jD1; j¤i

iD1

k

m m X 1 X nC1 f m x C mn x1j 1i n n!1 m iD1

D lim

jD1; j¤i

Cf

m X iD1

;

m X iD1

Cf

m X iD1

lim

n!1

mn x1i 2f

m X

mnC1 x1i ;

iD1 m X f mnC1 xki C mn xkj jD1; j¤i m X mn xki 2f mnC1 xki

k

iD1

1 '.mn x11 ; ; mn x1m ; ; mn xk1 ; ; mn xkm / mn

D0 for all x11 ; ; x1m ; ; xk1 ; ; xkm 2 A and so m m m m X X X X H mxi C xj C H xi D 2H mxi iD1

jD1; j¤i

iD1

iD1

for all x1 ; ; xm 2 A. So H is additive. By a similar method to above, we get H.mx/ D H.m x/

(6.9)

234


for all 2 T1 and x 2 A. Thus one can show that the mapping H W A ! B is C-linear. It follows from (6.3), (6.4) and (6.8) that k.H.x1 y1 / H.x1 /H.y1 /; ; H.xk yk / H.xk /H.yk //kk D lim

n!1

1 k.f .m2n x1 y1 / f .mn x1 /f .mn y1 /; m2n

; f .m2n xk yk / f .mn xk /f .mn yk //kk lim

n!1

1 .mn x1 ; mn y1 ; ; mn xk ; mn yk / m2n

D0 for all x1 ; y1 ; ; xk ; yk 2 A and so H.xy/ D H.x/H.y/ for all x; y 2 A. Thus H W A ! B is a homomorphism satisfying (6.5). This completes the proof. Corollary 6.2. Let ..Bk ; k kk / W k 1/ be a multi-Banach algebra. Let r < 1 and be nonnegative real numbers and f W A ! B be a mapping such that k.D f .x11 ; ; x1m /; ; D f .xk1 ; ; xkm //kk

m X jD1

kx1j krA C C

m X

kxkm krA

(6.10)

jD1

and k.f .x1 y1 / f .x1 /f .y1 /; ; f .xk yk / f .xk /f .yk //kk kx1 krA ky1 krA C C kxk krA kyk krA

(6.11)

for all 2 T1 and x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xm ; x; y1 ; ; yk 2 A. Then there exists a unique homomorphism H W A ! B such that k.f .x1 / H.x1 /; ; f .xk / H.xk //kk r r r r kx k C ky k C C kx k C ky k 1 1 k k A A A A m mr for all x1 ; ; xk 2 A.


235

Proof. The proof follows from Theorem 6.1 by taking '.x11 ; ; x1m ; ; xk1 ; ; xkm / WD

m X

kx1j krA C C

m X

jD1

kxkm krA

jD1

and .x1 ; y1 ; ; xk ; yk / WD kx1 krA ky1 krA C C kxk krA kyk krA for all x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xm ; x; y1 ; ; yk 2 A and L D mr1 .

Theorem 6.3. Let ..Bk ; kkk / W k 1/ be a multi-Banach algebra. Let f W A ! B be a mapping for which there are functions ' W Amk ! Œ0; 1/ and W A2k ! Œ0; 1/ such that lim mj '.mj x11 ; ; mj x1m ; ; mj xk1 ; ; mj xkm / D 0;

j!1

(6.12)


(6.13)

k.f .x1 y1 / f .x1 /f .y1 /; ; f .xk yk / f .xk /f .yk //kk

.x1 ; y1 ; ; xk ; yk /

(6.14)


j!1

(6.15)

for all 2 T1 and x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xm ; x; y1 ; ; yk 2 A. If there exists L < 1 such that m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ ' x11 ; 0; ; 0; x21 ; 0; ; 0; ; xk1 ; 0; ; 0 m

m

m

…„ ƒ ‚ …„ ƒ ‚ …„ ƒ L ‚ ' mx11 ; 0; ; 0; mx21 ; 0; ; 0; ; mxk1 ; 0; ; 0 m for all x11 ; x21 ; ; xk1 2 A, then there exists a unique homomorphism H W A ! B such that k.f .x1 / H.x1 /; ; f .xk / H.xk //kk m

m

m

‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ L ' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0 m mL for all x1 ; ; xk 2 A.

(6.16)

236


Proof. We consider the linear mapping J W X ! X such that Jg.x/ WD mg

x m

for all x 2 A. It follows from (6.6) that x x 1 k ; ; f .xk / mf f .x1 / mf k m m ‚ x

m

m

…„ ƒ ‚ …„ ƒ x2 1 ; 0; ; 0; ; 0; ; 0; ' m m m

m

‚ …„ ƒ xk ; ; 0; ; 0 m

m

(6.17)

m

‚ …„ ƒ L ‚ …„ ƒ ‚ …„ ƒ ' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0 m for all x1 ; ; xk 2 A. Hence we have d.f ; Jf /

L : m

By Theorem 1.3, there exists a mapping H W A ! B such that (1) H is a fixed point of J, i.e., H.mx/ D mH.x/

(6.18)

for all x 2 A. The mapping H is a unique fixed point of J in the set Y D fg 2 X W d.f ; g/ < 1g: This implies that H is a unique mapping satisfying (6.18) such that there exists C 2 .0; 1/ satisfying k.H.x1 / f .x1 /; ; H.xk / f .xk //kk m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ C' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0

for all x1 ; ; xk 2 A; (2) d.J n f ; H/ ! 0 as n ! 1. This implies the equality lim mn f

n!1

for all x 2 A;

x D H.x/ mn


(3) d.f ; H/

1 1L d.f ; Jf /,

237

which implies the inequality d.f ; H/

L ; m mL

which implies that the inequality (6.16) holds. The rest of the proof is similar to the proof of Theorem 6.1. This completes the proof. Corollary 6.4. Let ..Bk ; k kk / W k 1/ be a multi-Banach algebra. Let r > 1, be nonnegative real numbers and f W A ! B be a mapping such that k.D f .x11 ; ; x1m /; ; D f .xk1 ; ; xkm //kk

m X

kx1j krA C C

jD1

m X

kxkm krA

(6.19)

jD1

and k.f .x1 y1 / f .x1 /f .y1 /; ; f .xk yk / f .xk /f .yk //kk kx1 krA ky1 krA C C kxk krA kyk krA

(6.20)

for all 2 T1 and x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xm ; x; y1 ; ; yk 2 A. Then there exists a unique homomorphism H W A ! B such that k.f .x1 / H.x1 /; ; f .xk / H.xk //kk

mr

kx1 krA C ky1 krA C C kxk krA C kyk krA m

for all x1 ; ; xk 2 A. Proof. The proof follows from Theorem 6.3 by taking '.x11 ; ; x1m ; ; xk1 ; ; xkm / WD

m X jD1

kx1j krA C C

m X

kxkm krA ;

jD1

.x1 ; y1 ; ; xk ; yk / WD kx1 krA ky1 krA C C kxk krA kyk krA for all x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xm ; x; y1 ; ; yk 2 A and L D m1r .

238


6.1.2 Stability of Derivations in Multi-Banach Algebras Now, we prove the Hyers-Ulam stability of derivations in multi-Banach algebras for the following functional equation: D f .x1 ; ; xm / D 0 for all 2 T1 WD f 2 C W j j D 1g and x1 ; ; xm 2 A. Theorem 6.5. Let ..Ak ; k kk / W k 1/ be a multi-Banach algebra. Let f W A ! A be a mapping for which there exist the functions ' W Amk ! Œ0; 1/ and W A2k ! Œ0; 1/ such that lim mj '.mj x11 ; ; mj x1m ; ; mj xk1 ; ; mj xkm / D 0;

j!1

(6.21)


(6.22)

k.f .x1 y1 / f .x1 /y1 x1 f .y1 /; ; f .xk yk / f .xk /yk xk f .yk //kk

.x1 ; y1 ; ; xk ; yk /

(6.23)


(6.24)

j!1

for all 2 T1 and x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xm ; x; y1 ; ; yk 2 A. If there exists L < 1 such that m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ ' mx11 ; 0; ; 0; mx21 ; 0; ; 0; ; mxk1 ; 0; ; 0 m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ mL' x11 ; 0; ; 0; x21 ; 0; ; 0; ; xk1 ; 0; ; 0

for all x11 ; x21 ; ; xk1 2 A, then there exists a unique derivation ı W A ! A such that k.f .x1 / ı.x1 /; ; f .xk / ı.xk //kk m

m

m

‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ 1 ' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0 m mL for all x1 ; ; xk 2 A.

(6.25)


239

Proof. By the same reasoning as in the proof of Theorem 6.1, there exists a unique C-linear mapping ı W A ! A satisfying (6.24). The mapping ı W A ! A is given by ı.x/ D lim

n!1

f .mn x/ mn

(6.26)

for all x 2 A. It follows from (6.21), (6.24) and (6.26) that k.ı.x1 y1 / ı.x1 /y1 x1 ı.y1 /; ; ı.xk yk / ı.xk /yk xk ı.yk //kk D lim

n!1

1 k.f .m2n x1 y1 / f .mn x1 / mn y1 mn x1 f .mn y1 /; m2n

; f .m2n xk yk / f .mn xk / mn yk mn xk f .mn yk //kk lim m2n .mn x1 ; mn y1 ; ; mn xk ; mn yk / n!1

D0 for all x; y 2 A and so ı.xy/ D ı.x/y C xı.y/ for all x; y 2 A. Thus ı W A ! A is a derivation satisfying (6.23). This completes the proof. Corollary 6.6. Let ..Ak ; k kk / W k 1/ be a multi-Banach algebra. Let r < 1, be nonnegative real numbers and f W A ! A be a mapping such that k.D f .x11 ; ; x1m /; ; D f .xk1 ; ; xkm //kk

m X jD1

kx1j krA C C

m X

kxkj krA

(6.27)

jD1

and k.f .x1 y1 / f .x1 /y1 x1 f .y1 /; ; f .xk yk / f .xk /yk xk f .yk //kk (6.28) kx1 krA ky1 krA C C kxk krA kyk krA for all 2 T1 and x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xm ; x; y1 ; ; yk 2 A. Then there exists a unique derivation ı W A ! A such that k.f .x1 / ı.x1 /; ; f .xk / ı.xk //kk for all x 2 A.

kxk2r A m mr

240


Proof. The proof follows from Theorem 6.5 by taking '.x11 ; ; x1m ; ; xk1 ; ; xkm / WD

m X

kx1j krA C C

m X

jD1

kxkj krA ;

jD1

.x1 ; y1 ; ; xk ; yk / WD kx1 krA ky1 krA C C kxk krA kyk krA for all x; y 2 A and L D mr .

Remark 6.7. Let ..A ; k kk / W k 1/ be a multi-Banach algebra. Let f W A ! A be a mapping for which there exist the functions ' W Amk ! Œ0; 1/ and W A2k ! Œ0; 1/ such that k

lim mj '.mj x11 ; ; mj x1m ; ; mj xk1 ; ; mj xkm / D 0;

j!1

(6.29)


(6.30)

k.f .x1 y1 / f .x1 /y1 x1 f .y1 /; ; f .xk yk / f .xk /yk xk f .yk //kk

.x1 ; y1 ; ; xk ; yk /

(6.31)


j!1

(6.32)

for all 2 T1 and x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xm ; x; y1 ; ; yk 2 A. If there exists L < 1 such that m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ ' mx11 ; 0; ; 0; mx21 ; 0; ; 0; ; mxk1 ; 0; ; 0 m

m

m

‚ …„ ƒ L ‚ …„ ƒ ‚ …„ ƒ ' x11 ; 0; ; 0; x21 ; 0; ; 0; ; xk1 ; 0; ; 0 m for all x11 ; x21 ; ; xk1 2 A, then there exists a unique derivation ı W A ! A such that k.f .x1 / ı.x1 /; ; f .xk / ı.xk //kk m

m

m

‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ L ' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0 m mL for all x1 ; ; xk 2 A.

(6.33)

6.2 Ternary Jordan Homomorphisms and Derivations in Multi-C -Ternary. . .

241

Corollary 6.8. Let ..Ak ; k kk / W k 1/ be a multi-Banach algebra. Let r > 1, be nonnegative real numbers and f W A ! A be a mapping such that k.D f .x11 ; ; x1m /; ; D f .xk1 ; ; xkm //kk 1 0 m m X X @ kx1j krA C C kxkj krA A jD1

jD1

and k.f .x1 y1 / f .x1 /y1 x1 f .y1 /; ; f .xk yk / f .xk /yk xk f .yk //kk kx1 krA ky1 krA C C kxk krA kyk krA for all 2 T1 and x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xm ; x; y1 ; ; yk 2 A. Then there exists a unique derivation ı W A ! A such that k.f .x1 / ı.x1 /; ; f .xk / ı.xk //kk

mr

kxk2r A m

for all x 2 A. Proof. The proof follows from Remark 6.7 by taking '.x11 ; ; x1m ; ; xk1 ; ; xkm / WD

m X

kx1j krA C C

jD1

m X

kxkj krA ;

jD1

.x1 ; y1 ; ; xk ; yk / WD kx1 krA ky1 krA C C kxk krA kyk krA for all x; y 2 A and L D m1r .

6.2 Ternary Jordan Homomorphisms and Derivations in Multi-C -Ternary Algebras In this section, using the fixed point method, we prove the Hyers-Ulam stability of homomorphisms in multi-C -ternary algebras and derivations on multi-C -ternary algebras for the following additive functional equation: m m m m X X X X f mxi C xj C f xi D 2f mxi iD1

jD1; j¤i

iD1

iD1

for each m 2. Throughout this section, assume that A, B are C -ternary algebras.

242


6.2.1 Stability of Homomorphisms in Multi-C -Ternary Algebras For any mapping f W A ! B, we define D f .x1 ; ; xm / WD

m X iD1

m m m X X X f mxi C xj C f xi 2f mxi iD1

jD1; j¤i

iD1

for all 2 T1 WD f 2 C W j j D 1g and x1 ; ; xm 2 A. Using Theorem 1.3, we prove the Hyers-Ulam stability of homomorphisms in multi-C ternary algebras for the functional equation D f .x1 ; ; xm / D 0: Theorem 6.9. Let ..Bk ; k kk / W k 1/ be a multi-C -ternary algebra. Let f W A ! B be a mapping for which there exist the functions ' W Amk ! Œ0; 1/ and W A2k ! Œ0; 1/ such that lim mj '.mj x11 ; ; mj x1m ; ; mj xk1 ; ; mj xkm / D 0;

j!1

(6.34)


(6.35)

k.f .Œx1 ; y1 ; z1 / Œf .x1 /; f .y1 /; f .z1 /; ; f .Œxk ; yk ; zk / Œf .xk /; f .yk /; f .zk //kk

.x1 ; y1 ; z1 ; ; xk ; yk ; zk /

(6.36)

and lim m3j .mj x1 ; mj y1 ; mj z1 ; ; mj xk ; mj yk ; mj zk / D 0

j!1

for all 2 T1 and x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xk ; y1 ; ; yk ; z1 ; ; zk 2 A. If there exists L < 1 such that m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ ' mx11 ; 0; ; 0; mx21 ; 0; ; 0; ; mxk1 ; 0; ; 0 m

m

m

‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ mL' x11 ; 0; ; 0; x21 ; 0; ; 0; ; xk1 ; 0; ; 0

(6.37)


243

for all x11 ; x21 ; ; xk1 2 A, then there exists a unique homomorphism H W A ! B such that k.f .x1 / H.x1 /; ; f .xk / H.xk //kk m

m

m

‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ 1 ' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0 m mL

(6.38)

for all x1 ; ; xk 2 A. Proof. Consider the set X WD fg W A ! Bg and introduce the generalized metric on X as follows: d.g; h/ D inffC 2 RC W k.g.x1 / h.x1 /; ; g.xk / h.xk //kk m

m

m

‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ C'.x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0/; 8x1 ; ; xk 2 Ag; which .X; d/ is complete. Now, we consider the linear mapping J W X ! X such that Jg.x/ WD

1 g.mx/ m

for all x 2 A. Now, we have d.Jg; Jh/ Ld.g; h/ for all g; h 2 X. Letting D 1, xi1 D xi and xi2 D D xim D 0 (1 i k) in (6.35), we have k.f .mx1 / mf .x1 /; ; f .mxk / mf .xk //kk m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ ' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0

(6.39)

for all x1 ; ; xk 2 A. Thus we have k.f .x1 /

1 1 f .mx1 /; ; f .xk / f .mxk //kk m m m

m

m

‚ …„ ƒ 1 ‚ …„ ƒ ‚ …„ ƒ ' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0 m for all x1 ; ; xk 2 A. Hence d.f ; Jf / H W A ! B such that

1 . m


244



(6.40)

for all x 2 A. The mapping H is a unique fixed point of J in the set Y D fg 2 X W d.f ; g/ < 1g: This implies that H is a unique mapping satisfying (6.40) such that there exists C 2 .0; 1/ satisfying the following: k.H.x1 / f .x1 /; ; H.xk / f .xk //kk m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ C' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0

for all x1 ; ; xk 2 A; (2) d.J n f ; H/ ! 0 as n ! 1. This implies the equality f .mn x/ D H.x/ n!1 mn lim

(6.41)


1 : m mL

This implies that the inequality (6.38) holds. It follows from (6.34), (6.35) and (6.41) that m m m m X X X X H mx1i C x1j C H x1i 2H mx1i /; iD1

;

iD1

jD1; j¤i m X

H mxki C

iD1

m X

xkj C H

iD1

m X

xki 2H

iD1

jD1; j¤i

m m X 1 X nC1 n f m x C m x 1i 1j n!1 mn iD1

D lim

jD1; j¤i

Cf

m X iD1

m X mn x1i 2f mnC1 x1i ; iD1

m X iD1

mxki

k


;

m m X X f mnC1 xki C mn xkj iD1

Cf

m X

jD1; j¤i

m X mn xki 2f mnC1 xki

iD1

lim

n!1

245

iD1

k

1 '.mn x11 ; ; mn x1m ; ; mn xk1 ; ; mn xkm / D 0 mn

for all x11 ; ; x1m ; ; xk1 ; ; xkm 2 A and so m m m m X X X X H mxi C xj C H xi D 2H mxi iD1

jD1; j¤i

iD1

(6.42)

iD1

for all x1 ; ; xm 2 A. Thus H is additive. By the similar method, we have H.mx/ D H.m x/ for all 2 T1 and x 2 A. Thus one can show that the mapping H W A ! B is C-linear. It follows from (6.36), (6.37) and (6.41) that k.H.Œx1 ; y1 ; z1 / ŒH.x1 /; H.y1 /; H.z1 /; ; H.Œxk ; yk ; zk / ŒH.xk /; H.yk /; H.zk //kk 1 k.f .Œmn x1 ; mn y1 ; mn z1 / Œf .mn x1 /; f .mn y1 /; f .mn z1 /; m3n ; f .Œmn xk ; mn yk ; mn zk / Œf .mn xk /; f .mn yk /; f .mn zk //kk

D lim

n!1

lim

n!1

1 .mn x1 ; mn y1 ; ; mn xk ; mn yk / m3n

D0 for all x1 ; y1 ; ; xk ; yk 2 A and so H.Œx; y; z/ D ŒH.x/; H.y/; H.z/ for all x; y 2 A. Thus H W A ! B is a homomorphism satisfying (6.38). This completes the proof. Corollary 6.10. Let ..Bk ; k kk /W k 1/ be a multi-C -ternary algebra. Let r < 1, be nonnegative real numbers and f W A ! B be a mapping such that

246


k.D f .x11 ; ; x1m /; ; D f .xk1 ; ; xkm //kk 0 1 m m X X r r @ kx1j kA C C kxkj kA A jD1

(6.43)

jD1

and k.f .Œx1 ; y1 ; z1 / Œf .x1 /; f .y1 /; f .z1 /; ; f .Œxk ; yk ; zk / Œf .xk /; f .yk /; f .zk //kk kx1 krA ky1 krA kz1 krA C C kxk krA kyk krA kzk krA

(6.44)

for all 2 T1 and x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xk ; y1 ; ; yk ; z1 ; ; zk 2 A. Then there exists a unique homomorphism H W A ! B such that k.f .x1 / H.x1 /; ; f .xk / H.xk //kk

kx1 krA C C kxk krA r mm


m X jD1

kx1j krA C C

m X

kxkj krA ;

jD1

.x1 ; y1 ; z1 ; ; xk ; yk ; zk / WD kx1 krA ky1 krA kz1 krA C C kxk krA kyk krA kzk krA for all x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xk ; y1 ; ; yk ; z1 ; ; zk 2 A and L D mr1 . Theorem 6.11. Let ..Bk ; k kk / W k 1/ be a multi-C -ternary algebra. Let f W A ! B be a mapping for which there exist the functions ' W Amk ! Œ0; 1/ and W A2k ! Œ0; 1/ satisfying the inequalities (6.35) and (6.36) such that lim mj '.mj x11 ; ; mj x1m ; ; mj xk1 ; ; mj xkm / D 0

(6.45)

lim m3j .mj x1 ; mj y1 ; mj z1 ; ; mj xk ; mj yk ; mj zk / D 0

(6.46)

j!1

and j!1


247

for all 2 T1 and x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xm ; y1 ; ; yk ; z1 ; ; zk 2 A. If there exists L < 1 such that m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ ' x11 ; 0; ; 0; x21 ; 0; ; 0; ; xk1 ; 0; ; 0 m

m

m

…„ ƒ ‚ …„ ƒ ‚ …„ ƒ L ‚ ' mx11 ; 0; ; 0; mx21 ; 0; ; 0; ; mxk1 ; 0; ; 0 m for all x11 ; x21 ; ; xk1 2 A, then there exists a unique homomorphism H W A ! B such that k.f .x1 / H.x1 /; ; f .xk / H.xk //kk m

m

m

‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ L ' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0 m mL

(6.47)

for all x1 ; ; xk 2 A. Proof. We consider the linear mapping J W X ! X defined by Jg.x/ WD mg

x m

for all x 2 A. It follows from (6.39) that x x 1 k ; ; f .xk / mf f .x1 / mf k m m ‚ x

m

m

…„ ƒ ‚ …„ ƒ x2 1 ; 0; ; 0; ; 0; ; 0; ' m m m

m

m

‚ …„ ƒ xk ; ; 0; ; 0 m m

‚ …„ ƒ L ‚ …„ ƒ ‚ …„ ƒ ' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0 m

(6.48)

for all x1 ; ; xk 2 A. Hence we have d.f ; Jf /

L : m

By Theorem 1.3, there exists a mapping H W A ! B such that (1) H is a fixed point of J, i.e., H.mx/ D mH.x/ for all x 2 A. The mapping H is a unique fixed point of J in the set

(6.49)

248


Y D fg 2 X W d.f ; g/ < 1g: This implies that H is a unique mapping satisfying (6.49) such that there exists C 2 .0; 1/ satisfying the following: k.H.x1 / f .x1 /; ; H.xk / f .xk //kk m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ C' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0

for all x1 ; ; xk 2 A; (2) d.J n f ; H/ ! 0 as n ! 1. This implies the equality x D H.x/ lim mn f n!1 mn for all x 2 A; 1 (3) d.f ; H/ 1L d.f ; Jf /, which implies the inequality d.f ; H/

L ; m mL

which implies that the inequality (6.47) holds. The rest of the proof is similar to the proof of Theorem 6.9. This completes the proof. Corollary 6.12. Let ..Bk ; k kk / W k 1/ be a multi-C -ternary algebra. Let r > 1, be nonnegative real numbers and f W A ! B be a mapping such that satisfying (6.43) and (6.44). Then there exists a unique homomorphism H W A ! B such that k.f .x1 / H.x1 /; ; f .xk / H.xk //kk

mr

kx1 krA C C kxk krA m


m X jD1

kx1j krA

CC

m X

kxkj krA ;

jD1

.x1 ; y1 ; z1 ; ; xk ; yk ; zk / WD kx1 krA ky1 krA kz1 krA C C kxk krA kyk krA kzk krA for all x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xk ; y1 ; ; yk ; z1 ; ; zk 2 A and L D m1r .


249

6.2.2 Stability of Derivations in Multi-C -Ternary Algebras Now, we prove the Hyers-Ulam stability of derivations on multi-C -ternary algebras for the following functional equation: D f .x1 ; ; xm / D 0 for all 2 T1 WD f 2 C W j j D 1g and x1 ; ; xm 2 A. Theorem 6.13. Let ..Ak ; k kk / W k 1/ be a multi-C -ternary algebra. Let f W A ! A be a mapping for which there exist the functions ' W Amk ! Œ0; 1/ and W A3k ! Œ0; 1/ satisfying (6.34), (6.35) and (6.37) such that k.f .Œx1 ; y1 ; z1 / Œf .x1 /; y1 ; z1 Œx1 ; f .y1 /; z1 Œx1 ; y1 ; f .z1 /; ; f .Œxk ; yk ; zk / Œf .xk /; yk ; zk Œxk ; f .yk /; zk Œxk ; yk ; f .zk //kk

(6.50)

.x1 ; y1 ; z1 ; ; xk ; yk ; zk /

for all 2 T1 and x11 ; ; x1m ; ; xk1 ; ; xkm ; x1 ; ; xk ; y1 ; ; yk , z1 ; ; zk 2 A. If there exists L < 1 such that m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ ' mx11 ; 0; ; 0; mx21 ; 0; ; 0; ; mxk1 ; 0; ; 0 m m m ‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ mL' x11 ; 0; ; 0; x21 ; 0; ; 0; ; xk1 ; 0; ; 0

for all x11 ; x21 ; ; xk1 2 A, then there exists a unique derivation ı W A ! A such that k.f .x1 / ı.x1 /; ; f .xk / ı.xk //kk m

m

m

‚ …„ ƒ ‚ …„ ƒ ‚ …„ ƒ 1 ' x1 ; 0; ; 0; x2 ; 0; ; 0; ; xk ; 0; ; 0 m mL

(6.51)

for all x1 ; ; xk 2 A. Proof. By the same reasoning as in the proof of Theorem 6.9, there exists a unique C-linear mapping ı W A ! A satisfying (6.50) and the mapping ı W A ! A is given by ı.x/ D lim

n!1

f .mn x/ mn

(6.52)

250


for all x 2 A. It follows from (6.34), (6.37) and (6.52) that k.ı.Œx1 ; y1 ; z1 / Œı.x1 /; y1 ; z1 Œx1 ; ı.y1 /; z1 Œx1 ; y1 ; f .z1 /; ; ı.Œxk ; yk ; zk / Œı.xk /; yk ; zk Œxk ; ı.yk /; zk Œxk ; yk ; ı.zk //kk 1 D lim 3n k.f .Œmn x1 ; mn y1 ; mn z1 / Œf .mn x1 /; mn y1 ; mn z1 n!1 m Œmn x1 ; f .mn y1 /; mn z1 Œmn x1 ; mn y1 ; f .mn z1 /; ; f .Œmn xk ; mn yk ; mn zk / Œf .mn xk /; mn yk ; mn zk Œmn xk ; f .mn yk /; mn zk Œmn xk ; mn yk ; f .mn zk //kk lim m3n .mn x1 ; mn y1 ; mn z1 ; ; mn xk ; mn yk ; mn zk / n!1

D0 for all x1 ; ; xm ; y1 ; ; yk ; z1 ; ; zk 2 A and so ı.Œx; y; z/ D Œı.x/; y; z C Œx; ı.y/; z C Œx; y; ı.z/ for all x; y; z 2 A. Thus ı W A ! A is a derivation satisfying (6.50). This completes the proof. Corollary 6.14. Let ..Ak ; k kk / W k 1/ be a multi-C -ternary algebra. Let r 1 for r > 2 and p < 1 for r < 2. If f W X ! Y is an odd mapping such that k.Du f .x11 ; ; x1d /; ; Du f .xk1 ; ; xkd //kk

d X

kx1j kp C : : : C

jD1

d X

kxkj kp

jD1

for all u 2 U.A/, k 1 and x11 ; ; x1d ; ; xk1 ; ; xkd 2 X, then there exists a unique A-linear generalized additive mapping W X ! Y such that k..x1 / f .x1 /; ; .xk / f .xk //kk

rp1 .jjx1 jjp C C jjxk jjp / .rp1 2p1 /.d2 Cl d2 Cl2 C 1/

for all k 1 and x1 ; ; xk 2 X. Proof. Taking L D

2p1 rp1

and

'k .x11 ; ; x1d ; ; xk1 ; ; xkd / D

d X

kx1j kp C : : : C

jD1

d X

kxkj kp

jD1

for all k 1 and x11 ; ; x1d ; ; xk1 ; ; xkd 2 X in Theorem 6.17, we get the desired assertion. Theorem 6.19. Let r ¤ 2. Let f W X ! Y be an odd mapping for which there exists a function ' W X kd ! Œ0; 1/ such that 2j ' j!1 rj lim

rj rj rj rj x ; ; x ; ; x ; ; xkd 11 1d k1 2j 2j 2j 2j

D0

6.3 Generalized Additive Mappings and Isomorphisms in Multi-C -Algebras

257

and k.Du f .x11 ; ; x1d /; ; Du f .xk1 ; ; xkd //kk '.x11 ; ; x1d ; ; xk1 ; ; xkd / for all u 2 U.A/ and x11 ; ; x1d ; ; xk1 ; ; xkd 2 X. If there exists L < 1 such that d

‚r

d

…„ ƒ ‚ …„ ƒ r r r x11 ; x11 ; ; 0; x21 ; x21 ; ; 0; ' 2 2 2 2 d

d

‚ …„ ƒ r r ; xk1 ; xk1 ; ; 0 2 2

d

d

…„ ƒ ‚ …„ ƒ ‚ …„ ƒ r ‚ L' x11 ; x11 ; ; 0; x21 ; x21 ; ; 0; ; xk1 ; xk1 ; ; 0 2 for all x11 ; x21 ; ; xk1 2 X, then there exists a unique A-linear generalized additive mapping W X ! Y such that sup k..x1 / f .x1 /; ; .xk / f .xk //kk k1

sup k1

L ' x1 ; x1 ; 0; ; 0; „ ƒ‚ … 2.d2 Cl d2 Cl2 C 1/.1 L/ ; xk ; xk ; 0; ; 0 „ ƒ‚ …

d 2 times

d 2 times

for all x1 ; ; xk 2 X. Proof. Note that f .0/ D 0 and f .x/ D f .x/ for all x 2 X since f is an odd mapping. Let u D 1 2 U.A/. Putting xi1 D xi2 D x1 and xi3 D D xim D 0 (1 i k) in (6.56), we have rf 2 x1 2f .x1 /; ; rf 2 xk 2f .xk / r r k 1 ' x1 ; x1 ; 0; ; 0; ; xk ; xk ; 0; ; 0 : „ ƒ‚ … „ ƒ‚ … d2 Cl d2 Cl2 C 1 d 2 times

Letting t WDd2 Cl d2 Cl2 C 1, we have f .x1 / 2 f r x1 ; ; f .xk / 2 f r xk r 2 r 2 k

d 2 times

258


1 r r r r ' x1 ; x1 ; 0; ; 0; ; xk ; xk ; 0; ; 0 rt 2 2 „ ƒ‚ … 2 2 „ ƒ‚ … d 2 times

d 2 times

L ' x1 ; x1 ; 0; ; 0; ; xk ; xk ; 0; ; 0 „ ƒ‚ … „ ƒ‚ … 2t d 2 times

d 2 times

for all x1 ; ; xk 2 X. The rest of the proof is similar to the proof of Theorem 6.17. This completes the proof. Corollary 6.20. Let r < 2 and , p > 1 be positive real numbers or let r > 2 and , p < 1 be positive real numbers. Let f W X ! Y be an odd mapping such that k.Du f .x11 ; ; x1d /; ; Du f .xk1 ; ; xkd //kk

d X

kx1j kp C C

jD1

d X

kxkj kp

jD1

for all u 2 U.A/ and x11 ; ; x1d ; ; xk1 ; ; xkd 2 X. Then there exists a unique A-linear generalized additive mapping W X ! Y such that sup k..x1 / f .x1 /; ; .xk / f .xk //kk k1

sup k1

rp1 .jjx1 jjp C C kxk kp / .2p1 rp1 /.d2 Cl d2 Cl2 C 1/

for all x 2 X. Proof. Define '.x11 ; ; x1d ; ; xk1 ; ; xkd / D

d X

kx1j kp C C

jD1

and put L D

rp1 2p1

d X

kxkj kp

jD1

in Theorem 6.19. Then we get the desired result.

Now, we investigate the Hyers–Ulam stability of linear mappings for the case d D 2. Theorem 6.21. Let r ¤ 2. Let f W X ! Y be an odd mapping for which there exists a function ' W X 2k ! Œ0; 1/ such that rj ' j!1 2j lim

2j 2j 2j 2j x1 ; j y1 ; ; j xk ; j yk j r r r r

D0


259

and ux C uy 1 1 uf .x1 / uf .y1 /; rf r ux C uy k k ; rf uf .xk / uf .yk / k r '.x1 ; y1 ; ; xk ; yk /

(6.61)

for all u 2 U.A/ and x1 ; xk ; y1 ; yk 2 X. If there exists L < 1 such that '

2 2 2 2 2 2 x1 ; x1 ; x2 ; x2 ; ; xk ; xk r r r r r r

2 L' .x1 ; x1 ; x2 ; x2 ; ; xk ; xk / r

for all x1 ; xk 2 X. Then there exists a unique A-linear generalized additive mapping W X ! Y such that sup k..x1 / f .x1 /; ; .xk / f .xk //kk k1

sup k1

L '.x1 ; x1 ; ; xk ; xk / 2.1 L/

for all x1 ; ; xk 2 X. Proof. Let u D 1 2 U.A/. Putting xi D yi (1 i k) in (6.61), we have rf 2 x1 2f .x1 /; ; rf 2 xk 2f .xk / r r k '.x1 ; x1 ; ; xk ; xk / for all x 2 X and so f .x1 / r f 2 x1 ; ; f .xk / r f 2 xk 2 r 2 r k

1 '.x1 ; x1 ; ; xk ; xk / 2

for all x 2 X. The rest of the proof is the same as in the proof of Theorem 6.17. This completes the proof. Corollary 6.22. Let r > 2 and , p > 1 be positive real numbers or let r < 2 and , p < 1 be positive real numbers. Let f W X ! Y be an odd mapping such that

260


ux C uy 1 1 uf .x1 / uf .y1 /; rf r ux C uy k k uf .xk / uf .yk / ; rf k r

k X

.kxj kp C kyj kp /

jD1

for all u 2 U.A/ and x1 ; xk 2 X. Then there exists a unique A-linear generalized additive mapping W X ! Y such that sup k..x1 / f .x1 /; ; .xk / f .xk //kk sup k1

k1

k X rp1 jjxj jjp rp1 2p1 jD1

for all x1 ; xk 2 X. Proof. Define '.x1 ; y1 ; ; xk ; yk / D

k X

.jjxj jjp C jjyj jjp /

jD1

and apply Theorem 6.21. Then we get the desired result.

Theorem 6.23. Let r ¤ 2. Let f W X ! Y be an odd mapping for which there exists a function ' W X 2k ! Œ0; 1/ such that 2j lim j ' j!1 r and

rj rj rj rj x ; y ; ; x ; yk 1 1 k 2j 2j 2j 2j

D0

ux C uy 1 1 uf .x1 / uf .y1 /; rf r ux C uy k k uf .xk / uf .yk / ; rf k r '.x1 ; y1 ; ; xk ; yk /

(6.62)

for all u 2 U.A/ and x1 ; xk ; y1 ; yk 2 X. If there exists L < 1 such that '

r r r r r r x1 ; x1 ; x2 ; x2 ; ; xk ; xk L' .x1 ; x1 ; x2 ; x2 ; ; xk ; xk / 2 2 2 2 2 2 2

r

for all x1 ; xk 2 X. Then there exists a unique A-linear generalized additive mapping W X ! Y such that


261

sup k..x1 / f .x1 /; ; .xk / f .xk //kk k1

sup k1

1 '.x1 ; x1 ; ; xk ; xk / 2.1 L/

for all x1 ; ; xk 2 X. Proof. Let u D 1 2 U.A/. Putting xi D yi (1 i k) in (6.62), we have rf 2 x1 2f .x1 /; ; rf 2 xk 2f .xk / r r k '.x1 ; x1 ; ; xk ; xk / for all x1 ; ; xk 2 X and so f .x1 / 2 f r x1 ; ; f .xk / 2 f r xk r 2 r 2 k 1 r r r r x1 ; x1 ; ; xk ; xk ' r 2 2 2 2 1 L'.x1 ; x1 ; ; xk ; xk / 2 for all x1 ; ; xk 2 X. The rest of the proof is similar to the proof of Theorem 6.17. This completes the proof. Corollary 6.24. Let r > 2 and , p > 1 be positive real numbers or let r < 2 and , p < 1 be positive real numbers. Let f W X ! Y be an odd mapping such that ux C uy 1 1 uf .x1 / uf .y1 /; rf r ux C uy k k ; rf uf .xk / uf .yk / k r

k X

.kxj kp C kyj kp /

jD1

for all u 2 U.A/ and x1 ; xk 2 X. Then there exists a unique A-linear generalized additive mapping W X ! Y such that sup k..x1 / f .x1 /; ; .xk / f .xk //kk sup k1

for all x1 ; xk 2 X.

k1

k X rp1 kxj kp 2p1 rp1 jD1

262


Proof. Define '.x1 ; y1 ; ; xk ; yk / D

k X

.kxj kp C kyj kp /

jD1


6.3.2 Isomorphisms in Unital Multi-C -Algebras Assume that A and B are unital multi-C -algebras with the unit e. Let U.A/ be the set of unitary elements in A. Now, we investigate C -algebra isomorphisms in unital multi-C -algebras. Theorem 6.25. Let r ¤ 2. Let h W A ! B be an odd bijective mapping satisfying h

2n uy D h u h.y/ rn rn

2n

for all u 2 U.A/, y 2 A and n 0 for which there exists a function ' W Akd ! Œ0; 1/ such that j 2 2j 2j 2j rj lim ' j x11 ; ; j x1d ; ; j xk1 ; ; j xkd D 0; j!1 2j r r r r k.D h.x11 ; ; x1d /; ; D h.xk1 ; ; xkd //kk '.x11 ; ; x1d ; ; xk1 ; ; xkd / and n n n n h 2 u h 2 u1 ; ; h 2 u h 2 uk 1 k rn rn rn rn k 2n 2n 2n 2n ' n u1 ; ; n u1 ; ; n uk ; ; n uk „r ƒ‚ r … „r ƒ‚ r … d times

d times

for all 2 S1 WD f 2 C W jj D 1g, u1 ; ; uk 2 U.A/, n 0 and n n x11 ; ; xkd 2 A. Assume that limn!1 2r n h. 2rn e/ is invertible. Then the odd bijective mapping h W A ! B is a C -algebra isomorphism. Proof. Consider the multi-C -algebras A and B as left Banach modules over the unital multi-C -algebra C. By Theorem 6.17, there exists a unique C-linear generalized additive mapping H W A ! B such that


263

sup k.h.x1 / H.x1 /; ; h.xk / H.xk //kk k1

sup k1

' x1 ; x1 ; 0; ; 0; ; xk ; xk ; 0; ; 0 „ ƒ‚ … „ ƒ‚ … 2.d2 Cl d2 Cl2 C 1/ 1

d 2 times

d 2 times

for all x1 ; ; xk 2 A in which H W A ! B is given by r n 2n h nx n!1 2n r

H.x/ D lim

for all x 2 A. The rest of the proof is easy. This completes the proof.

Corollary 6.26. Let r > 2 and , p > 1 be positive real numbers or let r < 2 and , p < 1 be positive real numbers. Let h W A ! B be an odd bijective mapping satisfying h

2n

2n uy D h u h.y/ rn rn

for all u 2 U.A/, y 2 A and n 0 such that k.D h.x11 ; ; x1d /; ; D h.xk1 ; ; xkd //kk

d X .kx1j kp C C kxkj kp / jD1

and n n n n h 2 u h 2 u1 ; ; h 2 u h 2 uk 1 k n n n n r r r r

k

2pn kd pn r for all 2 S1 , u 2 U.A/, n 0 and x11 ; ; xkd 2 A. Assume that limn!1 2r n h. 2rn e/ is invertible. Then the odd bijective mapping h W A ! B is a C -algebra isomorphism. n

n

Proof. Define '.x11 ; ; x1d ; ; xk1 ; ; xkd / D

d X .jjx1j jjp C C jjxkj jjp / jD1


264


6.4 Additive Functional Inequalities in Proper Multi-CQ -Algebras In this section, we approximate the following additive functional inequality: dC1 dC1 X X f .x1i /; ; f .xki // . iD1

k

iD1

PdC1 x PdC1 x iD1 1i iD1 ki ; ; mf mf k m m

(6.63)

for all x11 ; ; xk dC1 2 X, where d 2 is a fixed integer. Also, we investigate homomorphisms in proper multi-CQ -algebras and derivations on proper multiCQ -algebras associated with the above additive functional inequality.

6.4.1 Stability of C-Linear Mappings in Multi-Banach Spaces Now, we investigate the Hyers-Ulam stability of C-linear mappings in multi-Banach spaces associated with the multi-additive functional inequality (6.63). In this section, we assume that .X; k k/ and .Y; k k/ are Banach spaces such that .X k ; k kk / and .Y k ; k kk / are multi-Banach spaces. Lemma 6.27. Let f W X ! Y be a mapping satisfying (6.63) in which f .0/ D 0. Then f is additive. Proof. Letting x3 D D xdC1 D 0 and replacing x1 by x and x2 by x in (6.63), we have kf .x/ C f .x/k kmf .0/k D 0 for all x 2 X. Hence f .x/ D f .x/ for all x 2 X. Replacing x1 by x, x2 by y and x3 by x y and putting x4 D D xdC1 D 0 in (6.63), we have kf .x/ C f .y/ f .x C y/k D kf .x/ C f .y/ C f .x y/k kmf .0/kY D0 for all x; y 2 X. Thus we have f .x C y/ D f .x/ C f .y/ for all x; y 2 X. This completes the proof.

6.4 Additive Functional Inequalities in Proper Multi-CQ -Algebras

265

Theorem 6.28. Let f W X ! Y be a mapping. If there exists a function ' W X kdCk ! Œ0; 1/ satisfying the following: dC1 dC1 X X f .x1i /; ; f .xki // . iD1

iD1

k

PdC1 x PdC1 x iD1 1i iD1 ki ; ; mf mf k m m C '.x11 ; ; x1 dC1 ; ; xk1 ; ; xk dC1 /

(6.64)

and '.x Q 11 ; ; x1 dC1 ; ; xk1 ; ; xk dC1 / WD

1 X

sup dj ' dj1 x11 ; ; dj1 x1 dC1 ; ; dj1 xk1 ; ; dj1 xk dC1

jD0 k1

2, be nonnegative real numbers and f W A ! B be a mapping satisfying (7.16), (7.17) and (7.18). Then there exists a unique nonArchimedean C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB

j2jr

kxkrA j2j

(7.25)

for all x 2 A. Proof. The proof follows from Theorem 7.3 by taking '.x; y/ D

.x; y/ WD .kxkrA C kykrA /;

.x/ WD kxkrA

for all x; y 2 A and L D j2j1r .

Theorem 7.5. Let f W A ! B be an odd mapping for which there exist the functions '; W A2 ! Œ0; 1/ and W A ! Œ0; 1/ satisfying (7.2), (7.3) and (7.4). If there exists L < 1 such that '.x; 3x/ j2jL'

x 3x ; 2 2

for all x 2 A and (7.5), (7.6) and (7.7) hold then there exists a unique nonArchimedean C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB

1 '.x; 3x/ j2j j2jL

(7.26)

for all x 2 A. Proof. Consider the set X WD fg W A ! Bg and introduce the generalized metric on X as follows: d.g; h/ D inffC 2 RC W kg.x/ h.x/kB C'.x; 3x/; 8x 2 Ag; which .X; d/ is complete. Now, we consider the linear mapping J W X ! X defined by Jg.x/ WD

1 g.2x/ 2

7.1 Stability of Jensen Type Functional Equations: The Fixed Point Approach

311

for all x 2 A. Now, we have d.Jg; Jh/ Ld.g; h/ for all g; h 2 X. Letting D 1 and replacing y by 3x in (7.2), we have kf .2x/ 2f .x/kB '.x; 3x/

(7.27)

for all x 2 A and so 1 1 '.x; 3x/ f .x/ f .2x/ B 2 j2j for all x 2 A. Hence d.f ; Jf / H W A ! B such that

1 . j2j



(7.28)

for all x 2 A. The mapping H is a unique fixed point of J in the set Y D fg 2 X W d.f ; g/ < 1g: This implies that H is a unique mapping satisfying (7.28) such that there exists C 2 .0; 1/ such that kH.x/ f .x/kB C'.x; 3x/ for all x 2 A; (2) d.J n f ; H/ ! 0 as n ! 1. This implies the equality f .2n x/ D H.x/ n!1 2n lim


1 : j2j j2jL

This implies that the inequality (7.26) holds. The rest of the proof is similar to the proof of Theorem 7.1. This completes the proof.

312

7 Stability of Functional Equations in Non-Archimedean Banach Algebras

Corollary 7.6. Let r < 12 , be nonnegative real numbers and f W A ! B be an odd mapping such that kD f .x; y/kB kxkrA kykrA ; kf .xy/ f .x/f .y/kB

kxkrA

kykrA

(7.29) (7.30)

and kf .x / f .x/ kB kxk2r A

(7.31)

for all 2 T1 and x; y 2 A. Then there exists a unique non-Archimedean C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB

3r kxk2r A j2j j2j2r

(7.32)


.x; y/ WD kxkrA kykrA ;

.x/ WD kxkrA

for all x; y 2 A and L D j2j2r1 .

Theorem 7.7. Let f W A ! B be an odd mapping for which there exist the functions '; W A2 ! Œ0; 1/ and W A ! Œ0; 1/ satisfying (7.2), (7.3), and (7.4). If there exists L < 1 such that '.x; 3x/

1 L'.2x; 6x/ j2j

for all x 2 A, and also (7.20), (7.21) and (7.22) hold, then there exists a unique non-Archimedean C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB

L '.x; 3x/ j2j j2jL

for all x 2 A. Proof. We consider the linear mapping J W X ! X defined by x Jg.x/ WD 2g 2 for all x 2 A. It follows from (7.27) that x 3x L x '.x; 3x/ f .x/ 2f . / ' ; 2 B 2 2 j2j

(7.33)


for all x 2 A. Hence d.f ; Jf / H W A ! B such that

L . 2

313



(7.34)

for all x 2 A. The mapping H is a unique fixed point of J in the set Y D fg 2 X W d.f ; g/ < 1g: This implies that H is a unique mapping satisfying (7.34) such that there exists C 2 .0; 1/ such that kH.x/ f .x/kB C'.x; 3x/ for all x 2 A; (2) d.J n f ; H/ ! 0 as n ! 1. This implies the equality x lim 2n f n D H.x/ n!1 2 for all x 2 A; 1 d.f ; Jf /, which implies the inequality (3) d.f ; H/ 1L d.f ; H/

L ; 2 2L

which implies that the inequality (7.33) holds. The rest of the proof is similar to the proof of Theorem 7.1. Corollary 7.8. Let r > 1, be nonnegative real numbers and f W A ! B be an odd mapping satisfying (7.29), (7.30) and (7.31). Then there exists a unique nonArchimedean C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB

j2j2r

kxk2r A j2j

(7.35)


.x; y/ WD kxkrA kykrA ;


.x/ WD kxkrA

314


7.1.2 Stability of Derivations in Non-Archimedean C -Algebras Assume that A is a non-Archimedean C -algebra with the norm k kA . Now, we prove the Hyers-Ulam stability of derivations in non-Archimedean C algebras for the functional equation D f .x; y/ D 0. Theorem 7.9. Let f W A ! A be a mapping for which there exist the functions '; W A2 ! Œ0; 1/ such that kD f .x; y/kA '.x; y/

(7.36)

and kf .xy/ f .x/y xf .y/kA

.x; y/

(7.37)

for all 2 T1 and all x; y 2 A. If there exists L < 1 such that x '.x; 0/ j2jL' ; 0 2 for all x 2 A, (7.5) and (7.6) hold. Then there exists a unique non-Archimedean derivation ı W A ! A such that kf .x/ ı.x/kA

L '.x; 0/ 1L

(7.38)

for all x 2 A. Proof. By the same reasoning as in the proof of Theorem 7.1, there exists a unique involution C-linear mapping ı W A ! A satisfying (7.38). Also, the mapping ı W A ! A is given by f .2n x/ n!1 2n


kı.xy/ ı.x/y xı.y/kA 1 D lim kf .4n xy/ f .2n x/ 2n y 2n xf .2n y/kA n!1 j4jn 1 lim .2n x; 2n y/ n!1 j4jn D0 for all x; y 2 A. Then we have ı.xy/ D ı.x/y C xı.y/ for all x; y 2 A. Thus ı W A ! A is a derivation satisfying (7.38). This completes the proof.


315

Corollary 7.10. Let r < 1, be nonnegative real numbers and f W A ! A be a mapping such that kD f .x; y/kA .kxkrA C kykrA /

(7.39)

kf .xy/ f .x/y xf .y/kA .kxkrA C kykrA /

(7.40)

and

for all 2 T1 and x; y 2 A. Then there exists a unique derivation ı W A ! A such that kf .x/ ı.x/kA

j2jr kxkrA j2j j2jr

(7.41)


.x; y/ WD .kxkrA C kykrA /

for all x; y 2 A and L D j2jr1 .

Remark 7.11. Let f W A ! A be a mapping for which there exist the functions '; W A2 ! Œ0; 1/ satisfying (7.36) and (7.37). If there exists L < 1 such that '.x; 0/

1 L'.2x; 0/ j2j

for all x 2 A, (7.20) and (7.21) hold, then there exists a unique derivation ı W A ! A such that kf .x/ ı.x/kA

L '.x; 0/ j2j j2jL

(7.42)

for all x 2 A. Corollary 7.12. Let r > 2, be nonnegative real numbers and f W A ! A be a mapping satisfying (7.39) and (7.40). Then there exists a unique derivation ı W A ! A such that kf .x/ ı.x/kA for all x 2 A.

kxkrA j2jr j2j

(7.43)

316


Proof. The proof follows from Remark 7.11 by taking '.x; y/ D



Remark 7.13. For the inequalities controlled by the product of powers of norms, one can obtain similar results to Theorems 7.5, 7.7 and Corollaries 7.6, 7.8.

7.1.3 Stability of Homomorphisms in Non-Archimedean Lie C -Algebras A non-Archimedean C -algebra C endowed with the Lie product Œx; y WD

xy yx 2

on C is called a non-Archimedean Lie C -algebra (see [224, 225, 227]). Definition 7.14. Let A and B be non-Archimedean Lie C -algebras. A C-linear mapping H W A ! B is called a non-Archimedean Lie C -algebra homomorphism if H.Œx; y/ D ŒH.x/; H.y/ for all x; y 2 A. Throughout this section, assume that A is a non-Archimedean Lie C -algebra with the norm k kA and B is a non-Archimedean Lie C -algebra with the norm k kB . Now, we prove the Hyers-Ulam stability of homomorphisms in nonArchimedean Lie C -algebras for the functional equation D f .x; y/ D 0. Theorem 7.15. Let f W A ! B be a mapping for which there are functions '; W A2 ! Œ0; 1/ satisfying (7.2) such that kf .Œx; y/ Œf .x/; f .y/kB

.x; y/

(7.44)

for all x; y 2 A. If there exists L < 1 such that '.x; 0/ j2jL'

x 2

;0

for all x 2 A, and also (7.5) and (7.6) hold, then there exists a unique nonArchimedean Lie C -algebra homomorphism H W A ! B satisfying (7.8).


317

Proof. By the same reasoning as in the proof of Theorem 7.1, there exists a unique C-linear mapping H W A ! B satisfying (7.8). Also, the mapping H W A ! B is given by f .2n x/ n!1 2n

H.x/ D lim for all x 2 A. It follows from (7.44) that

kH.Œx; y/ ŒH.x/; H.y/kB D lim

1 kf .4n Œx; y/ Œf .2n x/; f .2n y/kB j4jn

lim

1 .2n x; 2n y/ j4jn

n!1

n!1

D0 for all x; y 2 A. Then we have H.Œx; y/ D ŒH.x/; H.y/ for all x; y 2 A. Thus H W A ! B is a non-Archimedean Lie C -algebra homomorphism satisfying (7.8). This completes the proof. Corollary 7.16. Let r < 1, be nonnegative real numbers and f W A ! B be a mapping satisfying (7.16) such that kf .Œx; y/ Œf .x/; f .y/kB .kxkrA C kykrA /

(7.45)

for all x; y 2 A. Then there exists a unique non-Archimedean Lie C -algebra homomorphism H W A ! B satisfying (7.19). Proof. The proof follows from Theorem 7.15 by taking '.x; y/ D



Remark 7.17. Let f W A ! B be a mapping for which there exist the functions '; W A2 ! Œ0; 1/ and W A ! Œ0; 1/ satisfying (7.2), (7.5), (7.6) and (7.44). If there exists L < 1 such that '.x; 0/

1 L'.2x; 0/ j2j

for all x 2 A, then there exists a unique non-Archimedean Lie C -algebra homomorphism H W A ! B satisfying (7.23).

318


Corollary 7.18. Let r > 2, be nonnegative real numbers and f W A ! B be a mapping satisfying (7.16) and (7.45). Then there exists a unique non-Archimedean Lie C -algebra homomorphism H W A ! B satisfying (7.25). Proof. The proof follows from Remark 7.17 by taking '.x; y/ D



Remark 7.19. For the inequalities controlled by the product of powers of norms, one can obtain similar results to Theorems 7.5, 7.7 and their corollaries.

7.1.4 Stability of Non-Archimedean Lie Derivations in C -Algebras First, we give the following definition on the non-Archimedean Lie derivation in a non-Archimedean Lie C -algebra. Definition 7.20. Let A be a non-Archimedean Lie C -algebra. A C-linear mapping ı W A ! A is called a non-Archimedean Lie derivation if ı.Œx; y/ D Œı.x/; y C Œx; ı.y/ for all x; y 2 A. Assume that A is a non-Archimedean Lie C -algebra with the norm k kA . Now, we prove the Hyers-Ulam stability of non-Archimedean Lie derivations on non-Archimedean Lie C -algebras for the functional equation D f .x; y/ D 0: Theorem 7.21. Let f W A ! A be a mapping for which there exists a function '; W A2 ! Œ0; 1/ satisfying (7.5), (7.6) and (7.36) such that kf .Œx; y/ Œf .x/; y Œx; f .y/kA

.x; y/

(7.46)

for all x; y 2 A. If there exists L < 1 such that '.x; 0/ j2jL'

x 2

;0

for all x 2 A. Then there exists a unique non-Archimedean Lie derivation ı W A ! A satisfying (7.38).


319

Proof. By the same reasoning as in the proof of Theorem 7.1, there exists a unique involution C-linear mapping ı W A ! A satisfying (7.38). Also, the mapping ı W A ! A is given by f .2n x/ n!1 2n


kı.Œx; y/ Œı.x/; y Œx; ı.y/kA D lim

1 kf .4n Œx; y/ Œf .2n x/; 2n y Œ2n x; f .2n y/kA j4jn

lim

1 .2n x; 2n y/ j4jn

n!1

n!1

D0 for all x; y 2 A. Then we have ı.Œx; y/ D Œı.x/; y C Œx; ı.y/ for all x; y 2 A. Thus ı W A ! A is a non-Archimedean Lie derivation satisfying (7.38). This completes the proof. Corollary 7.22. Let r < 1, be nonnegative real numbers and f W A ! A be a mapping satisfying (7.39) such that kf .Œx; y/ Œf .x/; y Œx; f .y/kA .kxkrA C kykrA /

(7.47)

for all x; y 2 A. Then there exists a unique non-Archimedean Lie derivation ı W A ! A satisfying (7.41). Proof. The proof follows from Theorem 7.15 by taking '.x; y/ D

.x; y/ WD .kxkrA C kykrA /;


.x/ WD kxkrA

Remark 7.23. Let f W A ! A be a mapping for which there exist functions '; W A2 ! Œ0; 1/ and W A ! Œ0; 1/ satisfying (7.20), (7.21), (7.22), (7.36) and (7.46). If there exists L < 1 such that '.x; 0/

1 L'.2x; 0/ j2j

for all x 2 A, then there exists a unique non-Archimedean Lie derivation ı W A ! A satisfying (7.42).

320


Corollary 7.24. Let r > 2, be nonnegative real numbers and f W A ! A be a mapping satisfying (7.39) and (7.47). Then there exists a unique non-Archimedean Lie derivation ı W A ! A satisfying (7.43). Proof. The proof follows from Remark 7.23 by taking '.x; y/ D



Remark 7.25. For the inequalities controlled by the product of powers of norms, one can obtain similar results to Theorems 7.5, 7.7 and their corollaries.

7.2 Stability for m-Variable Additive Functional Equations Using the fixed point method, we prove the Hyers-Ulam stability of homomorphisms and derivations on non-Archimedean C -algebras and non-Archimedean Lie C -algebras for the following additive functional equation: m m m m X X X X f mxi C xj C f xi D 2f mxi iD1

iD1

jD1; j¤i

(7.48)

iD1

for each m 2. For any mapping f W A ! B, we define D f .x1 ; ; xm / WD

m X iD1

f mxi C

m X jD1; j¤i

xj

m m X X Cf xi 2f mxi iD1

iD1

for all 2 T1 WD f 2 C W j j D 1g and x1 ; ; xm 2 A.

7.2.1 Stability of Homomorphisms and Derivations in Non-Archimedean C -Algebras Now, we prove the Hyers–Ulam stability of homomorphisms in non-Archimedean C -algebras for the functional equation D f .x1 ; ; xm / D 0. Theorem 7.26. Let f W A ! B be a mapping for which there exist the functions ' W Am ! Œ0; 1/, W A2 ! Œ0; 1/ and W A ! Œ0; 1/ such that jmj < 1 is far from zero and kD f .x1 ; ; xm /kB '.x1 ; ; xm /; kf .xy/ f .x/f .y/kB

.x; y/

(7.49) (7.50)

7.2 Stability for m-Variable Additive Functional Equations

321

and kf .x / f .x/ kB .x/

(7.51)

for all 2 T1 and x1 ; ; xm ; x; y 2 A. If there exists L < 1 such that '.mx1 ; ; mxm / jmjL'.x1 ; ; xm /; .mx; my/ jmj2 L .x; y/

(7.52) (7.53)

and

.mx/ jmjL .x/

(7.54)

for all x; y; x1 ; ; xm 2 A, then there exists a unique non-Archimedean C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB

1 '.x; 0; ; 0/ jmj jmjL

(7.55)

for all x 2 A. Proof. It follows from (7.52), (7.53), (7.54) and L < 1 that 1 '.mn x1 ; ; mn xm / D 0; n!1 jmjn lim

lim

n!1

(7.56)

1 .mn x; mn y/ D 0 jmj2n

(7.57)

1

.mn x/ D 0 jmjn

(7.58)

and lim

n!1

for all x; y; x1 ; ; xm 2 A. Let us define ˝ to be the set of all mappings g W A ! B and introduce a generalized metric on ˝ as follows: d.g; h/

(7.59)

D inffk 2 .0; 1/ W kg.x/ h.x/kB < k.x; 0; ; 0/; 8x 2 Ag; which .˝; d/ is a generalized complete metric space. Now, we consider the function J W ˝ ! ˝ defined by Jg.x/ D m1 g.mx/ for all x 2 A and g 2 ˝. Note that, for all g; h 2 ˝, d.g; h/ < k H) kg.x/ h.x/kB < k.x; 0; ; 0/

322


1 1 k .mx; 0; ; 0/ H) g.mx/ h.mx/ < B m m jmj 1 1 H) g.mx/ h.mx/ < kL.mx; 0; ; 0/ B m m H) d.Jg; Jh/ < kL:

(7.60)

From this, it is easy to see that d.Jg; Jk/ Ld.g; h/ for all g; h 2 ˝, that is, J is a self-function of ˝ with the Lipschitz constant L. Putting D 1, x D x1 and x2 D x3 D D xm D 0 in (7.49) we have kf .mx/ mf .x/kB .x; 0; ; 0/

(7.61)

1 1 .x; 0; ; 0/ f .x/ f .mx/ B m jmj

(7.62)

for all x 2 A. Then

for all x 2 A, that is, d.Jf ; f /

1 < 1: jmj

Now, from the fixed point method, it follows that there exists a fixed point H of J in ˝ such that 1 f .mn x/ n!1 mn

H.x/ D lim

(7.63)

for all x 2 A since limn!1 d.J n f ; H/ D 0. On the other hand, it follows from (7.49), (7.56) and (7.63) that 1 kD H.x1 ; ; xm /kB D lim n Df .mn x1 ; ; mn xm / n!1 m B 1 lim .mn x1 ; ; mn xm / n!1 jmjn D 0: By the similar method as in above, we have H.mx/ D H.m x/

(7.64)


323

for all 2 T1 and x 2 A. Thus one can show that the mapping H W A ! B is C-linear. It follows from (7.50), (7.57) and (7.63) that kH.xy/ H.x/H.y/kB D lim

1 kf .m2n xy/ f .mn x/f .mn y/kB jmj2n

lim

1 .mn x; mn y/ jmj2n

n!1

n!1

D0

(7.65)

for all x; y 2 A and so H.xy/ D H.x/H.y/ for all x; y 2 A. Thus H W A ! B is a homomorphism satisfying (7.55). Also, by (7.51), (7.58), (7.63) and the similar method, we have H.x / D H.x/ . This completes the proof. Corollary 7.27. Let r > 1, be nonnegative real numbers and f W A ! B be a mapping such that kD f .x1 ; ; xm /kB .kx1 krA C kx2 krA C C kxm krA /; kf .xy/ f .x/f .y/kB .kxkrA kykrA /

(7.66)

and kf .x / f .x/ kB kxkrA for all 2 T1 and x1 ; ; xm ; x; y 2 A. Then there exists a unique non-Archimedean C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB

kxkrA jmj jmjr

(7.67)

for all x 2 A. Proof. The proof follows from Theorem 7.26 by taking '.x1 ; ; xm / WD .kx1 krA C kx2 krA C C kxm krA /; .x; y/ WD .kxkrA kykrA /;

.x/ WD kxkrA for all x1 ; ; xm ; x; y 2 A and L D jmjr1 .

(7.68)

324


Now, we prove the Hyers-Ulam stability of derivations on non-Archimedean C -algebras for the functional equation D f .x1 ; ; xm / D 0: Remark 7.28. Let f W A ! A be a mapping for which there exist the functions ' W Am ! Œ0; 1/, W A2 ! Œ0; 1/ and W A ! Œ0; 1/ such that jmj < 1 is far from zero and kD f .x1 ; ; xm /kA '.x1 ; ; xm /; kf .xy/ f .x/y xf .y/kA

.x; y/

(7.69)

and kf .x / f .x/ kA .x/ for all 2 T1 and x1 ; ; xm ; x; y 2 A. If there exists L < 1 such that (7.52), (7.53) and (7.54) hold, then there exists a unique non-Archimedean C -algebra derivation ı W A ! A such that kf .x/ ı.x/kA

1 '.x; 0; ; 0/ jmj jmjL

for all x 2 A.

7.2.2 Stability of Homomorphisms and Derivations in Non-Archimedean Lie C -Algebras Now, we prove the Hyers-Ulam stability of homomorphisms in non-Archimedean Lie C -algebras for the functional equation D f .x1 ; ; xm / D 0. Theorem 7.29. Let f W A ! B be a mapping for which there exist the functions ' W Am ! Œ0; 1/ and W A2 ! Œ0; 1/ such that (7.49) and (7.51) hold and kf .Œx; y/ Œf .x/; f .y/kB

.x; y/

(7.70)

for all 2 T1 and x; y 2 A. If there exists L < 1 and (7.52) and (7.53) hold, then there exists a unique non-Archimedean Lie C -algebra homomorphism H W A ! B such that (7.55) hold.


325

Proof. By the same reasoning as in the proof of Theorem 7.26, we can find the mapping H W A ! B given by f .mn x/ n!1 mn

H.x/ D lim

(7.71)

for all x 2 A. It follows from (7.53) and (7.71) that kH.Œx; y/ ŒH.x/; H.y/kB D lim

1 kf .m2n Œx; y/ Œf .mn x/; f .mn y/kB jmj2n

lim


(7.72)

H.Œx; y/ D ŒH.x/; H.y/

(7.73)

n!1

n!1

D0 for all x; y 2 A and so

for all x; y 2 A. Thus H W A ! B is a non-Archimedean Lie C -algebra homomorphism satisfying (7.55). This completes the proof. Corollary 7.30. Let r > 1, be nonnegative real numbers and f W A ! B be a mapping such that kD f .x1 ; ; xm /kB .kx1 krA C kx2 krA C C kxm krA /; kf .Œx; y/ Œf .x/; f .y/kB kxkrA kykrA

(7.74)

and kf .x / f .x/ kB kxkrA for all 2 T1 and x1 ; ; xm ; x; y 2 A. Then there exists a unique non-Archimedean Lie C -algebra homomorphism H W A ! B such that kf .x/ H.x/kB

kxkrA jmj jmjr

(7.75)

for all x 2 A. Proof. The proof follows from Theorem 7.29 and the method similar to Corollary 7.27.

326


Definition 7.31. Let A be a non-Archimedean Lie C -algebra. A C-linear mapping ı W A ! A is called a non-Archimedean Lie derivation if ı.Œx; y/ D Œı.x/; y C Œx; ı.y/ for all x; y 2 A. Now, we prove the Hyers-Ulam stability of derivations on non-Archimedean Lie C -algebras for the functional equation D f .x1 ; ; xm / D 0: Theorem 7.32. Let f W A ! A be a mapping for which there exist the functions ' W Am ! Œ0; 1/ and W A2 ! Œ0; 1/ such that (7.49) and (7.51) hold and kf .Œx; y/ Œf .x/; y Œx; f .y/kA

.x; y/

(7.76)

for all x; y 2 A. If there exists L < 1 and (7.52) and (7.53) hold, then there exists a unique non-Archimedean Lie derivation ı W A ! A such that such that (7.55) holds. Proof. It is straight forward to show, there exists a unique C-linear mapping ı W A ! A satisfying (7.55) and the mapping ı W A ! A is given by f .mn x/ n!1 mn

ı.x/ D lim

(7.77)

for all x 2 A. It follows from (7.53) and (7.75) that kı.Œx; y/ Œı.x/; y Œx; ı.y/kA D lim

1 kf .m2n Œx; y/ Œf .mn x/; mn y Œmn x; f .mn y/kA jmj2n

lim


n!1

n!1

(7.78)

D0 for all x; y 2 A and so ı.Œx; y/ D Œı.x/; y C Œx; ı.y/ for all x; y 2 A. Thus ı W A ! A is a non-Archimedean Lie derivation satisfying (7.55). This completes the proof.

References

1. Abellanas, L., Alonso, L.: A general setting for Casimir invariants. J. Math. Phys. 16, 1580–1584 (1975) 2. Abramov, V., Kerner, R., Le Roy, B.: Hypersymmetry: a Z3 -graded generalization of supersymmetry. J. Math. Phys. 38, 1650–1669 (1997) 3. Aczél, J.: The general solution of two functional equations by reduction to functions additive in two variables and with aid of Hamel-bases. Glasnik Mat.-Fiz. Astronom. Drustvo Mat. Fiz. Hrvatske 20, 65–73 (1965) 4. Aczél, J.: Lectures on Functional equations and Their Applications. Academic, New York/London (1966) 5. Aczél, J.: A Short Course on Functional Equations. D. Reidel Publishing Company, Dordrecht (1987) 6. Aczél, J., Dhombres, J.: Functional Equations in Several Variables. Cambridge University Press, Cambridge (1989) 7. Agarwal, R.P., Cho, Y.J., Park, C., Saadati, R.: Approximate homomorphisms and derivation in multi-Banach algebras. Comment. Math. 51, 23–38 (2011) 8. Agarwal, R.P., Cho, Y.J., Saadati, R., Wang, S.: Nonlinear L-fuzzy stability of cubic functional equations. J. Inequal. Appl. 2012, 19 pp. (2012) 9. Alli, G., Sewell, G.L.: New methods and structures in the theory of the multi-mode Dicke laser model. J. Math. Phys. 36, 5598–5626 (1995) 10. Almira, J.M., Luther, U.: Inverse closedness of approximation algebras. J. Math. Anal. Appl. 314, 30–44 (2006) 11. Amyari, M., Moslehian, M.S.: Approximately ternary semigroup homomorphisms. Lett. Math. Phys. 77, 1–9 (2006) 12. J. An: On an additive functional inequality in normed modules over a C -algebra. J. Korea Soc. Math. Educ. Ser. B. Pure Appl. Math. 15, 393–400 (2008) 13. An, J., Cui, J., Park, C.: Jordan -derivations on C -algebras and JC -algebras. Abstr. Appl. Anal. 2008, Article ID 410437 (2008) 14. Andrews, L.C.: Special Functions for Engineers and Applied Mathematicians. MacMilan, New York (1985) 15. Antoine, J.P., Inoue, A., Trapani, C.: O -dynamical systems and -derivations of unbounded operator algebras. Math. Nachr. 204, 5–28 (1999) 16. Antoine, J.P., Inoue, A., Trapani, C.: Partial -Algebras and Their Operator Realizations. Kluwer Academic Publishers, Dordrecht (2002) 17. Aoki, T.: On the stability of the linear transformation in Banach spaces. J. Math. Soc. Jpn. 2, 64–66 (1950)

© Springer International Publishing Switzerland 2015 Y.J. Cho et al., Stability of Functional Equations in Banach Algebras, DOI 10.1007/978-3-319-18708-2

327

328

References

18. Ara, P., Mathieu, M.: Local Multipliers of C -Algebras. Springer, London (2003) 19. Baak, C.: Cauchy-Rassias stability of Cauchy-Jensen additive mappings in Banach spaces. Acta Math. Sin. 22, 1789–1796 (2006) 20. Baak, C., Boo, D., Rassias, Th.M.: Generalized additive mapping in Banach modules and isomorphisms between C -algebras. J. Math. Anal. Appl. 314, 150–156 (2006) 21. Bae, J., Park, W.: Approximate bi-homomorphisms and bi-derivations in C -ternary algebras. Bull. Korean Math. Soc. 47, 195–209 (2010) 22. Bae, J., Park, W.: Generalized Ulam-Hyers stability of C -ternary algebra 3-homomorphisms for a functional equation. J. Chungcheong Math. Soc. 24, 147–162 (2011) 23. Bagarello, F.: Applications of topological -algebras of unbounded operators. J. Math. Phys. 39, 6091–6105 (1998) 24. Bagarello, F.: Fixed point results in topological -algebras of unbounded operators. Publ. RIMS Kyoto Univ. 37, 397–418 (2001) 25. Bagarello, F.: Applications of topological -algebras of unbounded operators to modified quons. Nuovo Cimento B 117, 593–611 (2002) 26. Bagarello, F., Karwowski, W.: Partial -algebras of closed linear operators in Hilbert space. Publ. RIMS Kyoto Univ. 21, 205–236 (1985); 22, 507–511 (1986) 27. Bagarello, F., Morchio, G.: Dynamics of mean-field spin models from basic results in abstract differential equations. J. Stat. Phys. 66, 849–866 (1992) 28. Bagarello, F., Sewell, G.L.: New structures in the theory of the laser model II: microscopic dynamics and a non-equilibrium entropy principle. J. Math. Phys. 39, 2730–2747 (1998) 29. Bagarello, F., Trapani, C.: Almost mean field Ising model: an algebraic approach. J. Stat. Phys. 65, 469–482 (1991) 30. Bagarello, F., Trapani, C.: A note on the algebraic approach to the “almost” mean field Heisenberg model. Nuovo Cimento B 108, 779–784 (1993) 31. Bagarello, F., Trapani, C.: States and representations of CQ -algebras. Ann. Inst. H. Poincaré 61, 103–133 (1994) 32. Bagarello, F., Trapani, C.: The Heisenberg dynamics of spin systems: a quasi--algebras approach. J. Math. Phys. 37, 4219–4234 (1996) 33. Bagarello, F., Trapani, C.: CQ -algebras: structure properties. Publ. RIMS Kyoto Univ. 32, 85–116 (1996) 34. Bagarello, F., Trapani, C.: Morphisms of certain Banach C -modules. Publ. RIMS Kyoto Univ. 36, 681–705 (2000) 35. Bagarello, F., Trapani, C.: Algebraic dynamics in O -algebras: a perturbative approach. J. Math. Phys. 43, 3280–3292 (2002) 36. Bagarello, F., Inoue, A., Trapani, C.: Some classes of topological quasi -algebras. Proc. Am. Math. Soc. 129, 2973–2980 (2001) 37. Bagarello, F., Inoue, A., Trapani, C.: -derivations of quasi--algebras. Int. J. Math. Math. Sci. 21, 1077–1096 (2004) 38. Bagarello, F., Inoue, A., Trapani, C.: Exponentiating derivations of quasi--algebras: possible approaches and applications. Int. J. Math. Math. Sci. 2005, 2805–2820 (2005) 39. Bagarello, F., Trapani, C., Triolo, S.: Quasi--algebras of measurable operators. Studia Math. 172, 289–305 (2006) 40. Bahyrycz, A., Piszczek, M.: Hyperstability of the Jensen functional equation. Acta Math. Hung. 142, 353–365 (2014) 41. Bahyrycz, A., Brzde¸k, J., Piszczek, M., Sikorska, J.: Hyperstability of the Fréchet equation and a characterization of inner product spaces. J. Funct. Spaces Appl. 2013, Article ID 496361, 6 pp. (2013) 42. Bahyrycz, A., Brzde¸k, J., Le´sniak, Z.: On approximate solutions of the generalized Volterra integral equation. Nonlinear Anal. Real World Appl. 20, 59–66 (2014) 43. Baker, J.A.: The stability of certain functional equations. Proc. Am. Math. Soc. 112, 729–732 (1991)

References

329

44. Baktash, E., Cho, Y.J., Jalili, M., Saadati, R., Vaezpour, S.M.: On the stability of cubic mappings and quadratic mappings in random normed spaces. J. Inequal. Appl. 2008, Article ID 902187, 11 pp. (2008) 45. Banach, S.: Sur les operations dans les ensembles abstraits et leur application aux euations integrales. Fundam. Math. 3, 133–181 (1922) 46. Bazunova, N., Borowiec, A., Kerner, R.: Universal differential calculus on ternary algebras. Lett. Math. Phys. 67, 195–206 (2004) 47. Belaid, B., Elqorachi, E., Rassias, Th.M.: On the Hyers–Ulam stability of approximately Pexider mappings. Math. Inequal. Appl. 11, 805–818 (2008) 48. Benyamini, Y., Lindenstrauss, J.: Geometric Nonlinear Functional Analysis, Vol. 1. Colloquium Publications, vol. 48. American Mathematical Society, Providence (2000) 49. Birkhoff, G.: Orthogonality in linear metric spaces. Duke Math. J. 1, 169–172 (1935) 50. Bodaghi, A., Park, C.: Jordan -derivations and quadratic Jordan -derivations on real C algebras and real JC -algebras. Int. J. Geom. Methods Mod. Phys. 10(10), 1350051, 16 pp. (2013) 51. Bourgin, D.G.: Classes of transformations and bordering transformations. Bull. Am. Math. Soc. 57, 223–237 (1951) 52. Brešar, M., Zalar, B.: On the structure of Jordan -derivations. Colloq. Math. LXIII, 163–171 (1992) 53. Brillouët-Belluot, N., Brzdek, J., Ciepliński, K.: On some developments in Ulam’s type stability. Abstr. Appl. Anal. 2012, Article ID 716936, 41 pp. (2012) 54. Brown, L., Pedersen, G.: C -algebras of real rank zero. J. Funct. Anal. 99, 131–149 (1991) 55. Brzde¸k, J.: Remarks on hyperstability of the Cauchy functional equation. Aequ. Math. 86, 255–267 (2013) 56. Brzde¸k, J., Ciepliński, K.: A fixed point approach to the stability of functional equations in non–Archimedean metric spaces. Nonlinear Anal. 74, 6861–6867 (2011) 57. Brzde¸k, J., Sikorska, J.: A conditional exponential functional equation and its stability. Nonlinear Anal. 72, 2923–2934 (2010) 58. Brzde¸k, J., Chudziak, J., Páles, Zs.: A fixed point approach to stability of functional equations. Nonlinear Anal. 74, 6728–6732 (2011) 59. Brzde¸k, J., Ciepliński, K.: Hyperstability and superstability. Abstr. Appl. Anal. 2013, Article ID 401756, 13 pp. (2013) 60. Brzde¸k, J., C˘adariu, L., Ciepliński, K.: Fixed point theory and the Ulam stability. J. Funct. Spaces 2014, Article ID 829419, 16 pp. (2014) 61. Brzde¸k, J., Ciepliński, K., Le´sniak, Z.: On Ulam’s type stability of the linear equation and related issues. Discret. Dyn. Nat. Soc. 2014, Article ID 536791, 14 pp. (2014) 62. C˘adariu, L., Radu, V.: Fixed points and the stability of Jensen’s functional equation. J. Inequal. Pure Appl. Math. 4(1), Article 4 (2003) 63. C˘adariu, L., Radu, V.: On the stability of the Cauchy functional equation: a fixed point approach. In: Iteration Theory (ECIT ’02). Grazer Mathematische Berichte, vol. 346, pp. 43–52. Karl-Franzens-Universität, Graz (2004) 64. C˘adariu, L., Radu, V.: Fixed point methods for the generalized stability of functional equations in a single variable. Fixed Point Theory Appl. 2008, Article ID 749392, 15 pp. (2008) 65. Carlsson, S.O.: Orthogonality in normed linear spaces. Ark. Mat. 4, 297–318 (1962) 66. Castro, L.P., Ramos, A.: Hyers-Ulam-Rassias stability for a class of nonlinear Volterra integral equations. Banach J. Math. Anal. 3, 47–56 (2009) 67. Cauchy, A.L.: Cours d’analyse de l’Ecole Polytechnique. Vol. 1, Analyse algebrique. Debure, Paris (1821) 68. Cayley, A.: On the 34 concomitants of the ternary cubic. Am. J. Math. 4, 1–15 (1881) 69. Chahbi, A., Bounader, N.: On the generalized stability of d’Alembert functional equation. J. Nonlinear Sci. Appl. 6, 198–204 (2013) 70. Cho, Y.J., Saadati, R.: Lattictic non-Archimedean random stability of ACQ functional equation. Adv. Differ. Equ. 2011, 12 pp. (2011)

330

References

71. Cho, Y.J., Park, C., Saadati, R.: Functional inequalities in non-Archimedean Banach spaces. Appl. Math. Lett. 60, 1994–2002 (2010) 72. Cho, Y.J., Eshaghi Gordji, M., Zolfaghari, S.: Solutions and stability of generalized mixed type QC functional equations in random normed spaces. J. Inequal. Appl. 2010, Article ID 403101, 16 pp. (2010) 73. Cho, Y.J., Park, C., Rassias, Th.M., Saadati, R.: Inner product spaces and functional equations. J. Comput. Anal. Appl. 13, 296–304 (2011) 74. Cho, Y.J., Saadati, R., Shabanian, S., Vaezpour, S.M.: On solution and stability of a twovariable functional equations. Discret. Dyn. Nat. Soc. 2011, Article ID 527574, 18 pp. (2011) 75. Cho, Y.J., Kang, J.I., Saadati, R.: Fixed points and stability of additive functional equations on the Banach algebras. J. Comput. Anal. Appl. 14, 1103–1111 (2012) 76. Cho, Y.J., Kang, S.M., Sadaati, R.: Nonlinear random stability via fixed-point method. J. Appl. Math. 2012, Article ID 902931, 44 pp. (2012) 77. Cho, Y.J., Saadati, R., Vahidi, J.: Approximation of homomorphisms and derivations on nonArchimedean Lie C -algebras via fixed point method. Discret. Dyn. Nat. Soc. 2012, Article ID 373904, 9 pp. (2012) 78. Cho, Y.J., Rassias, Th.M., Saadati, R.: Stability of Functional Equations in Random Normed Spaces. Springer, New York (2013) 79. Cho, Y.J., Saadati, R., Yang, Y.O.: Random-ternary algebras and application. J. Inequal. Appl. 2015, 26 (2015) 80. Cholewa, P.W.: Remarks on the stability of functional equations. Aequ. Math. 27, 76–86 (1984) 81. Chu, H.Y., Kang, D.S., Rassias, Th.M.: On the stability of a mixed n-dimensional quadratic functional equation. Bull. Belg. Math. Soc.–Simon Stevin 15, 9–24 (2008) 82. Ciepliński, K.: Generalized stability of multi-additive mappings. Appl. Math. Lett. 23, 1291–1294 (2010) 83. Cieplinski, K.: Stability of multi-additive mappings in non-Archimedean normed spaces. J. Math. Anal. Appl. 373, 376–383 (2011) 84. Ciepliński, K.: On the generalized Hyers-Ulam stability of multi-quadratic mappings. Comput. Math. Appl. 62, 3418–3426 (2011) 85. Ciepliński, K.: Stability of multi-additive mappings in ˇ-Banach spaces. Nonlinear Anal. 75, 4205–4212 (2012) 86. Ciepliński, K.: Applications of fixed point theorems to the Hyers-Ulam stability of functional equations-a Survey. Ann. Funct. Anal. 3, 151–164 (2012) 87. Czerwik, S.: On the stability of the quadratic mapping in normed spaces. Abh. Math. Sem. Univ. Hambg. 62, 59–64 (1992) 88. Czerwik, P.: Functional Equations and Inequalities in Several Variables. World Scientific Publishing Company, River Edge/Hong Kong/Singapore/London (2002) 89. Czerwik, S.: Stability of Functional Equations of Ulam-Hyers-Rassias Type. Hadronic Press, Palm Harbor (2003) 90. Dales, H.G.: Banach Algebras and Automatic Continuity. London Mathematical Society Monographs, New Series, vol. 24, Oxford University Press, Oxford (2000) 91. Dales, H.G., Moslehian, M.S.: Stability of mappings on multi-normed spaces. Glasg. Math. J. 49, 321–332 (2007) 92. Dales, H.G., Polyakov, M.E.: Multi-normed spaces. Dissertationes Math. (Rozprawy Mat.) 488, 165 pp. (2012) 93. Daletskii, Y.L., Takhtajan, L.: Leibniz and Lie algebra structures for Nambu algebras. Lett. Math. Phys. 39, 127–141 (1997) 94. Dehghanian, M., Modarres Mosadegh, S.M.S., Park, C., Shin, D.Y.: C -ternary 3-derivations on C -ternary algebras. J. Inequal. Appl. 2013, 124 (2013) 95. Diaz, J., Margolis, B.: A fixed point theorem of the alternative for contractions on a generalized complete metric space. Bull. Am. Math. Soc. 74, 305–309 (1968) 96. Dixmier, J.: C -Algebras. North-Holland Publishing Company, Amsterdam/New York/Oxford (1977)

References

331

97. Ebadian, A., Ghobadipour, N., Baghban, H.: Stability of bi- -derivations on JB -triples. Int. J. Geom. Methods Mod. Phys. 9, Article ID 1250051, 12 pp. (2012) 98. Ekhaguere, G.O.S.: Partial W -dynamical systems. In: Current Topics in Operator Algebras, Proceedings of the Satellite Conference of ICM-90, pp. 202–217, World Scientific, Singapore (1991) 99. Epifanio, G., Trapani, C.: Quasi--algebras valued quantized fields. Ann. Inst. H. Poincaré 46, 175–185 (1987) 100. Eshaghi Gordji, M., Ghobadipour, N.: Stability of .˛; ˇ; /-derivations on Lie C -algebras. Int. J. Geom. Methods Mod. Phys. 7, 1093–1102 (2010) 101. Eshaghi Gordji, M., Khodaei, H.: Stability of Functional Equations. Lap Lambert Academic Publishing, Saarbrücken (2010) 102. Eshaghi Gordji, M., Khodaei, H.: A fixed point technique for investigating the stability of .˛; ˇ; /-derivations on Lie C -algebras. Nonlinear Anal. 76, 52–57 (2013) 103. Eshaghi Gordji, M., Savadkouhi, M.B.: Stability of cubic and quartic functional equations in non-Archimedean spaces. Acta Appl. Math. 110, 1321–1329 (2010) 104. Eshaghi Gordji, M., Cho, Y.J., Ghaemi, M.B., Majani, H.: Approximately quintic and sextic ˆ mappings form r-divisible groups into Serstnev probabilistic Banach spaces: fixed point method. Discret. Dyn. Nat. Soc. 2011, Article ID 572062, 16 pp. (2011) 105. Eshaghi Gordji, M., Ghaemi, M.B., Cho, Y.J., Majani, M.: A general system of EulerLagrange-type quadratic functional equations in Menger probabilistic non-Archimedean 2-normed spaces. Abstr. Appl. Anal. Vol. 2011, Article ID 208163, 21 pp. (2011) 106. Eshaghi Gordji, M., Ghaemi, B., Rassias, J.M., Alizadeh, B.: Nearly ternary quadratic higher derivations on non-Archimedean ternary Banach algebras: a fixed point approach. Abstr. Appl. Anal. 2011, Article ID 417187, 18 pp. (2011) 107. Eshaghi Gordji, M., Ramezani, M., Cho, Y.J., Baghani, H.: Approximate Lie brackets: a fixed point approach. J. Inequal. Appl. 2012, 125 (2012) 108. Eskandani, G.Z.: On the Hyers-Ulam-Rassias stability of an additive functional equation in quasi-Banach spaces. J. Math. Anal. Appl. 345, 405–409 (2008) 109. Eskandani, G.Z., Gavruta, P.: Hyers-Ulam-Rassias stability of pexiderized Cauchy functional equation in 2-Banach spaces. J. Nonlinear Sci. Appl. 5(Special issue), 459–465 (2012) 110. Eskandani, G.Z., Rassias, Th.M.: Hyers-Ulam-Rassias stability of derivations in proper JCQ -triples. Mediterr. J. Math. 10, 1391–1400 (2013) 111. Faiziev, V.A., Rassias, Th.M.: The space of pseudocharacters on semigroups. Nonlinear Funct. Anal. Appl. 5, 107–126 (2000) 112. Faiziev, V.A., Rassias, Th.M., Sahoo, P.K.: The space of . ; /–additive mappings on semigroups. Trans. Am. Math. Soc. 354, 4455–4472 (2002) 113. Fechner, W.: Stability of a functional inequalities associated with the Jordan-von Neumann functional equation. Aequ. Math. 71, 149–161 (2006) 114. Fleming, R.J., Jamison, J.E.: Isometries on Banach Spaces: Function Spaces. Monographs and Surveys in Pure and Applied Mathematics, vol. 129. Chapman & Hall/CRC, Boca Raton/London/New York/Washington, DC (2003) 115. Forti, G.L.: An existence and stability theorem for a class of functional equations. Stochastica 4, 22–30 (1980) 116. Forti, G.L.: Comments on the core of the direct method for proving Hyers-Ulam stability of functional equations. J. Math. Anal. Appl. 295, 127–133 (2004) 117. Forti, G.L.: Elementary remarks on Ulam-Hyers stability of linear functional equations. J. Math. Anal. Appl. 328, 109–118 (2007) 118. Fredenhagen, K., Hertel, J.: Local algebras of observables and pointlike localized fields. Commun. Math. Phys. 80, 555–561 (1981) 119. Gajda, Z.: On stability of additive mappings. Int. J. Math. Math. Sci. 14, 431–434 (1991) 120. Gao, Z.X., Cao, H.X., Zheng, W.T., Xu, L.: Generalized Hyers-Ulam-Rassias stability of functional inequalities and functional equations. J. Math. Inequal. 3, 63–77 (2009) 121. Gauss, C.F.: Theoria moyus corporum caelestium. Perthes–Besser, Hamburg (1809)

332

References

122. Gˇavruta, P.: On the stability of some functional equations. In: Stability of Mappings of HyersUlam Type, pp. 93–98. Hadronic Press. Palm Harbor (1994) 123. Gˇavruta, P.: A generalization of the Hyers-Ulam-Rassias stability of approximately additive mappings. J. Math. Anal. Appl. 184, 431–436 (1994) 124. Gˇavruta, P.: On a problem of G. Isac and Th.M. Rassias concerning the stability of mappings. J. Math. Anal. Appl. 261, 543–553 (2001) 125. Gˇavruta, P.: On the Hyers-Ulam-Rassias stability of the quadratic mappings. Nonlinear Funct. Anal. Appl. 9, 415–428 (2004) 126. Gilányi, A.: Eine zur Parallelogrammgleichung äquivalente Ungleichung. Aequ. Math. 62, 303–309 (2001) 127. Gilányi, A.: On a problem by K. Nikodem. Math. Inequal. Appl. 5, 707–710 (2002) 128. Goodearl, K.R.: Notes on Real and Complex C -Algebras. Shiva Mathematics Series IV. Shiva Publishing, Cheshire (1982) 129. Grabiec, A.: The generalized Hyers-Ulam stability of a class of functional equations. Publ. Math. Debr. 48, 217–235 (1996) 130. Gudder, S., Strawther, D.: Orthogonally additive and orthogonally increasing functions on vector spaces. Pac. J. Math. 58, 427–436 (1975) 131. Haag, R., Kastler, D.: An algebraic approach to quantum field theory. J. Math. Phys. 5, 848–861 (1964) 132. Hardy, G.H., Littlewood, J.E., Polya, G.: Inequalities, 2nd edn. Cambridge University Press, Cambridge (1952) 133. Hyers, D.H.: On the stability of the linear functional equation. Proc. Natl. Acad. Sci. USA 27, 222–224 (1941) 134. Hyers, D.H.: The stability of homomorphisms and related topics. In: Rassias, Th.M. (ed.), Global Analysis Analysis on Manifolds. Teubner-Texte zur Mathematik, pp. 140–153. B.G. Teubner, Leipzig (1983) 135. Hyers, D.H., Rassias, Th.M.: Approximate homomorphisms. Aequ. Math. 44, 125–153 (1992) 136. Hyers, D.H., Isac, G., Rassias, Th.M.: Stability of Functional Equations in Several Variables. Birkhäuser, Basel (1998) 137. Hyers, D.H., Isac, G., Rassias, Th.M.: On the asymptoticity aspect of Hyers-Ulam stability of mappings. Proc. Am. Math. Soc. 126, 425–430 (1998) 138. Isac, G., Rassias, Th.M.: On the Hyers-Ulam stability of additive mappings. J. Approx. Theory 72, 131–137 (1993) 139. Isac, G., Rassias, Th.M.: Stability of -additive mappings: applications to nonlinear analysis. Int. J. Math. Math. Sci. 19, 219–228 (1996) 140. Jacobson, N.: Lie Algebras. Dover, New York (1979) 141. James, R.C.: Orthogonality in normed linear spaces. Duke Math. J. 12, 291–302 (1945) 142. Jesen, B., Karpf, J., Thorup, A.: Some functional equations in groups and rings. Math. Scand. 22, 257–265 (1968) 143. Johnson, B.E.: Approximately multiplicative maps between Banach algebras. J. Lond. Math. Soc. 37, 294–316 (1988) 144. Jun, K., Kim, H.: On the stability of Appolonius’ equation. Bull. Belg. Math. Soc. Simon Stevin 11, 615–624 (2004) 145. Jun, K., Lee, Y.: A generalization of the Hyers-Ulam-Rassias stability of Jensen’s equation. J. Math. Anal. Appl. 238, 305–315 (1999) 146. Jun, K., Kim, B., Shin, D.: On Hyers-Ulam-Rassias stability of the Pexider equation. J. Math. Anal. Appl. 239, 20–29 (1999) 147. Jun, K.W., Kim, H.M., Rassias, J.M.: Extended Hyers-Ulam stability for Cauchy-Jensen mappings. J. Differ. Equ. Appl. 13, 1139–1153 (2007) 148. Jung, S.-M.: On the Hyers-Ulam-Rassias stability of approximately additive mappings. J. Math. Anal. Appl. 204, 221–226 (1996) 149. Jung, S.-M.: Hyers-Ulam-Rassias Stability of Functional Equations in Mathematical Analysis. Hadronic Press, Palm Harbor (2001)

References

333

150. Jung, S.-M.: Hyers-Ulam stability of a system of first order linear differential equations with constant coefficients. J. Math. Anal. Appl. 320, 549–561 (2006) 151. Jung, S.-M.: A fixed point approach to the stability of an equation of the square spiral. Banach J. Math. Anal. 1, 148–153 (2007) 152. Jung, S.-M.: A fixed point approach to the stability of a Volterra integral equation. Fixed Point Theory Appl. 2007, Article ID 57064, 9 pp. (2007) 153. Jung, S.-M.: A fixed point approach to the stability of isometries. J. Math. Anal. Appl. 329, 879–890 (2007) 154. Jung, S.-M.: A fixed point approach to the stability of a Volterra integral equation. Fixed Point theory Appl. 2007, Article ID 57064, 9 pp. (2007) 155. Jung, Y.S., Chang, I.S.: The stability of a cubic type functional equation with the fixed point alternative. J. Math. Anal. Appl. 306, 752–760 (2005) 156. Jung, S.-M., Kim, T.S.: A fixed point approach to the stability of the cubic functional equation. Bol. Soc. Mat. Mexicana 12, 51–57 (2006) 157. Jung, S.-M., Lee, Z.H.: A fixed point approach to the stability of quadratic functional equation with involution. Fixed Point Theory Appl. 2008, Article ID 732086, 11 pp. (2008) 158. Jung, S.-M., Rassias, Th.M.: Ulam’s problem for approximate homomorphisms in connection with Bernoulli’s differential equation. Appl. Math. Comput. 187, 223–227 (2007) 159. Jung, S.-M.: Hyers-Ulam stability of linear differential equations of first order. Appl. Math. Lett. 22, 70–74 (2009) 160. Jung, S.-M.: Hyers–Ulam–Rassias Stability of Functional Equations in Nonlinear Analysis. Springer Optimization and Its Applications, vol. 48. Springer, New York/Dordrecht/Heidelberg/London (2011) 161. Jung, S.-M., Rassias, Th.M.: Generalized Hyers–Ulam stability of Riccati differential equation. Math. Inequal. Appl. 11, 777–782 (2008) 162. Jung, S.-M., Rassias, J.M.: A fixed point approach to the stability of a functional equation of the spiral of Theodorus. Fixed Point Theory Appl. 2008, Article ID 945010, 7 pp. (2008) 163. Jung, S.-M., Rassias, M.Th.: A linear functional equation of third order associated to the Fibonacci numbers. Abstr. Appl. Anal. 2014, Article ID 137468, 7 pp. (2014) 164. Jung, S.-M., Kim, T.S., Lee, K.S.: A fixed point approach to the stability of quadratic functional equation. Bull. Korean Math. Soc. 43, 531–541 (2006) 165. Jung, S.-M., Popa, D., Rassias, M.Th.: On the stability of the linear functional equation in a single variable on complete metric groups. J. Glob. Optim. 59, 165–171 (2014) 166. Jung, S.-M., Rassias, M.Th., Mortici, C.: On a functional equation of trigonometric type. Appl. Math. Comput. 252, 294–303 (2015) 167. Kadison, R.V., Pedersen, G.: Means and convex combinations of unitary operators. Math. Scand. 57, 249–266 (1985) 168. Kadison, R.V., Ringrose, J.R.: Fundamentals of the Theory of Operator Algebras. Academic, New York (1983) 169. Kapranov, M., Gelfand, I.M., Zelevinskii, A.: Discriminants, Resultants and Multidimensional Determinants. Birkhäuser, Berlin (1994) 170. Kenari, H.M., Saadati, R., Park, C.: Homomorphisms and derivations in C -ternary algebras via fixed point method. Adv. Differ. Equ. 2012, 13 pp. (2012) 171. Kerner, R.: The cubic chessboard. Geometry and physics. Class. Quantum Gravity 14(1A), A203–A225 (1997) 172. Kerner, R.: Ternary algebraic structures and their applications in physics. Univ. P. and M. Curie, Paris (2000, preprint) 173. Khodaei, H., Eshaghi Gordji, M., Kim, S.S., Cho, Y.J.: Approximation of radical functional equations related to quadratic and quartic mappings. J. Math. Anal. Appl. 397, 284–297 (2012) 174. Khrennikov, A.: Non-Archimedean Analysis: Quantum Paradoxes, Dynamical Systems and Biological Models. Mathematics and Its Applications. vol. 427, Kluwer Academic Publishers, Dordrecht (1997)

334

References

175. Kim, G.H., Dragomir, S.S.: The stability of the generalized d’Alembert and Jensen functional equations II. Int. J. Math. Math. Sci. 2006, Article ID 43185, 12 pp. (2006) 176. Kochanek, T., Lewicki, M.: Stability problem for number-theoretically multiplicative functions. Proc. Am. Math. Soc. 135, 2591–2597 (2007) 177. Kuczma, M.: Functional equations on restricted domains. Aequ. Math. 18, 1–34 (1978) 178. Kuczma, M.: An Introduction to the Theory of Functional Equations and Inequalities. Uniwersytet Slaski, Warszawa/Krakow/Katowice (1985) 179. Lassner, G.: Algebras of unbounded operators and quantum dynamics. Physica 124 A, 471–480 (1984) 180. Lee, S.H., Park, C.: Hyers-Ulam-Rassias stability of isometric homomorphisms in quasiBanach algebras. J. Comput. Anal. Appl. 10, 39–51 (2008) 181. Lee, J.R., Park, C., Shin, D.Y.: On the stability of generalized additive functional inequalities in Banach spaces. J. Inequal. Appl. 2008, Article ID 210626, 13 pp. (2008) 182. Lee, J.R., Park, C., Shin, D.Y.: Stability of an additive functional inequality in proper CQ algebras. Bull. Korean Math. Soc. 48, 853–871 (2011) 183. Lee, Y.H., Jung, S.-M., Rassias, M.Th.: On an n-dimensional mixed type additive and quadratic functional equation. Appl. Math. Comput. 228, 13–16 (2014) 184. Legendre, A.M.: Elements de geometrie, Note II. Didot, Paris (1791) 185. Li, Y., Shen, Y.: Hyers-Ulam stability of nonhomogeneous linear differential equations of second order. Int. J. Math. Math. Sci. 2009, Article ID 576852, 7 pp. (2009) 186. Lu, G., Park, C.: Additive functional inequalities in Banach spaces. J. Inequal. Appl. 2012, 294 (2012) 187. Lungu, N., Popa, D.: Hyers-Ulam stability of a first order partial differential equation. J. Math. Anal. Appl. 385, 86–91 (2012) 188. Luxemburg, W.A.J.: On the convergence of successive approximations in the theory of ordinary differential equations, II. Koninkl, Nederl. Akademie van Wetenschappen, Amsterdam, Proc. Ser. A (5) 61; Indag. Math. 20, 540–546 (1958) 189. Mihe¸t, D.: The fixed point method for fuzzy stability of the Jensen functional equation. Fuzzy Sets Syst. 160, 1663–1667 (2009) 190. Mihe¸t, D., Radu, V.: On the stability of the additive Cauchy functional equation in random normed spaces. J. Math. Anal. Appl. 343, 567–572 (2008) 191. Mirzavaziri, M., Moslehian, M.S.: A fixed point approach to stability of a quadratic equation. Bull. Braz. Math. Soc. (N.S.) 37, 361–376 (2006) 192. Miura, T., Miyajima, S., Takahasi, S.E.: Hyers-Ulam stability of linear differential operator with constant coefficients. Math. Nachr. 258, 90–96 (2003) 193. Miura, T., Miyajima, S., Takahasi, S.E.: A characterization of Hyers-Ulam stability of first order linear differential operators. J. Math. Anal. Appl. 286, 136–146 (2003) 194. Miura, T., Jung, S.-M., Takahasi, S.E.: Hyers-Ulam-Rassias stability of the Banach space valued linear differential equations y0 D y. J. Korean Math. Soc. 41, 995–1005 (2004) 195. Mlesnite, O.: Existence and Ulam-Hyers stability results for coincidence problems. J. Nonlinear Sci. Appl. 6, 108–116 (2013) 196. Mohammadi, M., Cho, Y.J., Park, C., Vetro, P., Saadati, R.: Random stability of an additivequadratic-quartic functional equation. J. Inequal. Appl. 2010, Article ID 754210, 18 pp. (2010) 197. Montigny, M., Patera, J.: Discrete and continuous graded contractions of Lie algebras and superalgebras. J. Phys. A 24, 525–547 (1991) 198. Morchio, G., Strocchi, F.: Mathematical structures for long range dynamics and symmetry breaking. J. Math. Phys. 28, 622–635 (1987) 199. Mortici, C., Rassias, M.Th., Jung, S.-M.: On the stability of a functional equation associated with the Fibonacci numbers. Abstr. Appl. Anal. 2014, Article ID 546046, 6 pp. (2014) 200. Moslehian, M.S.: On the orthogonal stability of the pexiderized quadratic equation. J. Differ. Equ. Appl. 11, 999–1004 (2005) 201. Moslehian, M.S.: Almost derivations on C -ternary rings. Bull. Belg. Math. Soc.-Simon Stevin 14, 135–142 (2007)

References

335

202. Moslehian, M.S.: Superstability of higher derivations in multi-Banach algebras. Tamsui Oxf. J. Math. Sci. 24, 417–427 (2008) 203. Moslehian, M.S., Rassias, Th.M.: Stability of functional equations in non-Archimedean spaces. Appl. Anal. Discret. Math. 1, 325–334 (2007) 204. Moslehian, M.S., Rassias, Th.M.: Orthogonal stability of additive type equations. Aequ. Math. 73, 249–259 (2007) 205. Moslehian, M.S., Rassias, Th.M.: Generalized Hyers-Ulam stability of mappings on normed Lie triple systems. Math. Inequal. Appl. 11, 371–380 (2008) 206. Moslehian, M.S., Nikodem, K., Popa, D.: Asymptotic aspect of the quadratic functional equation in multi-normed spaces. J. Math. Anal. Appl. 355, 717–724 (2009) 207. Moszner, Z.: On the stability of functional equations. Aequ. Math. 77, 33–38 (2009) 208. Najati, A., Cho, Y.J.: Generalized Hyers-Ulam stability of the pexiderized Cauchy functional equation in non-Archimedean spaces. Fixed Point Theory Appl. 2011, Article ID 309026, 11 pp. (2011) 209. Najati, A., Park, C.: The pexiderized Apollonius-Jensen type additive mapping and isomorphisms between C -algebras. J. Differ. Equ. Appl. 14, 459–479 (2008) 210. Najati, A., Park, C.: Stability of a generalized Euler-Lagrange type additive mapping and homomorphisms in C -algebras II. J. Nonlinear Sci. Appl. 3, 123–143 (2010) 211. Najati, A., Rahimi, A.: A fixed point approach to the stability of a generalized Cauchy functional equation. Banach J. Math. Anal. 2, 105–112 (2008) 212. Najati, A., Ranjbari, A.: Stability of homomorphisms for a 3D Cauchy-Jensen type functional equation on C -ternary algebras. J. Math. Anal. Appl. 341, 62–79 (2008) 213. Najati, A., Rassias, Th.M.: Stability of homomorphisms and .; /-derivations. Appl. Anal. Discret. Math. 3, 264–281 (2009) 214. Najati, A., Kang, J.I., Cho, Y.J.: Local stability of the pexiderized Cauchy and Jensen’s equations in fuzzy spaces. J. Inequal. Appl. 2011, 78 (2011) 215. Najatim, A., Rassias, Th.M.: Stability of a mixed functional equation in several variables on Banach modules. Nonlinear Anal. 72, 1755–1767 (2010) 216. Nikoufar, E., Rassias, Th.M.: -centralizers on semiprime Banach *-algebras. Ukr. Math. J. 66, 300–310 (2014) 217. Novotný, P., Hrivnák, J.: On .˛; ˇ; /-derivations of Lie algebras and corresponding invariant functions. J. Geom. Phys. 58, 208–217 (2008) 218. O’Regan, D., Rassias, J.M., Saadati, R.: Approximations of ternary Jordan homomorphisms and derivations in multi-C* ternary algebras. Acta Math. Hung. 134, 99–114 (2012) 219. Pallu de la Barriére, R.: Algèbres unitaires et espaces d’Ambrose. Ann. Ecole Norm. Sup. 70, 381–401 (1953) 220. Park, C.: On the stability of the linear mapping in Banach modules. J. Math. Anal. Appl. 275, 711–720 (2002) 221. Park, C.: Linear functional equations in Banach modules over a C -algebra. Acta Appl. Math. 77, 125–161 (2003) 222. Park, C.: Modified Trif’s functional equations in Banach modules over a C -algebra and approximate algebra homomorphisms. J. Math. Anal. Appl. 278, 93–108 (2003) 223. Park, C.: On an approximate automorphism on a C -algebra. Proc. Am. Math. Soc. 132, 1739–1745 (2004) 224. Park, C.: Lie -homomorphisms between Lie C -algebras and Lie -derivations on Lie C algebras. J. Math. Anal. Appl. 293, 419–434 (2004) 225. Park, C.: Approximate homomorphisms on JB -triples. J. Math. Anal. Appl. 306, 375–381 (2005) 226. Park, C.: Homomorphisms between Lie JC -algebras and Cauchy-Rassias stability of Lie JC -algebra derivations. J. Lie Theory 15, 393–414 (2005) 227. Park, C.: Homomorphisms between Poisson JC -algebras. Bull. Braz. Math. Soc. 36, 79–97 (2005) 228. Park, C.: Isomorphisms between unital C -algebras. J. Math. Anal. Appl. 307, 753–762 (2005)

336

References

229. Park, C.: Hyers-Ulam-Rassias stability of a generalized Euler-Lagrange type additive mapping and isomorphisms between C -algebras. Bull. Belg. Math. Soc.–Simon Stevin 13, 619–631 (2006) 230. Park, C.: Isomorphisms between C -ternary algebras. J. Math. Phys. 47(10), 103512 (2006) 231. Park, C.: Isomorphisms between C -ternary algebras. J. Math. Anal. Appl. 327, 101–115 (2007) 232. Park, C.: Fixed points and Hyers-Ulam-Rassias stability of Cauchy-Jensen functional equations in Banach algebras. Fixed Point Theory Appl. 2007, Article ID 50175 (2007) 233. Park, C.: Hyers-Ulam-Rassias stability of homomorphisms in quasi-Banach algebras. Bull. Sci. Math. 132, 87–96 (2008) 234. Park, C.: Generalized Hyers-Ulam stability of functional equations: a fixed point approach. Taiwan. J. Math. 14, 1591–1608 (2010) 235. Park, C.: Square root and 3rd root functional equations in C -algebras: an fixed point approach. J. Nonlinear Anal. Optim. 2, 27–34 (2011) 236. Park, C.: Orthogonal stability of a cubic-quartic functional equation. J. Nonlinear Sci. Appl. 5(Special issue), 28–36 (2012) 237. Park, C.: Stability of bi- -derivations on JB -triples: revisited. Int. J. Geom. Methods Mod. Phys. 11(3), 1450015, 10 pp. (2014) 238. Park, C., An, J.S.: Isomorphisms in quasi-Banach algebras. Bull. Korean Math. Soc. 45, 111–118 (2008) 239. Park, C., Cui, J.: Generalized stability of C -ternary quadratic mappings. Abstr. Appl. Anal. 2007, Article ID 23282 (2007) 240. Park, C., Hou, J.: Homomorphisms between C -algebras associated with the Trif functional equation and linear derivations on C -algebras. J. Korean Math. Soc. 41, 461–477 (2004) 241. Park, C., Najati, A.: Homomorphisms and derivations in C -algebras. Abstr. Appl. Anal. 2007, Article ID 80630 (2007) 242. Park, C., Park, W.: On the Jensen’s equation in Banach modules. Taiwan. J. Math. 6, 523–531 (2002) 243. Park, C., Park, J.: Generalized Hyers-Ulam stability of an Euler-Lagrange type additive mapping. J. Differ. Equ. Appl. 12, 1277–1288 (2006) 244. Park, C., Rassias, Th.M.: Hyers-Ulam stability of a generalized Apollonius type quadratic mapping. J. Math. Anal. Appl. 322, 371–381 (2006) 245. Park, C., Rassias, Th.M.: Fixed points and generalized Hyers-Ulam stability of quadratic functional equations. J. Math. Inequal. 1, 515–528 (2007) 246. Park, C., Rassias, Th.M.: Homomorphisms between JC*-algebras. Studia Univ. “Babes– Bolyai”, Math. 53, 43–55 (2008) 247. Park, C., Rassias, Th.M.: Homomorphisms in C -ternary algebras and JB -triples. J. Math. Anal. Appl. 337, 13–20 (2008) 248. Park, C., Rassias, Th.M.: Homomorphisms and derivations in proper JCQ -triples. J. Math. Anal. Appl. 337, 1404–1414 (2008) 249. Park, C., Rassias, Th.M.: On the stability of orthogonal functional equations. Tamsui Oxf. J. Math. Sci. 24, 355–365 (2008) 250. Park, C., Rassias, J.M.: Stability of the Jensen-type functional equation in C -algebras: a fixed point approach. Abstr. Appl. Anal.2009, Article ID 360432, 17 pp. (2009) 251. Park, C., Rassias, Th.M.: Fixed points and stability of functional equations. In: Pardalos, P.M., Rassias, Th.M., Khan, A.A. (eds.), Nonlinear Analysis and Variational Problems. Springer Optimization and Its Applications, vol. 35, pp. 125–134. Springer, New York (2010) 252. Park, C., Saadati, R.: Approximation of a generalized additive mapping in multi-Banach modules and isomorphisms in multi-C -algebras: a fixed-point approach. Adv. Differ. Equ. 2012, 162, 14 pp. (2012) 253. Park, C., Cho, Y., Han, M.: Stability of functional inequalities associated with Jordan–von Neumann type additive functional equations. J. Inequal. Appl. 2007, Article ID 41820 (2007) 254. Park, C., An, J.S., Moradlou, F.: Additive functional inequalities in Banach modules. J. Inequal. Appl. 2008, Article ID 592504, 10 pp. (2008)

References

337

255. Park, C., Boo, D.H., Rassias, Th.M.: Approximately additive mappings over p-adic fields. J. Chungcheong Math. Soc. 21, 1–14 (2008) 256. Park, C., Lee, J.R., Shin, D.Y.: Fixed points and stability of functional equations associated with inner product spaces. In: Difference Equations and Applications, pp. 235–242, UgurBahesehir University Publishing Company, Istanbul (2009) 257. Park, C., Lee, J.R., Rassias, Th.M., Saadati, R.: Fuzzy -homomorphisms and fuzzy -derivations in induced fuzzy C -algebras. Math. Comput. Model. 54, 2027–2039 (2011) 258. Park, C., Kenary, H.A., Kim, S.O.: Positive-additive functional equations in C -algebras. Fixed Point Theory 13, 613–622 (2012) 259. Popa, D., Rasa, I.: On the Hyers-Ulam stability of the linear differential equation. J. Math. Anal. Appl. 381, 530–537 (2011) 260. Popa, D., Rasa, I.: Hyers-Ulam stability of the linear differential operator with nonconstant coefficients. Appl. Math. Comput. 291, 1562–1568 (2012) 261. Popa, D., Rasa, I.: The Fréchet functional equation with application to the stability of certain operators. J. Approx. Theory 164, 138–144 (2012) 262. Popovych, R., Boyko, V., Nesterenko, M., Lutfullin, M.: Realizations of real low–dimensional Lie algebras. J. Phys. A 36, 7337–7360 (2003) 263. Pourpasha, M.M., Rassias, Th.M., Saadati, R., Vaezpour, S.M.: The stability of some differential equations. Math. Probl. Eng. 2011, Article ID 128479, 15 pp. (2011) 264. Prastaro, A., Rassias, Th.M.: Ulam stability in geometry of PDE’s. Nonlinear Funct. Anal. Appl. 8, 259–278 (2003) 265. Radu, V.: The fixed point alternative and the stability of functional equations. Fixed Point Theory 4, 91–96 (2003) 266. Rand, D., Winternitz, P., Zassenhaus, H.: On the identification of Lie algebra given by its structure constants I. Direct decompositions, Levi decompositions and nil radicals. Linear Algebra Appl. 109, 197–246 (1988) 267. Rassias, Th.M.: On the stability of the linear mapping in Banach spaces. Proc. Am. Math. Soc. 72, 297–300 (1978) 268. Rassias, J.M.: On approximation of approximately linear mappings by linear mappings. J. Funct. Anal. 46, 126–130 (1982) 269. Rassias, J.M.: On approximation of approximately linear mappings by linear mappings. Bull. Sci. Math. 108, 445–446 (1984) 270. Rassias, Th.M.: New characterizations of inner product spaces. Bulletin des Sciences Mathematiques, 2ed serie, Paris 108, 95–99 (1984) 271. Rassias, Th.M.: On the stability of mappings. Rendiconti del Seminario Matematico e Fisico di Milano 58, 91–99 (1988) 272. Rassias, J.M.: Solution of a problem of Ulam. J. Approx. Theory 57, 268–273 (1989) 273. Rassias, Th.M.: Topics in Mathematical Analysis. A Volume Dedicated to the Memory of A. L. Cauchy. World Scientific Publishing Company, Singapore/Teaneck/London (1989) 274. Rassias, Th.M.: Problem 16: 2. Report of the 27th international symposium on functional equations. Aequationes Mathematicae, Bielsko-Bia la, Poland 39, 292–293 (1990) 275. Rassias, Th.M.: On a modified Hyers-Ulam sequence. J. Math. Anal. Appl. 158, 106–113 (1991) 276. Rassias, Th.M.: Inner Product Spaces and Applications. Pitman Research Notes in Mathematics Series, No. 376. Addison Wesley Longman, Essex (1997) 277. Rassias, Th.M.: Nonlinear Mathematical Analysis and Applications. Hadronic Press, Palm Harbor (1998) 278. Rassias, Th.M.: Approximation Theory and Applications. Hadronic Press, Palm Harbor (1998) 279. Rassias, Th.M.: On the stability of the quadratic functional equation and its applications. Studia Univ. Babes-Bolyai XLIII, 89–124 (1998) 280. Rassias, Th.M.: New Approaches in Nonlinear Analysis. Hadronic Press, Palm Harbor (1999) 281. Rassias, Th.M.: Functional Equations and Inequalities. Kluwer Academic Publishers, Dordrecht/Boston/London (2000)

338

References

282. Rassias, Th.M.: Survey on Classical Inequalities. Kluwer Academic Publishers, Dordrecht/Boston/London (2000) 283. Rassias, Th.M.: The problem of S. M. Ulam for approximately multiplicative mappings. J. Math. Anal. Appl. 246, 352–378 (2000) 284. Rassias, Th.M.: On the stability of functional equations in Banach spaces. J. Math. Anal. Appl. 251, 264–284 (2000) 285. Rassias, Th.M.: On the stability of functional equations and a problem of Ulam. Acta Appl. Math. 62, 23–130 (2000) 286. Rassias, Th.M.: On the stability of functional equations originated by a problem of Ulam. Mathematica 44, 39–75 (2002) 287. Rassias, Th.M.: On the stability of minimum points. Mathematica 45, 93–104 (2003) 288. Rassias, Th.M.: Functional Equations, Inequalities and Applications. Kluwer Academic Publishers, Dordrecht/Boston/London (2003) 289. Rassias, Th.M.: Handbook of Functional Equations – Functional Inequalities. Springer, New York (2014) 290. Rassias, Th.M.: Handbook of Functional Equations – Stability Theory. Springer, New York (2014) 291. Rassias, Th.M., Brzdek, J.: Functional Equations in Mathematical Analysis – In Honor of S. M. Ulam. Springer, New York (2012) 292. Rassias, Th.M., Šemrl, P.: On the behavior of mappings which do not satisfy Hyers–Ulam stability. Proc. Am. Math. Soc. 114, 989–993 (1992) 293. Rassias, Th.M., Šemrl, P.: On the Hyers-Ulam stability of linear mappings. J. Math. Anal. Appl. 173, 325–338 (1993) 294. Rassias, Th.M., Shibata, K.: Variational problem of some quadratic functionals in complex analysis. J. Math. Anal. Appl. 228, 234–253 (1998) 295. Rassias, Th.M., Tabor, J.: What is left of Hyers–Ulam stability? J. Nat. Geom. 1, 65–69 (1992) 296. Rassias, Th.M., Tabor, J.: Stability of Mappings of Hyers–Ulam Type. Hadronic Press, Palm Harbor (1994) 297. Rassias, Th.M., Toth, L.: Topics in Mathematical Analysis and Applications. Springer, New York (2014) 298. Rätz, J.: On approximately additive mappings. In: Beckenbach, E.F. (ed.), General Inequalities 2. International Series of Numerical Mathematics, vol. 47, pp. 233–251. Birkhäuser, Basel (1980) 299. Ravi, K., Thandapani, E., Senthil Kumar, B.V.: Solution and stability of a reciprocal type functional equation in several variables. J. Nonlinear Sci. Appl. 7, 18–27 (2014) 300. Rolewicz, S.: Metric Linear Spaces. PWN-Polish Scientific Publishers/Reidel, Dordrecht (1984) 301. Saadati, R., Sadeghi, Gh.: Approximate homomorphisms and derivations in proper JCQ*triples via a fixed point method. Expo. Math. 31, 87–97 (2013) 302. Saadati, R., Sadeghi, Gh., Rassias, Th.M.: Approximate generalized additive mappings in proper multi–CQ –algebras. Filomat 28 , 677–694 (2014) 303. Saadati, R., Cho, Y.J., Vahidi, J.: The stability of the quartic functional equation in various spaces. Comput. Math. Appl. 60, 1994–2002 (2010) 304. Saadati, R., Rassias, Th.M., Cho, Y.J.: Approximate .˛; ˇ; /-derivation on random Lie C algebras. RACSAM 109, 1–10 (2015) 305. Sahoo, P.K., Riedel, T.: Mean Value Theorem and Functional Equations. World Scientific Publishing Company, Singapore/River Edge/London/Hong Kong (1998) 306. Schwaiger, J.: Remark 12. Report on the 25th international symposium on functional equations. Aequationes Mathematicae, Hamburg-Rissen, Germany 35, 120–121 (1985) 307. Šemrl, P.: On Jordan -derivations and an application. Colloq. Math. 59, 241–251 (1990) 308. Šemrl, P.: Quadratic functionals and Jordan -derivations. Stuidia Math. 97, 157–163 (1991) 309. Sewell, G.L.: Quantum Mechanics and Its Emergent Macrophysics. Princeton University Press, Princeton/Oxford (2002)

References

339

310. Shilkret, N.: Non-Archimedian Banach algebras. Ph.D. thesis, Polytechnic University, ProQuest LLC (1968) 311. Shulman, E.V.: Group representations and stability of functional equations. J. Lond. Math. Soc. 54, 111–120 (1996) 312. Skof, F.: Proprietà locali e approssimazione di operatori. Rend. Sem. Mat. Fis. Milano 53, 113–129 (1983) 313. Smital, J.: On Functions and Functional Equations. Adam Hilger, Bristol/Philadelphia (1988) 314. Streater, R.F., Wightman, A.S.: PCT, Spin and Statistics and All That. Benjamin, New York (1964) 315. Sundaresan, K.: Orthogonality and nonlinear functionals on Banach spaces. Proc. Am. Math. Soc. 34, 187–190 (1972) 316. Takhtajan, L.: On foundation of the generalized Nambu mechanics. Commun. Math. Phys. 160, 295–315 (1994) 317. Thirring, W., Wehrl, A.: On the mathematical structure of the B.C.S.-model. Commun. Math. Phys. 4, 303–314 (1967) 318. Trapani, C.: Quasi--algebras of operators and their applications. Rev. Math. Phys. 7, 1303–1332 (1995) 319. Trapani, C.: Some seminorms on quasi--algebras. Studia Math. 158, 99–115 (2003) 320. Trapani, C.: Bounded elements and spectrum in Banach quasi -algebras. Studia Math. 172, 249–273 (2006) 321. Ulam, S.M.: A Collection of Mathematical Problems. Interscience Publishers, New York (1960) 322. Ulam, S.M.: Problems in Modern Mathematics. Wiley, New York (1960) 323. Upmeier, H.: Jordan Algebras in Analysis, Operator Theory, and Quantum Mechanics. Regional Conference Series in Mathematics No. 67. American Mathematical Society, Providence (1987) 324. Urs, C.: Ulam-Hyers stability for coupled fixed points of contractive type operators. J. Nonlinear Sci. Appl. 6, 124–136 (2013) 325. Vainerman, L., Kerner, R.: On special classes of n-algebras. J. Math. Phys. 37, 2553–2565 (1996) 326. Villena, A.R.: Derivations on Jordan–Banach algebras. Studia Math. 118, 205–229 (1996) 327. Wang, G., Zhou, M., Sun, L.: Hyers-Ulam stability of linear differential equations of first order. Appl. Math. Lett. 21, 1024–1028 (2008) 328. Wang, Z., Li, X., Rassias, Th.M.: Stability of an additive-cubic-quartic functional equation in multi-Banach spaces. Abstr. Appl. Anal. 2011, Article ID 536520, 11 pp. (2011) 329. Wyrobek, W.: Orthogonally additive functions modulo a discrete subgroup. Aequ. Math. 78, 63–69 (2009). Springer, New York (2009) 330. Zariski, O., Samuel, P.: Commutative Algebra. Van Nostrand, Princeton (1958) 331. Zeidler, E.: Nonlinear Functional Analysis and Its Applications I: Fixed-Point Theorems. Springer, New York (1986) 332. Zettl, H.: A characterization of ternary rings of operators. Adv. Math. 48, 117–143 (1983)

Index

Symbols .˛; ˇ; /-derivation, 124 .˛; ˇ; /-derivations, 52 -derivation, 11 -homomorphisms, 53 CQ -algebra, 16 C -algebra, 10 C -ternary 3-derivation, 207 C -ternary 3-homomorphism, 204 C -ternary algebra, 12 C -ternary algebra homomorphism, 12 C -ternary derivations, 216 C -ternary homomorphisms, 215 JB -algebra, 152 JB -triple, 13 JB -triple derivation, 13 JB -triple derivations, 218 JB -triple homomorphism, 13 JB -triple homomorphisms, 217 JB -triples, 220 JC -algebra, 64, 69, 152 JC -algebra homomorphism, 69 JC -algebras, 53, 151 J -triple, 13 -additive, 5 p-Banach algebra, 10, 36 p-Banach space, 10 p-adic, 14 p-norm, 10

A additive function, 1 additive functional equation, 51 additive functional inequality, 171, 178

anticommutator product, 64, 152 Aoki–Rolewicz theorem, 10 Apollonius type additive functional equation, 51, 63, 202, 209 Apollonius type additive mapping, 51, 63, 210 Apollonius’ identity, 63

B Banach algebra, 9 Banach contraction principle, 7 Banach fixed point theorem, 7 bi- -derivation, 219 bi- -derivations, 202, 219 biadditive function, 1

C Cauchy difference, 3 Cauchy functional equation, 1 Cauchy-Jensen functional equations, 20, 38 Cauchy-Rassias inequality, 53, 57, 60 complete non-Archimedean normed space, 14 complex normed algebra, 9

D derivation, 9–11 derivations on Banach algebras, 23 direct method, 3, 20

F fixed point method, 303

© Springer International Publishing Switzerland 2015 Y.J. Cho et al., Stability of Functional Equations in Banach Algebras, DOI 10.1007/978-3-319-18708-2

341

342 G generalized additive mapping, 85, 87, 91 generalized complete metric space, 8 generalized derivation, 36, 44 generalized Leibniz rule, 36 generalized metric, 8, 25, 143

H homomorphism, 9, 10, 26 Hyers-Ulam stability, 19, 24, 29, 31, 32, 36, 38, 44, 51–53, 85, 104, 110, 124, 165, 203 Hyers-Ulam stability of homomorphisms, 24 Hyers-Ulam-Rassias stability, 6

I involution, 10 isomorphisms, 20

J Jensen type functional equation, 51, 70 Jensen’s equation, 54 Jordan -derivation, 52, 110, 112, 114 Jordan algebra, 64, 152 Jordan derivation, 69 Jordan triple product, 152 Jordan triple system, 12 Jordan–von Neumann, 64 Jordan–von Neumann functional inequality, 63 Jordan-von Neumann functional equation, 192

L Lie C -algebra, 11, 64, 67 Lie C -algebra homomorphism, 11, 67 Lie C -algebras, 51 Lie derivation, 11, 68

M maximum multi-norm, 15 minimum multi-norm, 15 modulus of concavity, 9 multi-C ternary algebras, 229, 241 Multi-C -algebra, 16 multi-C -ternary algebra, 17 multi-CQ -algebra, 17 multi-Banach algebra, 16 multi-Banach space, 15 multi-convergent, 15 multi-norm, 14

Index multi-normed algebra, 16 multi-normed space, 15

N non-Archimedean C -algebras, 303 non-Archimedean Banach algebra, 14 non-Archimedean field, 13 non-Archimedean Lie C -algebra, 316 non-Archimedean Lie C -algebra homomorphism, 316 non-Archimedean Lie derivation, 318, 326 non-Archimedean norm, 13 non-Archimedean normed space, 13 non-Archimedean normed spaces, 303 normed A-module, 185 normed modules, 186

O observables, 152

P positive, 143 positive-additive functional equation, 144 positive-additive mapping, 144 proper CQ -algebra homomorphism, 11 proper CQ -algebras, 11, 16, 166 proper multi-CQ -algebras, 230, 264

Q quadratic functional equation of Apollonius type, 63, 210 quadratic Jordan -derivations, 115 quadratic mapping of Apollonius type, 63 quasi--algebra, 16 quasi-Banach algebra, 10, 35 quasi-Banach algebras, 19, 31 quasi-Banach space, 9 quasi-norm, 9 quasi-normed algebra, 10 quasi-normed space, 9

S square root functional equation, 52, 130 square root mapping, 131 stable, 3, 7 superstability, 52, 110, 124, 202, 209 superstable, 7

Index T theorem of the alternative, 8 topological, 16 trivial valuation, 13

343 U ultrametric, 13 unital C -algebra, 12, 51, 53 unital Banach algebra, 9

Stability of Functional Equations in Banach Algebras

Stability of functional differential equations

Stability of functional differential equations

Cohomology in Banach Algebras

Stability of Solutions of Differential Equations in Banach Space

Stability of Solutions of Differential Equations in Banach Space

Stability of solutions of differential equations in Banach space

Perturbations of Banach Algebras

Stability analysis of impulsive functional differential equations

Perturbations of Banach algebras

Continuous semigroups in Banach algebras

Hyers-Ulam- Rassias Stability of Functional Equations in Nonlinear Analysis

Hyers-Ulam-Rassias stability of functional equations in nonlinear analysis

Ultrametric Banach algebras

Algebraic equations in commutative Banach algebras and related questions

Banach algebras: An Introduction

Ultrametric Banach Algebras

Ultrametric Banach algebras

Linear Equations in Banach Spaces

Differential Equations in Banach Spaces

Differential equations in Banach spaces

Differential equations in Banach spaces

M-Ideals in Banach Spaces and Banach Algebras

Banach Algebras and the General Theory of *-Algebras: Volume 1, Algebras and Banach Algebras (Encyclopedia of Mathematics and its Applications)

A Course in Commutative Banach Algebras

Functional equations

Banach Algebras and the General Theory of *-Algebras: Volume 1, Algebras and Banach Algebras (Encyclopedia of Mathematics and its Applications)

Banach Algebras and the General Theory of *-Algebras: Volume 1, Algebras and Banach Algebras (Encyclopedia of Mathematics and its Applications)

Mathematics - A Course In Commutative Banach Algebras

A course in commutative Banach algebras

Stability of nonautonomous differential equations

Stability of Functional Equations in Banach Algebras

Stability of functional differential equations

Stability of functional differential equations

Cohomology in Banach Algebras

Stability of Solutions of Differential Equations in Banach Space

Stability of Solutions of Differential Equations in Banach Space

Stability of solutions of differential equations in Banach space

Perturbations of Banach Algebras

Stability analysis of impulsive functional differential equations

Perturbations of Banach algebras

Continuous semigroups in Banach algebras

Hyers-Ulam- Rassias Stability of Functional Equations in Nonlinear Analysis

Hyers-Ulam-Rassias stability of functional equations in nonlinear analysis

Ultrametric Banach algebras

Algebraic equations in commutative Banach algebras and related questions

Banach algebras: An Introduction

Ultrametric Banach Algebras

Ultrametric Banach algebras

Linear Equations in Banach Spaces

Differential Equations in Banach Spaces

Differential equations in Banach spaces

Differential equations in Banach spaces

M-Ideals in Banach Spaces and Banach Algebras

Banach Algebras and the General Theory of *-Algebras: Volume 1, Algebras and Banach Algebras (Encyclopedia of Mathematics and its Applications)

A Course in Commutative Banach Algebras

Functional equations

Banach Algebras and the General Theory of *-Algebras: Volume 1, Algebras and Banach Algebras (Encyclopedia of Mathematics and its Applications)

Banach Algebras and the General Theory of *-Algebras: Volume 1, Algebras and Banach Algebras (Encyclopedia of Mathematics and its Applications)

Mathematics - A Course In Commutative Banach Algebras

A course in commutative Banach algebras

Stability of nonautonomous differential equations

Recommend Documents