Lecture Notes in
Physics
Edited by J. Ehlers, MLinchen, K. Hepp, Zerich, H. A. Weidenm~ller, Heidelberg, and J. Zitta...
212 downloads
652 Views
7MB 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
Lecture Notes in
Physics
Edited by J. Ehlers, MLinchen, K. Hepp, Zerich, H. A. Weidenm~ller, Heidelberg, and J. Zittartz, K61n Managing Editor: W. BeiglbSck, Heidelberg
47 Pade Approximants Method and Its Applications to Mechanics
Edited by H. Cabannes
Springer-Verlag Berlin. Heidelberg. New York 197 6
Editor Prof. Henri Cabannes Universit@ Pierre et Marie Curie Mecanique Theorique Tour 66, 4, Place Jussieu 7 5 0 0 5 Paris/France
Library of Congress Cataloging in Publication Data
Main entry under title: Fade approximants method and its applications .to mechanics. (Lecture notes in physics ; 47) Bibliography: p. Includes index. I. Pade approximant. 2. Fluid mechanics. I. Cabannes, Henri. II. Series. QC20.7.P3P35 532' .01' 515 75-46504
ISBN 3-540-07614-X Springer-Verlag Berlin - Heidelberg • New York ISBN 0-387-07614-X Springer-Verlag New York • Heidelberg • Berlin This ,,york is subject to copyright, All rights are reserved, whether the whole or part of the material is concerned, specifically those of translation, re* printing, re-use of illustrations, broadcasting, reproduction by photocopying machine or similar means, and storage in data banks. Under § 54 of the German Copyright Law where copies are made for other than private use, a fee is payable to the publisher, the amount of the fee to be determined by agreement with the publisher. © by Springer-Verlag Berlin • Heidelberg 1976 Printed in Germany Printing and binding: Offsetdruckerei Beltz, Hemsbach/Bergstr.
Henri Pad~
(1863 - 1953)
Henri Euggne Pad~ was born in Abbeville
(France) on Dec.
17th 1863.
Admitted in 1883 to the Ecole Normale Supgrieure, he left it in 1886 with the highest teacher's degree
(Agr~gation)
in Mathematics.
After teaching at the classi-
cal secondary school in Limoges, Carcassonne and Montpellier, in 1889 in order to study in Germany,
first in Leipzig,
he was granted a leave
then in G~ttingen.
On June
21st 1892, before the University of Paris, he defended his doctorate thesis on the approximate representation of a function by rational fractions. Henri Pad~ was appointed lecturer at the Faculty of Sciences of Lille in 1897, Professor of Rational and Applied Mechanics at the Faculty of Sciences of Poitiers
in 1902 and Professor of Mechanics at the Faculty of Bordeaux in 1903. In
1906, he was elected Dean of the Faculty of Sciences of Bordeaux and became Laureate with the major prize for mathematical on the report of Emile Picard.
sciences awarted by the Academy of Sciences
In 1908 he was named Rector of the Academy of
Besan~on, he was then the youngest rector in France.
In 1917 he became rector of the
Academy of Dijon and in 1923 rector of the Academy of Aix-Marseilles, kept until he retired in 1934. Henri Pad~ died in 1953 at the age of 89.
an office he
Henri
Pad~
(1863
-
1953)
/
•
~
i ¸f
.f~'~o~
~.-""~INISTI~ftE • . .
f
-~":,,, ii,.9'~
h,
~Z~f '
REPUBLIQUE .....
',~"4" \'~
),.~
5'~-AN GA I S E
L'INSTttUCTION PUBLIQ~f,~ e*:x~;'~Gbecause
E~N)(~)
the moments
J~> we choose,
~k are dependent
.
(11-4)
More precisely we have 4)
~N)(q~)
(11-5)
I~> In that way, we generate interesting
variational
principles
fact is that if we now increase E (N+I) I % L ~)
In other words, nal principle
Ei
and E (2) are the ground state eigenvalue O
O
and eigenvector of P2 H P2, where P2 is the projector onto the two-dimensional space ~(2) spanned by the vectors
I~> and HI~ > . This remark can be completely
generalized, and is linked to the fact that the Lanczos method fer matrices, and its generalization to Hilbert space the tri-diagonalisation Jacobi method, are deeply linked with the theory of Pad~ Approximation
6)
.
To end up this section we give the formulae generalizing (IV-2) for the N th approximation. Introducing the normalized polynomials
:
~O.....~N
~(E,N)
A(E,N)
>
0.
(iv-4)
~/GNGN_ I
~ . . . . . g2N The L th approximated eigenfunction at order N is given by : N-I
(Iv-5) L
j=O
The
I~-N$ are obtained by diagonalizing the Hermitian matrix ILl =
PN H PN
(IV-6)
25 where PN is the projector onto the N-dimensional space spanned by the vectors I~>, HI~>,..o HN'I]~>. While the
~L(N)\ar / e orthogonal but not normalized,the
of orthonormal vectors because it is known that the
A(H,j) I~> are a set
A(E,N) form a set of ortho-
gonalized polynomials with respect to the spectral measure of the operator H 5).
V - APPLICATION TO A SYSTEM OF N FERMIONS. To apply effectively the Pad4-Ritz principle (IV-5) it is necessary to find a way to compute easily the moments
~k
"
~o
(v-l)
In the Z-body problem the moments are given Z-uple integrals which are difficult or even impossible to compute if k is large. However, there are cases in which the moments can be reduced to entirely algebraic expressions. An example is given in ref. (5)~ for the case of the most general d-dimensional anharmonic oscillator, by choosing a suitable variational test vector I~>
•
Here we want to consider the case of Z fermions interacting via a two body Z potential : Z ~ 2 ~ ~pi)2 HZ
"
~ i =I
P2 ~
+ i
I i<jsZ
~ ~ V( Iri'rj ])
i-I 2M
(V-2)
where we have subtracted the center of mass energy. We shall suppose that V, the two-body potential is a superposition of Gaussian potentials : .
V( l~i-~j I)
"
OlS
2
Pp ([~i-~j ]2)
e- ~P(~i-~J)
(V-B)
p=l (where
P (x) is a polynomial in x~With such a superposition, we can build up a P large class of phenomenological potentials, The interesting case of hard cores will be postponed to a further section. We must now choose the test vector in such a way that I) The Z-uple integral reduces to an algebraic calculation. ..... ~ - - ~ - ~ - ~ 5 ~ z 9 - y ~ - ~ 5 5 ~ 5 ~ - ~ - ~ [ ~ 9 ~ 9 - ~ 5 - ~ - - 5 ~ z - f ~
illed.
~]~> is a given vector~ by using "m~trix" Pad4 Approxlment most of the results exposed here generalize to the case where ]~> is a set of vectors I~i>, ]~2> o.o l~R>, generating a "model" space.
2B We shall choose i~> to be an arbitrary sum of Slater determinants built up with the states of the harmonic oscillator. -+
-+
-~
-+
%0o(rl, t) o..go(rz, t) -+
P I
-+
o ,
be denoted by Yi"
We now describe a method
f o r computing Yi' f o r i = 1 , 2 ..... M.
Assume that Yi is Known and that the values of the derivatives exist for k > o .
Then,consider
s i [ h ] =Yi + h ' f [ x i ) Pi Let r. = - - b e z qi
f[k)[x i]
the series s. defined as follows l
h2 h3 f " [ x _ ) +~T!'f'(xi) +~!' i + ""
the Pad@ apprcximant
of s, of order [m,n]. z
[2)
This implies
that r i is an element of R m and that n
si[h).qi[h) -pi[h] =OChm+n+ki+lj
[3]
for some integer value of K i, which is as high as possible.
If qi[xi+1) # o then the value of Yi+1 is defined as follows
Yi+l =ri[xi+l]
"
(4)
Since Yo is Known and since r i exists for every i, this technique allows us to compute yl,y 2, ....YM'
The value YM can be considered
as an approximate value for y[b) or I. section that lim YM =I h~ o
It will be seen in the next
if certain conditions
ere satisfied.
73
A variant of the above technique has been used by P.J.S. Watson in [9], for the case where h = b - a
or M = I.
In the next exampie
we illustrate what Kind of formulas we get if [4) is
used
to
compute approximate vaiues for I.
Example
Let m =n = 1 then the Pad@ approximant r i of order (1,1)
of (2) is the rational function associated with the irreducible form of ~ q
where
f'(x.) p=yi.f[xi)
÷ h. [f2[xi) - Y i . T
]
and
~'[x i) q =f[x i ] -h.-----~----.
The formula [4) gives 2.f2(x.) Yi+q = Yi + h . 2 . f ( x i
) - h.f'(x 1 i)
for
i=0,1
. . . . . M-1
It is clear that YM is a nonlinear combination of values of f and its first derivative at the points x,. z As a consequence of [5) we also get the following formula for approximate integration between x i and xi+ 1 :
Xi+l
2.f2(x. )
1
fxl f[t).dt ~ h.2.f(xl ) _h.f,(xi )
(5)
74
4. Convergence
properties
In this section some convergence
properties
of the method described
in section 3 will be given. Consider any one-step method of the following form
Yi+l =yi + h'g(xi'h)
for i=0,1
for computing a solution of (I) numerically.
. . . . . M-1
(6)
Assume that g(x,h) is
defined and continuous for every (x,h) in [a,b] x [o,h o] , where h ° is some positive real number.
Under this condition
the following
result can be proved.
Theorem 1.
Let Yi be d e f i n e d by (6), then
x in
[a,~
if
and o n l y i f
lim
h~ o x=xi
Yi =y(x) f o r every
g[x,o) =f(x).
This property is a special case of a theorem about the convergence o% one-step methods for the numerical equations
solution of ordinary differential
Csee [4 ,p.?1).
Due to the definition
of r. in section 3, it is clear that l
be written in the form [6) with g{x,h) = f { x ) + h .
g(x,h) = [ri(x) -yi ] /h or
f'(x] + 3 ~~ . f"(x) .....
g[x,o) = f i x ) , consequently
[4) can
This implies that
theorem I can be applied and
~yM
=y{b)"
About the order of convergence we can prove the following result, the case where K. =o in (3) for i =0,1,2 .....M-1. l the case of normal Pad~ approximants.
for
This case is called
75
Let r. be the normal Pad@ approximant l
Theorem 2.
of s. of order &
[m,n) and Y i + l be d e f i n e d by [ 4 ) , then Y ( X i + l ) - Y i + l ~O(hm+n÷1) as h ~ o ,
f o r i = 0 , 1 , 2 . . . . . M-1.
A proof of this theorem is given in [10].
5. One-step methods w i t h o u t usin~ d e r i v a t i v e s
Consider any one-step method of the form (6) for computing of (I).
In order to find the value of g[xi,h)
that derivatives
of f must be computed.
it might be possible
This is e.g. the case if (6)
is derived by using the method in section 3. with m + n > l . compute derivatives less interesting. another expression,
a solution
of the integrend can be compllcated
The need to and numerically
Therefore one could try to replace g(xl,h) without derivatives,
the same order of convergence.
in (6) by
hoping to Keep a method with
This technique
has successfully
been
applied in some cases.
If e.g. the derivative quotient,
in (5) is replaced
by its forward difference
we get
2.f2(x.) 1 Yi+l = Yi + h • 3.f(xi ) _ f(Xi+l 3
It
can be proved (see [10]) t h a t t h i s
(7)
o n e - s t e p method has the same
o r d e r o f convergence as the method d e f i n e d by using ( 5 ) .
The f o r m u l a
(7) can a l s o be c o n s i d e r e d as a n o n l i n e a r method f o r a p p r o x i m a t e integration
of f between x i end xi+ 1 , namely as
Xi÷l I xi
f(t).dt
2.f2(Xl) ~ h.3.f(xi ) _f(xi+l
)
76
6. RemarKs
In [10] some numerical examples are given, illustrating the usefulness of the nonlinear techniques derived in this paper.
For "smooth"
integrands the classical linear methods give in general better results than the nonlinear techniques.
If the integrand has a pole near the
interval of integration than the nonlinear techniques can give better results.
Due to possible singularities in formulas of type C5) and
(7), care must be taken in applying these formulas.
Difficulties can
sometimes be avoided by a careful choice of the stepsize h.
Other nonlinear techniques for numerical integration are considered in [10 I.
Author's address Wuytack L. Department of Mathematics University of Antwerp Universiteitsplein I B-26]0 WILRJK (Belgium)
77
References
[I] CHISHOLM, J.S.R.: Applications of Pad6 approximation to numerical integration.
The Rocky Mountain Journal of Mathematics 4
C1972),159-167.
[2] CHISHOLM, J.S.R,; GENZ, A. and ROWLANOS, G.E.:
Accelerated
convergence of sequences of quadrature approximations. Journal of Computational Physics 10(1972),264-307.
[3] DAVIS, P.J.; RABINOWITZ, P.: Numerical integration. Publ. Co., London,
[41 FOX, L.:
Blaisdell
1967.
Romberg integration for a class of singular integrands.
The Computer Journal 10C1967),87-93.
[5] HENRICZ, P.:
Discrete variable methods in ordinary differential
equations.
[61 KAHANER,
O.K.:
John Wiley & Sons, New York, 1962.
Numerical quadrature by the s-algorithm.
Mathematics of Computation 26(1972),689-693.
[71LYNESS,J,N.;
NINHAM, B.W.:
expansions.
Numerical quadrature and asymptotic
Mathematics of Computation 21(1967),162-176.
[61 PIESSENS, R.; MERTENS,
I.; BRANDERS, M.:
Automatic integration
of functions having algebraic end point singularities. Angewandte InformatiK (1974), 65-66.
[91 WATSON, P.J.S.:
Algorithms for differentiation and Integration.
In "Pad~ approximants and their applications" MORRIS, P.R., editor),
[I01WUYTACK,
93-98.
(GRAVES-
Academic Press, London,1971.
L.: Numerical integration by using nonlinear techniques.
Journal of Computational and Applied Mathematics. To appear.
DETERMINATION
OF SHOCK WAVES
BY CONVERGENCE ACCELERATION by Pr.
Max BAUSSET
-
TOULON
(France)
-MAY 1 9 7 5 -
The difficulties
of the determination
of stationary detached shocks
arise from the global character of the problem to be solved as it is impossible to determine shock waves in the vicinity of a point. On the contrary for attached shocks determination These difficulties non-s~at~anary
is possible step by step from the vertex of the body. can be avoided by considering
flow. If a body situated in a motionless
so that the field of initial velocities causes a shock wave the determination
is not null,
shock problems
in a
fluid is set into motion
this motion immediately
of which in the vicinity of the initial
time is a local problem. The evolution of the shocks corresponding
to this motion in the vicinity
of the starting point of the body is presented here. On analytical of the stationary detached shocks waves related to an analytical be obtained by a process of convergence
accelaration,
representation
convex body can
then by passing on to the
limit when the permanent motion is reached. The data will be supposed to be such that operations
can be considered as possible within the fields where it is being
operated.
-
E~E~_~_~_~e_!~_~_~e~i~!~e_S!~i~
:
The space is related to orthonormated
fixed axes.
The equation of the body is : .%~ s h o c k
[Y (1)
x = f(yz) +
~(t)
oI
[.i o''x body
82
in which f is supposed to be uniform and the body occupies
the region
:
x ~
f (y z) +
~ (t).
The initial data are such that : (2)
$(O) = O
~' (0) >
0
This body is plunged into a non viscous compressible, supposedly perfect fluid. This motionless representing
respectively
the pressure,
and and
the density and ratio of specific heats.
The latter will be supposed to be constant throughout ristic quantities
non heat-conducting
is defined by quantities p, p
of this fluid (velocity,
the motion.
The characte-
pressure and density) will be desi-
gnated as V, p and p at the point of the spatio-temporal
coordinates
x, y, z,
and t. The fluid and the body are motionless before t = O. Since at the initial time
~' (0) > 0 the motion of the body immediately
propagates
(3)
itself through the fluid.
x =
F (y z t )
For reasons of calculations write (4)
causes a schock wave which
Its equation will be :
symmetry which will appear below,
let us
: ~ (x y z t) (x y
z t)
~
f (y z) +
=
F
so that when considering
(y z t)
$(t) - x - x
covector r = ] x, y, z ] the vectors normal at each
instant to any point of the body and the shock are defined by :
(5)
N,
= ~r *
=
N~
= ~r ~
=
[-1,
Fy,
F'zl
f'z In these conditions,
the normal number of Mach M at time t at any point
of the shock wave and the upstream number of Mach at infinite by :
32~ are defined
83
6t (6)
= _ _ ~'(0)
M (~) =
l -
C(~r~
c designating Let us write
x ~r~) 2
2 the speed of sound so that c p = yp at any point x, y, z, t. 2 (y + I) ~ = y - 1 and relate pressure and density to their va-
lours at infinite
introducing
P (x y z t) =
will be written
under
dr+ t
(7)
:
R (x y z t) =
the continuous
equations the form
motions
of dynamics,
of the considered
fluid are
mass and energy conservations
which
:
2 ! - ~ 2" R. ~ r P = O I +~
-2 c
dtR - R.. T r (~r v) = O
(I - 2 )
dt
dimensions
P
In these conditions, defined by the classic
without
desigmating
R. dtP
(I + 2 )
the corpuscular
The solution and the boundary
p. dt R = O
derivative
of shock problems
of limits
In the absence which
+
and Tr
the trace of matrix
~r V.
is a result of the study of this system
on the body and on the shock wave.
of viscosity,
leads to the equations
the body is necessarily
a stream surface,
:
(x y z t) = 0 (8)
6r~ x V~ = ~t ~
in which
V~
is the value of V on the body
The classic motion,
conditions
of shock phenomena
mass and energy are preserved
Theae usual conditions
~ = O.
while
can be expressed
mean that the quantities
crossing
the surface
by the equations
:
of
of the wave.
84
~ (xy
z t)
=0
(9) 6t#
-- 2
V~
~r ~ ~r ~ x ~r ~
P
+2 ~2
=
~r ~
--
'c
R~
= 21(
~2~r~-I
6t~
--
~/
2
I 7t° _ 2,)
+
2 (6 t *)
in which
V~ ) P#
and R~
are the values of V, P and R on shock ~
At any point of the spatio~temporal the shock and for t to equations
~ 0
equations
= 0.
region included between the body and
(7) are identities in x, y, z and t. Added
(8) and (9), they permit the calculation of the partial n-order
derivatives of quantities V, P, R and
~ at time t = 0.
In effect, if one place oneself at the initial time when the fluid is motionless everythere except on the body which is set into motion, the position of the shock wave coincides with the position of the body. Thus one has for t = 0 the relations
(lo)
:
~ (x y z o) =
which will be designated
~ (xy
~
=
z o)
F(y z 0) = f (y z)
~o"
O
They entail the following equations V~
= V o
: R~
$o
= o
R $o
(11) P$
o
= P
~o
~
r
~
o
= ~
r
~o
85
From the given equations (8) and (9) considered at time t = O one deduces the following relations :
6r~ °
x
= (6t~)t = O =
V~o
~'(O)
(12) 6r~ ° x V~
= - (| - 2 )
l~--]--, t - 0
o
c
0r~ ° x ~ J o
(6r¢ ° x
6r#o)
(6t~)t=O
If one notices that -.(6t$)t_O = F' (y z O) one deduces by elimination of V the t initial value of the shock velocity at any point :
(13)
F' (y z O)
t
~'(O)
+ (~
2(1_ 2 )
(O)
(1_~2)
)2
+
6r~ °
x
--
6r~ °
- Second order derivatives : . . . . . . . . . . . . . . . . . . . . . . . .
The total number of the partial n-order derivatives of a function A(x y z t) being C n n+3' one will be brought to consider the table : Jl
(14)
6n rn
6nA II = II 6xn
A ( x y z t )
~n A ;
6Y n
6nA ;
6zn
which will he a line or a matrix 3 x 3 according as A has a scalary or vectorial value. Relations (|l) then permit to calculate the spatial derivatives of (x y z t) for t = 0 in the form :
n
(15) ~r n (6r~ o) =
In
-~x - n (6r~o) ; ~ y
n
n
(6r# o) ; ~--~ (6r~ o)
1
from which the initial values of all the spatial partial derivatives of the function representing the shock wave :
(16)
6 p+q F (y z O) 6y p
6z q
=
dP+q f (~ z) 6y p
~z q
can be deduced. But concerning the temporal or spatio temporal derivatives of F it is necessary
86
to use the derivatives of the boundary of limits on the body and the shock. If one designates by ~ one differential which according to the case can be
~r'
~t' or dr, one deduces from equations (8) and (9) the following linear
equations in relation to
~(~t~), ~(~t~) and
~(~r~) :
+ ( x y z t) = 0
(17)
~r ~ x
6V#
=
6(~t ~) -
6(~r~) x
v~
(x y z t) = 0
V~ = ~. ~r ~
x
~(~t ¢)
+
~. 6r ¢
x
~r ¢ "
~(~r ¢)
+ Y "
~(~r ~)
(18)
P~ = el. ~(6t ~) +
81 •
~(~r ~)
R~ = e2" 6(6t~) +
82"
~(~r ~)
in which to following quantities which are dependent only on the first order derivatives of ~ that are known at the initial time are :
= - (I - 2),
~ 2 ~r ¢ x 6r¢ + (6t~)2 ~ #
x
8 = 2(I- 2),
~r ¢ "(6t#)2
~r ¢
y = (I- 2 ) 6r~
x
~r ~ .(6t~)2
6 ¢ t
~1 =
x 6r ~
.... -2 (6 r ~ x 6 ~¢)2 c
(19) 8L = - 2(|+ 2 ) c --2
Y2 =
2c2"
(dt~) 2 (~r ~ x ~r ~ )2
(| - 2 )
r (6t0) 2
. 6r ~
2 6r ~ x dr~ ~2 = -2 ~2(I- ~ ) (~t~)4
87
According to the choice of differential 6, equations lines and matrices
(|8) include the scaleries,
: t!
~t(~t ~) =
#t 2
~t(~r ¢)
(20) ~r co
(y Z O)
one can write : ~ ~
Ft2(Y z O) '~
f~r ~
3!a2+ 1
x ~J>
2|
(1 + p 2 ) , £ 1 ( f )
F't (y z O) ~
+ (1- !a2).~2(f )
One can then see that, for an analytical body, approximation (36) remains valid for t
> O. It is clear that such is not the case if the body's velocity is
~ndeterminate.
92
-
D~_t_ermination_of_d_et_a~_hed_s_ta_tio_~_a_~z_s_h~h~
:
It is possible to infer from the foregoing results approximate formular determining by the movements
the positions of the detached stationary shocks created
of blunt analytical bodies in translation motion.
In effect, when the second derivative
~"(t) vanishes
for t ->
the motion of the body tends to become unifo©m. It is then logical to admit that the motion of the fluid in relation to a mark linked to the body tends to become stationary. stationary
shock results from the knowledge
lary logical to admit that the stationary
Consequently
the position of a
of F(y z t) for t--~oo. It is simi-
flow is reached all the more rapidly
as the velocity limit of the body is reached more rapidly. One will place oneself in the case when is chosen so that : t
~< 0 :
$ (t) = O
t > : ~ (t) = t. ~' (0)
In any space of time when F(y z t) is an analytical variable dependent on time, one can write
: oo
(37)
F (y z t) ~
~
aj(yz).t j
j=O and one has indicated a calculation process of a. coefficients. J co ral methods allowing to replace the analytical function Ya.t j
There are seveby a sequence
L
of functions
converging towards F(yz)
in the field of
O j
convergence of the entire sequence a.t j and which are definite whatever 3 value of t is and which have a limit when time increases indefinitely. Within the framework of the
(38)
A
= n
n ~ 0
a.t j J
n ~
e-algorithm
the
theory supposes
0
is a converging sequence whose terms are dependent as parameters. Among others this theory proposes
to replace this sequence by another converging
more rapidly towards the same limit or converging
to a more extensive field
than the initial sequence. PADE'S defined by :
n-order diagonal approximation
fused in (36) for n = 17 is
93
t n 1A l ) ..............
t ° An
a 2 ) .................
an+ l
an+l
a2n
Pn
(t)
t °
Qn
(t)
tnAo
al
(39)
i
t
n
a1
an which
leads
in p a r t i c u l a r
Ql(t)
a2
................
an+ !
an+l
~ ................
a2n
~
lim
P (t) A n--5-----= (-I) n
t =o
Qn(t )
A
=
the c o f a c t o r
ao
,
ald ........
an
a2
~
a 3 ~ .......
an+ |
an
~
an+|) ......
a2n
of a
o
in
(41)
A
n
•
A.
As to SttANKS'S t r a n s f o r m a t i o n s , sequence
.....
2 ao(ala 4 - a~) - a 1 ( a l a 4 - a 2 a 3) + a 2 ( a l a 3 - a 2) 2 a2a 4 - a 3
Q2(t)
and 6 b e i n g
~
~..•......•
to
lim t =
:
t
a2
P2(t) (4O)
n-1
~
: 2 aoa 2 - a i
P1(t) lim t ~
with
~ ...............
In, n| A n =
the sequence
B
n
=
they
are
defined
:
An+l"
An-l-
An+l + A n - I
A2 n
- 2 An
n
>
1
by associating
to
94
The process
can be iterated by considering _
C
(42)
= Bn+| Bn-I
B2 n
n
the sequence
:
~ 2
n
Bn+ 1 + Bn_ 1 - 2B n and so forth. In each of these iterations t
->
the first term B! C 2 ... has a limit for
oo
lim
2 aoa 2 - a I
B I (t)
a2 (43)
2 aoa 2 - a I
lim C2(t )
22 (ala 3 - a2) 2 a4a 2 - a 3
a2 One can then see that PADE'S
a4 2 a2
approximation
If,I] and SHANKS'S
first
order B 1 have the same limit which we shall call first order approximation. On the contrary,
second order approximation
differ.
- ! l ~ ! ~ _ ~ l ~ 2 ~ _ ~ ! _ ~ ! ~ ! ~ l _ ~ h ~
When the studied the choice between
Concerning the stationary
method
numerical
which is the case here,
and another
results
the approximation
detached wave, point
Placing
is determinate,
an approximation
a comparison with accurate
an indefinite
function
:
relative
of function
the preceding
can only be made through to know particular
cases.
x = F(y z t) representing
calculations
have been conducted
to
(x, y z) only up to n = 2.
oneself
at a mark linked
to the body,
it follows
from (35)
and : _
(44)
F't (y z O) -
~2
M 2
+
1 -
~2
1
~'(0) = c.
~Sr f x 8 rf ~'2 M
that in the vicinity
(45)
of the initial
time one can write
F (y z t) = f (y z) + ~t -~2
M2 + 1 - -~2
:
~l~ + tSF, ' t2 (y z O)
M in which
the expression
of the second derivative
$"(O) = O. In these conditions gives a position
of the detached
PADE'S
follows
(or SHANKS'S)
stationary
shock
:
from (26) with
first order approximation
95 (46)
x = f(y z) -
2( ~2 M2+ I - ~2)2.(M2 + 1).d 4 M 2. A 8 . ~
2 (f) -
(~r f x 6r f)
I
M 2. (M 2 + l)~A6.~l(f)
In this expression the velocity of the body and the conditions of the motionless fluid interven through the normal number of Mach which plays the role of an auxiliary parameter.
There results a certain complication even
for very simple analytical bodies. As shown above the permanent motion will be reached all the more rapidly as the body's velocity is high. In this hypothesis
the preceding rela-
tion is consideratly simplified and leads to the approximation
(47)
x = f (y z) + (I +
2 ~2(3 ~2 + I) ~ 2 ) . £ I ( f ) + (I- ~2).£2(f)
:
+ ~ ~22>
The field of validity of (46) remains to be specified in relation to ~4~as well as to r. As in the case of (47) it is linked to that of PADE'S approximations
for the accelerated sequence.
It can easily be verified that if analytical function f(y z) is chosen as uniform and convex everywhere,
the field of definition of (47) remains limi-
ted. The application of (46) or (47) to the two dimensional flows (or revolutions)
leads for the body's curve
(48) ~(Y) =
~(Y) -
x =
~ (y) to
2 ~2(3 ~2+ I)(I + ~,2)2y + ~ (I + ~2)y . ~,, _ 2y- ~"-0 '2 + ~ (I+ ~2).(|+ ~,2)~,
I ~
and : 2 (~2 M 2 + I -~2)2(M 2 + I),C4(I + ~,2) . y (49)
(y)
=
~(y)
-
I[M2 A8 - 2(M2 + I) A6]-y.~'2~"+ M2(M2 + l)A6(l+~'2)y~"l + ~ M 2 (M 2 + I) A 6 (I + ~ , 2 ) ~
with
M = ~ k, and k+m, for some integral m.
Additional
storage for whole sets of field data is required, and it varies from I to 4 sets, depending on the order of the transform and ways the eigen-value estimates are handled. In passing, we note that if the component,
~process
is strictly applied for each
i.e. at each grid point, not only more storage is required but the
redundant eigen-value estimates
implicit in such process
sistency and delay the approach to the limit.
We find the
would lead to incon~process
coverges
much more slowly in most cases. V.
EXAMPLE:
A DIRECHLET PROBLEM
In a study described in our report 11, we have tested this cyclic-transform technique on line-relaxation methods applied to a model Dirichlet problem, using
107
various relaxation parameters and sweep directions.
These numerical experiments
show that a reduction in iteration number by a factor of three to five is generally possible.
In the example of Fig. 2, a line Gauss-Seidel procedure is applied to
the system based on a 9-point central difference scheme, for which neither the optimum
relaxation parameter, nor the spectral
to the best of our knowledge.
radius,
A typical convergence history of the unaccelerated
results, using a 1/30th mesh, is shown as a solid curve. is the iteration number k.
is theoretically known
The abscissa of the graph
The accelerated result based on a 2nd-order transform
is shown in short dash with circles, which approaches the limit within 1% in 30 iterations~as compared to VI.
EXAMPLES:
three to four hundred for the unaccelerated one.
TRANSONIC THIN AIRFOIL PROBLEMS
We shall study below the results of application in transonic small-disturbance theory governed by the von K~rm~n equation.21 The discussion
is confined to the
flow over a symmetric circular arc airfoil, which has an embedded supersonic region.
The basic program to be accelerated is one similar to that of Murman and
Cole I, using an x-mesh of 2½% chord, and a y-mesh near the wing 2% chord.
For the
result shown in Figure 3a, the relaxation parameter is taken to be 1.4 in the subsonic region and
0.9 in the supersonic region.
This slide gives the conver-
gence history for the velocity perturbation near the mid chord.
The unaccelerated
result shown in solid curve takes 140 iterations to approach the limit within I%; cyclic acceleration using the first-order transform presented in thin solid curve takes 60 iterations for the same accuracy.
For the results using 2nd-order trans-
form, only data at the end of each cycle are shown in circles; this takes only 40 iterations to reach the limit within I%. The results in Figure 4b differ from the preceeding one in that, here, a uniform relaxation parameter~uJ= regions.
0.95, is used in the supersonic and subsonic
The convergence rate for the unaccelerated program in solid curve is
low, as expected, taking 400 iterations or more to reach the limit within I%. This is to be compared with the 65 and 30 iterations for the two accelerated solutions. We have also studied the acceleration of transonic solutions involving circulation,
i.e. airfoil at incidence.
The basic line-relaxation program is
the same as before, except for a doubling in the number of grid points to account for the asymmetry and the use of a somewhat different pair of relaxation parameters. One sees from Fig. 5 that the use of the first-order transform in solid curve achieves a convergence within I% at 150 iterations for the circulation, whereas the unaccelerated one may take more than 400. We would like to emphasize that the above examples
involve shock waves which
are "captured", so to speak, by the numerical procedure - thanks to the "numerical
108
viscosity" inherent in the computer program. near the shock is lost.
Because of this, the flow detail
A shock-fitting method, which modifies the computer
program to fit the shock as a surface of discontinuity, has been developed. 11 The natural question to be asked is whether acceleration and shock-fitting techniques can work together.
The answer is an affirmative one.
In Figure 5,
we present the shock and sonic boundaries from our iterative, shock-fitting solution, computed for a slightly supersonic Mach number.
The unaccelerated
result obtained after 240 iterations compares well with that obtained by Magnus & Yoshihara, who used a shock capturing method based on an unsteady approach.
With
acceleration based on a 2nd-order transform, the very same shock-fitting solution is recovered in 64 iterations. VII.
CONCLUDING REMARKS In summary, our study with the transonic flow and other examples show that
the cyclic acceleration techniques based on sequence transforms may effectively increase the convergence rate and the efficiency of the relaxation methods, with minimal programming and storage changes.
A reduction by a factor of three to
five in computer time is possible, with and without shock-fitting.
In fact,
where accurate description for the shock is important, the time saved by acceleration with shock fitting can be 6 to 36 fold.
One observes that the above
demonstration involves only the use of some of the most rudimentary forms of sequence transforms.
With an increase in data storage capacity (or facility),
it should be possible to employ the more sophisticated higher order transforms and their recurrence relations which are discussed in other parts of this Proceeding.
In the meantime, possibilities for reducing the data storage require-
ment for the higher-order transforms do exist.
This is supported by a study
described in Appendix A below for an iterative procedure applied to a linear system, making use of Wynn's recursive relations for the APPENDIX A.
~-algorithm.
IMPLEMENTATION OF WYNN'S ~-ALGORITHM FOR APPLICATIONS TO ITERATIVE MATRIX EQUATIONS
In Ref. 22, Wynn uses the ~-algorithm as an acceleration technique for iterative vector and matrix problems. +
The effective use of the transforms in
the cyclic iterative method discussed in the text, as well as the corresponding elements in the
~-algorithm,
are limited, in practice, by the increased storage
requirement for the higher-order transforms •
However, the possibility for using
higher-order transforms without the increas ing storage remains, and is confirmed below for a linear system.
This is accompl ished through application of Wynn's
+The symbol " ~ " employed in this Appendix, is not to be confused with the error vector " E~" used in the text.
109 rhombus rule, and other identities for a linear iteratlve equation system. The 22 result provides an alteration from Wynn's origlnal procedure with a substantial savings in data storage. Wynn's Recursive Relation Applied to Vectors and Matrices The power of the
~-algorithm lies in the fact, established through many
examples, that if the sequence
~o,
~ , ~z .....
~x ....... i.e.,
is slowly convergent, then the numerical convergence of the sequence 4
, "'" ,
to the l i m i t rapid.
~25
,
...
,
(or a n t i l i m i t ) ,
with which sequence { O x t
In the s c a l a r c a s e , the q u a n t i t i e s
$,l
~_f
is a s s o c i a t e d ,
~(k? s a t i s f y
E~÷,~ =
-z
,
_(o)
i.e.,
E(k}
E~ ,
is f a r more
the rhombus r u l e 15
I @
(A.l) which is closely related to the Shanks' transform.
As is well known, the elements
generated in this manner in the E-algorithm may be identified with those on the upper half of the Pade Table (cf. Sec.
II in text), hence, those in ShankS' e n
transform , 2n
,n (A.2)
J with
~(k)
= ~
~
E~ )
and setting
In cases in which
=
0
~:)x is a vector or a matrix, the algorithm is still mean-
ingful, provided the inverse of the entity is consistently defined.
Wynn has
considered the following alternative definitions. (i)
Primitive Inverse: independently;
In this case, each component is considered
it amounts to a simultaneous application of the scalar
~-algorithm to components of the array. (ii)
The Samelson Inverse of a Vector:
In this case, the inverse of the
~ : (X, XI,....,XN) is taken (after K. Samelson) to be
vector
N
X-'~(~ where (iii)
~
Xj X J ' ~ ' c a ' ~ ' ' ' ' ' ' ~ " ) '
is the complex conjugate of
~
(A.3) .
The Normally Defined Inverse of a Square Matrix:
This was not
recommended for large systems. Wynn discusses in Ref. 22 applications of the algorithm to numerical analyses, including boundary-value problems,
initial-value problems, Fredholm and Voltera
integral equations, and differential equations.
110
~/ynn's_Procedure for Acceleratincj Relaxation Solutions Of particular
interest are Wynn's application to the acceleration of the
Jacobi and Gauss-Seidel
relaxation methods for iterative solution of large systems 22
of linear algebraic equations.
For subsequent discussion,
it is convenient
cIk) to arrange the array of ~s into
the familiar pattern suggested by the rhombus rule, illustrated at the middle of the page, where the original column near the left.
sequence
{~kl
is given on the first non-zero
With the identification
given by Eq. (A.2) elements on each
column corresponds to those belonging to Shanks'
e - transform of the same order n The diagonal elements in the Pad~
(with the order increasing towards the right).
Table rn, n are identified with elements on the "roof top" with even subscript, i.e.,
_(o) Eis .
with
_--'-~o~c
0
o
-r~
~J
o
~2
E~
~ ~
,
~
~s
E?
In Wynn's applications,
a sequence of vectors or matrices
is obtained from
an iterative procedure for a linear system, say,
and stored as
E(~) before the acceleration
if we have three iterates be determined as
E~)
Do
J
~
procedure
, and
~2
For example,
according to the rhombus rule (which in this case is ident-
ifiable with Pad~'s rll or Shanks ~ e I { ~ l } i.e.,
is applied.
, a better estimate will then
if one wishes to obtain
E~) ~
).
with n ~
If more resolution
is needed,
4, more iterates (with k ~ 4)
have to be generated from Eq. (A.4) and stored. Underlying this procedure
is the assumption that at the end point (towards
the right) of the application of the rhombus rule, one shall arrive at (or near) the limit ~
satisfying the equation
(~) ~ (~ (~ + ~
This assumption can indeed be justified. components of ~ ,
(A.4b)
In fact,
inasmuch as the number of
say N, is finite, the exact solution ~
Do and 2N (and only 2N) successive
iterates,
i.e.,
can be predicted from
(~o ' ~ ' ~2~-'.',~,-..,
~,
111 ~A/+I ) . . . . > ~)2N-I, ~)aN, using rhombus rule.
This follows from Eq. (A.4a),
f o r whicha corollary of the Cayley-Hamilton theorem (cf. of Ref. l l )
Eq. (3.9) on p. 5,
gives ~/' N
where p j ' s
j:o /=0 (A. 5a~ in the c h a r a c t e r i s t i c polynormal of the i t e r a t i v e
are the c o e f f i c i e n t s
matrix Q N ] I ('"X'-~j) )=1
--
~ + ~,/~.'e Pz .A-='F''"
+~,V/~.N .
(A.Sb)
Now, the right hand member of Eq. (A.5a) is precisely eN(~),__ identifiable with E ZN 4o
while the
until
a matchin9
is obtained.
RESULTS A c c o r d i n 9 t o t h e method d e v e l o p e d i n S e c t i o n s 4 . ,
?ormed a n u m e r i c a l s u r v e y o£ t h e models i n t r o d u c e d The r e s u l t s
a r e summarized i n s e v e r a l t a b l e s
v e n i e n c e we have f i x e d variables
x, 4
the time
is finite
For t h e v i s c o e l a s t i c we have computed:
t=T
(O~x~cT,
5.,
in Sections 2,,
and f i g u r e s .
so t h a t
we have p e r 3.
For c o n -
t h e r a n g e oP t h e
O~4~T).
waves i n a S t a n d a r d L i n e a r S o l i d
(see (2.10)
197 (i)
the Green's Function
and t h e s o l u t i o n s (ii)
of the following
boundary v a l u e p r o b l e m s :
~ ( )
R.(t)=l
0
S
1 = ~ S
Ro(S) =
(ill)
i
I
s[i + ~ ( s ) ] ~ (iv) (v)
i
Ro(t) = e-at
Ro(s)
R o ( t ) = e "at cos ~ t
Ro(s)-
In T a b l e s
=
s
I - Vl we compare t h e s e r i e s
a=O.
In t h i s
R(x,t)=e_t/2 and a t than
(4.2)
results.
check For t h e P.A.
case t h e e x a c t s o l u t i o n
digits
accuracy
i s a c h i e v e d For ( i i i )
is explicitly
i s Found f o r
known 171:
T ~ SO
u s i n g no more
P.A,
In F i g u r e s 1-4 t h e r e s p o n s e s t o several
For t h e
i o { ~ ( t 2 _ c~ ) : } 2
least a five (12/12]
and P.A. s o l u t i o n s
i n o r d e r t o check t h e method we do a l s o
long t i m e and c o n v o l u t i o n
An even more s t r i n g e n t when
+~
(s + a ) 2 + 2
above boundary v a l u e p r o b l e m s ; quote t h e
s+~
values of
(ii)
and ( i i i )
are plotted
for
T.
For t h e t h e r m o e l a s t i c
waves t h e F o l l o w i n g
boundary v a l u e
is
considered:
go(t) = 0
~o(S) = 0
I Eo(t)=1
go(S)=~
In T a b l e s VII, VIII we e x h i b i t comparing t h e s e r i e s ,
P.A.
the thermal (separately
g+ # g - , E+ , E-) and Ion9 t i m e relevant
results
transient
and e l a s t i c
r e s p o n s e s by
computed on t h e e x p a n s i o n o£
1161 s o l u t i o n s .
In F i g u r e s 5 - 8 some
a r e shown.
From t h e p r e v i o u s examples plays a crucial
S
role
if
wave p r o b l e m s .
we are
we can i n f e r interested
that
t h e Padd method
i n a 91obal
solution
of
198 APPENDIX The r e s p o n s e o f a S . L , S .
R (x,t) = I
t o any i n p u t
For l a r g e v a l u e s o f
t
(A.i)
~ (s)
with
g i v e n by
we a p p r o x i m a t e ( A . 1 )
point method. FOr the Greenes ~unction result
i s g i v e n by
err(s) go(S) d s
f(S):S~-c--"~ Vl+~(s~ ,
where
to(t)
No(S): I
(2.10).
using the saddle
and the standard
reads
R ( x , t ) = [ 2 = t l f " ( ~ ) l ] -½ e- t l f ' ( ~ ) l where
s
i s d e f i n e d by
For i n p u t s most r e l e v a n t close to point
(ii) (iii)
f'(~)=O
( i i ) and ( i i i )
contribution
s=O.Replacin 9
contribution
and must be computed n u m e r i c a l l y ,
Ro(s)
to
has a p o l e a t t h e o r i g i n
(A.1)
f(s)
(A.2)
by
i s o b t a i n e d when t h e ~(s)= f'(O)s+
can be a n a l y t i c a l l y
f"(O)s 2
and t h e
saddle pointis the saddle
e v a l u a t e d and we g e t :
R(x,t) -~ ½Erfc[~W] R(x,t) -~ x V~ x+ct
whereW=-(1-
x
ct
V-a
Erfc[W~
)-(2
(A.4)
1,,:a.
x
a
c t 2 ~a
Ca
For t h e r m o e l a s t i c i n 1161 by t a k i n g
(A.3)
)-½
waves long t i m e a p p r o x i m a t i o n s
the
limit
as
s~O
of the transform
have been d e r i v e d solutions
and
read: O(x,t)
= - ~1
+~
1-
1+
E(x,t) = where
Z=x
V1 +e/?-
Erf
[Z]
Err [Z] ¢-t
•
(A.5) (A.6)
199 REFERENCES Ill
F. MAINARDI & G. TURCHETTI, Mechanics Res. Comm., in press.
1 2 1 F . MAINARDI & G. TURCHETTI, t o be p u b l i s h e d . 131 S. C. HUNTER, i n Progress in S o l i d Mechanics, e d i t e d by SNEDDON & HILL, Vol.
I , p. 3, N o r t h - H o l l a n d , Amsterdam, 1960.
141 R, M. CHRISTENSEN, Theory of V i s c o e l a s t i c i t y ,
Acad. Press, N.Y.,
1971. 151 J. D. ACHENBACH, "Wave Propagation in E l a s t i c S o l i d s " ,
North-
Holland, Amsterdam, 1973. 161G. DOETSCH, "Theory and A p p l i c a t i o n of the Laplace Transform", Sprlnger-Verla9,
N.Y., 1974.
171 E. M. LEE & T. KANTER, J. Appl. Phys. 24, 1115 (1956). 181 S, KALISKY, B u l l . Acad. Pol. S c i . 13, 409 (1965). 191 E. B. POPOV, J, Appl. Math. Mech. (PMM) 31, 349 (1967). I10) M. W. LORD & g. SHULMAN, J. Mech. Phys. S o l i d s 15, 299 (1967). Illl
J. D. ACHENBACH, J. Mech. Phys. S o l i d s 16, 273 (1968).
1121 H. PADE', Thesis Ann. Ecole Nor. ~, Suppl, 1 (1892). 1131 G, A. BAKER, J. Adv. Theor, Phys. ~, 1 (1965). 1141 J. NUTTAL, J. Math. Anal. and Appl. 31, 147 (1970). 1151 J, ZINN-JUSTIN, Physics Reports (Sect. C, Phys, L e t t , ) ~, 55 ( 1 9 g l ) , I161 F. R. NORWOOD & W. E. WARREN, Quart. J. Mech. Appl. Math, 22, 283
(1969).
200
TABLE CAPTIONS The meaning o f t h e symbols used i n t h e T a b l e s X :
distance
TAU :
time elapsed from the wave Front ;
SERIES :
r e s u l t s of the p a r t i a l sums
in
~
is the Following:
= t - x/c
(NS) of the series s o l u t i o n
,
NS :
number of terms in the p a r t i a l sums
ERS :
estimated accuracy of the series s o l u t i o n defined by
PADE :
diagonal
II_,(NS-I)I (NS)
[NP/NP]
Pad6 approximants
series s o l u t i o n in ~
computed on the
,
NP :
order of P . A . .
ERP:
estimated accuracy of P.A. d e f i n e d by
CONVOLUTION: c o n v o l u t i o n piecewise
LONG TIME :
NS=2NP+I I
II"
L.NP..- 1/NP- I ] I [NP/NP] I
of the input with the Green's f u n c t i o n computed using the series when ERS~ 10-5 ,
ERS > 10-5 ~ ERP
ERP> 10-5 .
Remark t h a t
the P.A. when
and the long time approximation when
A Gauss quadrature w i t h
NO
long t i m e a p p r o x i m a t i o n o f t h e s o l u t i o n
V i s c o e l a s t i c waves in a S.L.S. w i t h
a=.S
Table
I
Input ( i ) ,
Table
II
Input
(ii)
Table
III
Input
(iii)
T a b l e IV
Input
(iv)
with
a = ,1 ,
Table V
Input
(v)
with
a : . I , ~:T0,
for
points
is used.
(see A p p e n d i x ) .
T=30.
Green's f u n c t i o n
Thermal and s t r a i n waves w i t h
NG=8
~ = .03,
NG=20
~=1.3
(NS~21, N P ~ I O ) , Table VI
T h e r m a lwaves f o r
Eo(t)=l
go(t)=O
T a b l e VII
Elastic
Eo(t )= 1
go(t) = 0
waves f o r
for
T = 5
201
TABLE I : S,LS(i ) T = 3 0
TAU
SERIES
NS
0
2.593E-03
1
2
2,123E-02
10
4
5.042E-02
6 8 I0
ERS
PADE t
NP
2.593E-03
0
1.E-06
2.124E-02
4
3.E-04
2.28E-02
12
7,E-06
5.042E-02
5
6. E-05
5.30E-02
7.222E-02
16
9.E-06
7.221E-02
7
3 • E-06
7.53 E-02
7.451 E-02
19
1,E-05
7.451E-02
9
3 • E-08
7.74E-02
5.973 E-02
23
4.E-07
5.973E-02
11
3.E-10
6.19E-02
12
3. 867E-02
25
3.E-05
3.867E-02
12
1. E-I 0
4 . OOE-02
14
2.059E-02
25
6.E-03
2,062E-02
12
I . E-O8
2.13E-02
16
9.012E-03
25
2.E-02
9.124E-03
12
6. E-07
9.45E-03
18
2.021E-02
25
3.E+O0
3.344E-03
12
I.E-06
3,47E-03
20
1.579E+00
25
2.E+O0
1.003E-03
12
1.E-05
1.04E-03
22
3.145E+02
25
2.E+O0
2.400E-04
12
3. E - 0 4
2.50E-04
24
2.965E+04
25
2.E+O0
4.911E-05
12
1,E-O1
4.57E-05
26
1,516E+06
25
2.E+O0
4.538E-06
12
2.E-01
5,75E-06
28
3.984E+07
25
2.E+O0
2.080E-07
12
7,E-01
3.85E-07
.0
ERP .0
LONG TIME --
202
TABLE fl : SLS(~ii) T = 3 0 TAU
SER IES
NS
0
5,531E-04
1
3
5,893E-02
11
6
2. 767E-01
9
£RS
NP
5,531E-04
0
9. E-06
5,893 E-02
5
2. E-05
9.27E-02
16
4.E-06
2. 767E-01
7
5,E-07
2,49 E-Ol
5. 858E-01
2o
8. E-06
5.858£-01
9
6. E-08
5,22E-01
12
8. 288E-01
25
4. E-06
8.288E-01
12
3. E-11
8,16E-01
15
9,497E-01
2.5
2. £-04
9.497E-01
12
2. E-09
9,72E-01
18
1 . 001 E+O0
25
4. E-02
9,898E-01
12
5. E-09
9• 99E-01
21
2,117 E+01
z5
2. E+O0
9.986E-01
12
7.E-08
I • OOE+O0
24
2,788E+04
25
2. £+00
9.999E-01
12
7, E-07
I , OOE+O0
27
9,103 E+06
25
2, £+00
1, O00E+O0
12
I , E-06
1 • OOE+O0
ERP
LONG TI ME
.0
ERP
LONG TI ME
PADE'
.0
TABLE Ill : SL.,S(iii) T=..~0 SER I ES
NS
o
5,531E-04
1
3
4,980E-02
12
6
2,182E-01
9
TAU
ERS ,0
PADE'
NP
5,531E-04
0
,0
9. E-07
4.980E-02
5
2,E-05
7.34E-02
15
5.E-06
2.182E-01
7
3,E-08
1.87E-01
4,402E-01
2O
8.E-06
4,402E-01
9
4,E-08
3.71E-01
12
6,031E-01
25
5,E-06
6,031E-01
12
7, E-11
5.88E-01
15
6,786E-01
25
4, E-05
6,786E-01
12
6,E-12
6.91E-01
18
7,073E-01
25
3. E-02
7,018E-01
12
8. E-09
7,07E-0l
21
-3,681E+01
25
2, E+O0
7 • 065 E-01
12
2, E-09
7,07E-01
24
-9,250E+04
25
2.E+O0
7,071E-01
12
2. £-06
7.07E-01
27
-6,870E+07
25
2.£+00
7,071E-01
12
3.£-06
7,07E-01
203
TABLE
TAU
SER I ES
NS
0
5 • 531 E-04
1
3
5,279E-02
12
6
2.224E-01
9
IV
:
ERS
SLS(iv) T=30 PADE'
NP
5,531E-04
0
7, E-07
5,279E-02
5
5.E-05
5.279E-02
14
9, E-06
2,224E-01
6
5, E-05
2,224E-01
4. 054E-01
21
3.E-06
4.054E-01
10
2.E-09
4,054E-01
12
4. 638E-01
25
I.E-05
4.63gE-01
12
2,E-lO
4,638E-01
15
3,972E-01
25
6.E-04
3,972E-01
12
1,E-08
3.972E-01
18
3.01 OE-01
25
2, E-O1
2,867E-01
12
3.E-08
2.867E-01
21
2,240E+01
25
2,E+O0
1.902E-01
12
1,E-07
1,902E-01
24
3 , 1 0 9 E+04
25
2, E+O0
1,223E-01
12
7. E-06
1,223E-01
27
1. 018E+07
25
2. E+O0
7. 808E-02
12
9. E-06
7,808E-02
PADE'
NF'
ERP
CONVOLUTI ON
5,531E-04
0
.0
TABLE
V
:
ERP .0
CONVOLUTI ON
5.531E-04
SLS(v) T=.~O
SER I ES
NS
0
5,531E-04
1
3
4,835E-02
12
4, E-07
4. $35E-02
5
2, E-04
4.835E-02
6
t . 597E-01
t6
8, E-06
1,597E-01
7
4, E-06
1,597E-01
9
1.575E-01
21
4.E-06
1,575E-01
10
5,E-08
1.575E-01
12
-1.583E-02
25
3, E-05
-1,583E-02
12
5, E-06
-1.583E-02
15
-1,400E-01
25
2. E-03
-1.399E-01
12
2. E-05
-1,399E-01
18
-6.051E-02
25
6. E-01
-7,035E-02
12
2, E-04
-7,035E-02
21
2,030E+O1
25
2,E+O0
5,592E-02
12
4.E-03
5,590E-02
24
2,749E+04
25
2, E+O0
6,375E-02
12
5, E-03
6,556E-02
27
8,939E+06
25
2, E+O0 - 2 . 870E-02
12
1, E+O0
-1.381E-02
TAtl
ERS
.0
,0
5,531 E-04.
204 TABLE Vl
: THERMAL WAVES T= PADE"
ERP
LONG TI ME
X
SERIES
5,24
-5,358E-02
O.
-5.358E-02
O.
4.81
-4.851E-02
1, E-06
- 4 . 851 E-02
3. E-06
-2,56E-02
4.18
-3,807E-02
7, E-06
- 3 . 807 E-02
I , E-05
-2.39E-02
4.18
-2,276E-02
I,E-05
-2,276E-02
2. E-05
-2,39E-02
3.54
-2,011E-02
4, E-06
-2,011E-02
8, E-06
-2.17E-02
2.90
-1,708E-02
5. E-06
- I , 708E-02
3, E-07
-1,89E-02
2,27
-1,373E-02
7.E-05
-I,373E-02
6. E-07
-1,55E-02
1.63
-1,010E-02
4. E-03
- 1 . 009 E-02
9 • E-06
-1.16E-02
.99
-6.707E-03
2, E-01
- 6 . 241 E-03
1, E-04
-7.30E-03
.35
-7.372E-03
4. E+O0
-2,263 E-03
3, E-03
-2.67E-03
ERS
w ~
TABLE VII : ELASTIC WAVES T = ~ X
SERIES
ERS
PADE'
ERP
LONG TIME
5.24
5.369E-01
O.
5.369E-01
O.
4.81
7. 709E-01
4. E-06
7. 709E-01
6. E-05
9,74E-01
4.18
9.276E-01
4. E-06
9,276E-01
1.E-05
9.76E-01
4.18
9,785E-01
4. E-06
9. 785E-01
1. E-05
9.76E-01
3.54
9,811E-01
9.E-06
9.811E-01
1.E-05
9,78E-01
2.90
9 • 840E-01
4. E-06
9. 840E-01
8. E-07
9.81E-01
2.27
9.872E-01
4;E-06
9.872E-01
7, E-09
9.84E-01
1.63
9.906E-01
3. E-04
9.906E-01
8.E-07
9.88E-01
.99
9,950E-01
1.E-02
9,942E-01
9, E-06
9.93E-01
.35
1. O08E+O0
3. E-01
9,979E-01
7, E-05
9,97E-01
205
FIGURE CAPTIONS I - 2
The response o f a S . L . S .
with
a = .5
For i n p u t ( i i )
and
T=l,3,5;10,30w50.
3- 4
The same as F i g u r e s
1 - 2 for
S- 6
Thermal and e l a s t i c
waves w i t h
Eo(t)=Z, 7- 8
go(t)=O
The same as F i g u r e s
and 5- 6
input
(iii)
s = .03 ,
~= 1 . 3
T=2.5. For
~= .03 ,
~ = .9
for
input
206
R, 1
a=O,5
,
I
i 2
SLS .
i 4
2o
Fig. l
Fig.2
a=0,5
40
R~ 1 SLS a=0,5
SLS a=0,5
t.50
2
4
Fig.3
Malnardl~
Turc
20 Fig.4
het
tl
40
207
E
=0,03 (3 = 1 , 3
= 0,03 (3 = 1,3
t=5
-0,1
I
0
4
t~5
t=2
I
6
I 4
x
Fig. 5
I 8
E
F\ = 0,03 /3=0.9
C = 0,03 /3 = 0 , 9
t=2
I 4
Fig. 7
Mainardi-Turchetti
x)~
Fig. 6
I 0
x~
t=5
,I,
4
Fig. 8
0
x
APPLICATION OF METHODS FOR ACCELERATION OF CONVERGENCE TO THE CALCULATION OF SINGULARITIES OF TRANSONIC FLOWS* Andrew H. Van Tuyl Naval Surface Weapons Center White Oak Laboratory Silver Spring, Maryland 20910 USA
i.
Introduction Initial value problems in gas dynamics which lead to transonic flows include
the inverse blunt body problem, given and the body w h i c h w o u l d
in which a bow shock wave in a uniform flow is produce it is calculated,
and the inverse calculation
of nozzle flows starting from data given on the centerline.
Each of these problems
can be expressed as an initial value problem for a second order quasi-linear differential
equation satisfied by the stream function.
When the initial curve and initial data are such that the initial curve is noncharacteristic,
it follows from the Cauchy-Kowalewskl
theorem [i, page 39] that
the initial value problem can be solved in terms of power series in the neighborhood of a given point of the initial curve.
However,
the region of convergence
series obtained may be too small for practical use, due to the occurrence singularities,
either real or complex, near the initial curve.
of the of
This was found by
Van Dyke [2] in the case of the inverse blunt body problem, where a limiting line** (envelope of characteristics) flow.
occurs in the upstream analytic continuation
of the
This limiting line lies closer to the shock than the distance between the
latter and the body, and hence, a power series solution in the neighborhood
of a
point of the shock diverges at the body and cannot be used directly to calculate the flow there. In [3], Leavitt has calculated
the shape and position of this limiting
near the axis of symmetry by a modification
line
of a method due to Domb [4], starting
* This work was supported by the Naval Surface Weapons Center Independent Fund. **Also called limit line.
Research
210
from the power solution of the inverse blunt body problem in the neighborhood the nose of the shock.
The location of the limiting line was then used to trans-
form the series so that convergence was obtained at the body. Schwartz
More recently,
[5] has used Domb's method to calculate limiting lines in the flows
produced by parabolic and paraboloidal number.
of
Various modifications
shocks in a free stream of infinite Mach
and extensions
of Domb's method have been applied to
problems of statistical mechanics by Domb, Sykes, Fisher, and others
([6], for
example). Limiting lines in solutions of the inverse blunt body problem have also been calculated by Garabedian and his students
([7] and [8]) by use of Garabedian's
method of complex characteristics. Limiting lines may also occur in nozzle flows obtained from given centerline distributions
of velocity or Math number.
region of convergence
As in the case of blunt body flows, the
of a power series solution may be restricted by a limiting
line even though the point about which the solution is obtained lles in the subsonic region.
In nozzle design,
given centerline distribution
it is of practical interest to know if a
leads to a limiting line which lies between a
desired streamline and the axis of symmetry. A procedure for calculation of limiting lines will be described, from a power series solution, are used.
in which methods for acceleration
This procedure involves the ratio of successive
power series, as in Domb's method,
of convergence
coefficients
and a necessary requirement
the extent of the region of convergence
starting
of a
is therefore that
in the direction of at least one of the
coordinate axes should be determined by a limiting line.
Sequences are constructed
which converge to points on a limiting line and to its order k ~i.
With the
assumption that the single power series used in this calculation has only one singularity on its circle of convergence, sequence transformations,
including
it is proved that certain nonlinear
the e 1(s) transformation
defined by Shanks
([9], page 39) accelerate the convergence of these sequences.
211
The results obtained hold also for analytic initial value problems for other equations or systems of equations in two independent variables, when the given equation or system of equations can be replaced by a characteristic system in two independent variables.
In particular, limiting lines can be calculated by the
present method in the one-dimensional unsteady flow produced by a given piston motion.
Finally, the present method is also applicable to some of the series
occurring in [6].
2.
Limiting Lines of Order k In both the inverse blunt body problem and the inverse calculation of nozzle
flows, the stream function @ satisfies a quasi-linear second order partial differential equation of the form a~xx + b~xy + C~yy + d = 0, where the coefficients are analytic functions of their arguments.
(2.1) The independent
variables denote cartesian coordinates in the two-dimensional case and cylindrical coordinates in the axially symmetric case.
As in [i], pp. 491-493,
(2.1) can be
replaced by the system of characteristic equations y~ = h I x Y8 = h2 x8 d pe + h2q e + ~ x
= 0
(2.2)
P8 + hlq8 + ~a x8 = 0 ~
- px
- q y~
=
0
where h I and h 2 are the roots of the equation ah 2 - bh + c = 0.
(2.3)
Real values of = and 8 correspond to values of x and y for which (2.1) is hyperbolic.
It follows from the Cauchy-Kowalewski theorem that the solution of an
analytic initial value problem for (2.2) in the real eB-plane is analytic.
212
Given a solution of (2.2) which is analytic in a domain D of the real ~B-plane, the functions x(~,B) and Y(~,B) define a mapping which is one-to-one in any portion of D in which the Jacobian J = x yB-xBy ~ does not vanish.
Let k ~ l
be an integer, and let J and its derivatives of order up to and including k-i vanish along a curve C in D.
Then the image of C in the xy-plane is defined to be
a limiting line of order k. The well-known result ([i0], for example) that a regular arc of a limiting line of the first order is an envelope of one of the families of characteristics can also be shown to hold for limiting lines of order k > l .
As in the case of
limiting lines of first order, characteristics of the second family have infinite curvature at the limiting line for k>l. Finally, we can prove also that the behavior of flow quantities in the neighborhood of a limiting line of order k ~ l Theorem i.
is given by the following theorem:
Let a solution of (2.1) have a limiting line of order k ~ l
with
the equation x = Xo(Y), where Xo(Y) is analytic for yl 2.
When bo=0 but bl~O ,
an+l/a n =
=
a - --~]{
_
l+2/m
+ j--liE e.n -~-lJ
+ O(n - N - l )
, m ~ 2
1 + E e.n-J/m-i + 0(n-N/m-l) j=l 3
m > 2.
A similar but more complicated theorem can be proved when there are several singularities on IzI=l, with a different value of m associated with each.
With
more than one singularity, however, the transformations used here become less effective.
4.
Sequences for Calculation of Limitin Z Lines Given an analytic initial value problem for an equation of the form of (2.1),
let the origin be taken at a point of the initial curve, and let (Xo,0) be a point on a limiting line of order k ~i.
Let F(x,y) be any one of the dependent
variables, such as the density or pressure in the inverse blunt body problem, and
215
as described in the introduction, let F(z,0) be analytic for Izl ~ x ° except at z = x . This assumption, that only one singularity lies on the circle of o convergence, appears to be verified in special cases which have been calculated. We have F(z,o) = ~
(4.1)
anzn , Izl < x
n= 0
o '
and by Theorem i,
(4.2)
F(z,o) = n=O~ b n ( l - xZ--) n/(k+l)o
in a cut neighborhood of z = x O .
It follows from Corollary 1 of Theorem 2
that when bo#0 in (4.2), (i
l+i/(k+l)) an = x n+l an+ I o
i + N-I ~ c.n-j-I + O(n -N-I) 1 J=l 3
k = I (4.3)
=
X
0
i + ~ c.n-j/(k+l)-I + O(n-N/(k+l)-l) , k > 2 j=l 3
From (4.3), we obtain rn
=
(Sn-l)(n+l)(n+2) = ~3
I
1
i + ~ d.n -3" + 0(n-N) j=l 3
k = i
= (l+i/(k+l)) I + N-I ~ d.n-j/(k+l) + 0 (n_N/(k+l ) )} , k ~ 2 j=l 3 2 when bo~0, where Sn=an+2an/an+ 1. N o t i n g t h a t ( n + 2 ) S n - ( n + l ) t e n d s t o u n i t y
(4.4)
as
n-~o, we see that the sequence t
= n
(Sn-l)(n+l)(n+2) (n+2)Sn_(n+l)
(4.5)
tends to the same limit as the left hand side of (4.4), and we can show that it has the same asymptotic form as (4.4) as n-~.
Denoting the coefficients of
the asymptotic expression for tn by ej, we find that e I = d I when k ~ 2 . When k = i, however, we have d I = e I + 3/2. Thus, while tn and rn have the same rate of convergence, one may be asymptotically more accurate than the other.
216
When bo=0 but bl#0 , sequences similar to the preceding can be obtained by use of Corollary 2 of Theorem 2.
By consideration of both the limit and the rate of
convergence of tn, it is possible to determine k and to decide whether or not
b
o
5.
vanishes.
Nonlinear Sequence Transformations Let A
be a given sequence, and let
n
B
= n
An+lAn 1-An
.
(5.1)
An+ l+An- i- 2An
The sequence Bn is equal to the first order transform el(An ) defined by Shanks [9] and to the sequence e~ n) o b t a i n e d by t h e e - a l g o r i t h m of Wynn [11], referred to as the Aitken 6 2 process. sequence A
n
Under certain conditions, a convergent
is transformed to a sequence B
converges more r a p i d l y .
and i s a l s o
In particular,
n
which has the same limit as A
the latter
n
and
holds when lim IAAn+I/AAnI#I. n..+~
When lim ' n'IAAn+l/AAI=l but lim AAn/ABn=S@I, the transformed sequence B converges n_>~ ~ n with the same rapidity as A . For this case, Shanks ([9], page 39) has defined n the transform
e
~S)(A n
)=
sB -A n n s-l"
(5.2)
A transformation equivalent to this, called Un, was also introduced by Lubkin [12]. Finally, Lubkin ([12], page 229) has introduced a more general transformation which is expressed in the present notation by AA BnA--~
- An n
Wn =
AA n AB n
(5.3) 1
With the preceding definitions, we can verify the following theorems by direct calculation:
217
Theorem 3.
Let a given sequence have the form
A
as n-~o for all N ~ I .
n
N-I = A + E a.n-J-~ + O(n-N-=) . j 3=o
Then
N-I e I(c~+l) (An) = A + ~ b.n -j-~-2 + 0(n-N-~-2). j=o 3
Corollary.
With the notation e~Sl)(e~S2)(An)) = e(Sl)e~S2)(An ), we have
e(~+2r+l)^(~+2r-l) ... e(e+l)(An ) = A + 0(n -=-2r-2) =i
Theorem 4.
Let a given sequence have the form
A n
as n-~o for all N ~ 2
N-I = A + ~ a.n -j/m-~ + 0(n -N/m-~) j=l 3
where m = 2,3, . . . .
,. , e(l+e+i/m) i ~'~n)
Then
. = A + N-I ~ 5.n -3/m-~ + 0(n_N/m_~) j=2 3
for N = 3,4, . . . .
Corollary.
(lq~+r/m) (l+e+(r-l))/m e (l+~+I/m) ( A ) el el "'" I n
= A + 0(n-(r+l)/m-=).
Theorems 3 and 4 have direct application to the sequences of Section 4. In particular, with a
An
(i=
3/2~
n__ n---~" an+ I
218
in the case k = i, we see that
el(2r+l)el(2r-l) ... e ~ 3 ) ( A )
= x
+ 0(n -2r-2) o
We see that the sequences A respectively,
n
in Theorems 3 and 4 must have e > 0 and e > -i/m,
in order to be convergent.
However,
the theorems do not specify the
sign of ~, and the corollaries show that the iterated transformations converge to A for sufficiently large values of r whether or not A Finally, the following theorems for W Theorem 5.
= A + n
as n ->~ for N = 1,2, . . . .
Then
W
= A + n
are nearly identical with the preceding:
N-I ~ a.n -j-~ + O(n -N-e) j=o
N-I ~ c.n -j-e-2 + O(n-N-~-2), j=o J
Let a given sequence have the form
A
= A + n
N-1 ~ a.n -j/m-~ + 0(n-N/m-a). j=l 3
as n ->~ for N = 2,3,... where m = 2,3, . . . .
W
= A + n
for N = 3,4 . . . .
converges.
Let a given sequence have the form
A
Theorem 6.
n
n
Then
N-I 0 (n-N/m-a) ~ c.n -j/m-~ + j=2 3
Corollaries which are exactly parallel to the corollaries of
Theorems 3 and 4 clearly hold.
219
6.
Numerical Examples The sequences of Section 4, together with the sequence transformations of
Section 5, have been applied to the calculation of limiting lines of first order in the inverse blunt body problem and in the calculation of nozzle flows.
The
calculation of the axial point on the upstream limiting line in the case of a paraboloidal shock at a free stream Mach number of 2 is shown in Table I, which was calculated by use of 39 coefficients of the power series for the density. Calculations were carried out on the CDC 6500 in double precision, using the method of [13]. The sequences r and t converge to 3/2 and are accelerated by n n the transformation e~ , 2) -
showing that the upstream limiting line is of first
order, and that b #0 in the expansion of the density in the neighborhood of the o limiting line.
Values of rn, tn, and e~2)(tn ) are given in Table 2, and show that
t is preferable to r in this case. n n
When the same calculations are repeated
using the power series for the stream function, it is found that rn and tn approach the limit 5/2.
Thus, it appears that bo=0 while bl#0 in the expansion of the
stream function in the neighborhood of the axial point of the limiting line. The rates of convergence shown in Section 5 cannot usually be realized beyond the transformation e1(3) when using single precision, because of loss of significance due to subtraction.
However, the accuracy obtained by use of e~ 3)- in
single precision has been found to be sufficient in all examples calculated. Agreement of successive terms of the sequence to 5 or 6 significant figures is usually obtained. Finally, calculations of limiting lines in nozzle flows have been carried out in single precision on the CDC 6500 using e~ 3).
The result of such a calculation
is shown in Figure 1 in the axially symmetric case, where u = 1 + (%/3/2)x on the centerline.
This centerline velocity distribution is the same as that of case (f)
of [14] after a change of reference quantities.
The sonic line, limiting
characteristic, and streamlines were calculated by the method of [15]. All
220
calculations
were carried out using 25 terms of the series in y, and convergence
to 5 or more figures was found except near the limiting line.
With the use of
double precision, the more rapidly convergent transformation e(5)e(3)(A ~ would i i n" give additional aceuracy. The portion of the streamline ~ = 0.2 to the left of the limiting characteristic can be taken as the wall of a nozzle in the subsonic and transonic regions, since the flow is analytic on the limiting characteristic up to and on the streamline.
However, the streamline ~ = 0.45 meets the limiting line above its
point of tangency with the limiting characteristic, and cannot be used as part of a nozzle contour.
The dashed portion of the limiting characteristic belongs to a
second solution of the flow equations having the limiting line shown.
The two
solutions correspond to the same analytic solution of equations (2.2) in the •
i
~8-plane, but on opposite sides of the curve J =, 0 which corresponds to the limiting line.
7.
References i.
R. Courant and D. Hilbert, Methods of Mathematical Physics, Vol. II, Interscience Publishers, New York, 1962.
2.
M. D. Van Dyke, "A Model of Supersonic Flow Past Blunt Axisymmetric Bodies with Application to Chester's Solution," Journal of Fluid Mechanics, Vol. 3, (1958), pp. 515-522.
3.
J. A. Leavitt, "Computational Aspects of the Detached Shock Problem," AIAA Journal, Vol. 6 (1968), pp. 1084-1088.
4.
C. Domb, "Order-Disorder Statistics, II. A Two Dimensional Model," Proceedings of the Royal Society of London, Series A, Vol. 199 (1949), pp. 199-221.
5.
L . W . Schwartz, "Hypersonic Flows Generated by Parabolic and Paraboloidal Shock Waves," Physics of Fluids, Vol. 17 (1974), pp. 1816-1821.
6.
M. E. Fisher, "The Nature of Critical Points," Lectures in Theoretical Physics, Vol. VlIC, University of Colorado Press, Boulder, 1964, pp. 1-159.
7.
P. R. Garabedian, "Global Structure of Solutions of the Inverse Detached Shock Problem," Problems of Hydrodynamics and Continuum Mechanics, published by the Society for Industrial and Applied Mathematics, 1969, pp. 318-322.
221
8.
E. V. Swenson, "Geometry of the Complex Characteristics in Transonic Flow," Comm. Pure and App. Math., Vol. 21 (1968), pp. 175-185.
9.
D. Shanks~ "Non-Linear Transformations of Divergent and Slowly Convergent Sequences," Journal of Mathematics and Physics, Vol. 34 (1955), pp. 1-42.
10.
R. Courant and K. O. Friedrichs, Supersonic Flow and Shock Waves, Interscience Publishers, New York, 1948.
Ii.
P. Wynn, "On a Device for Computing the em(Sn) Transform," MTAC, Vol. i0 (1956), pp. 91-96.
12.
S. Lubkin, "A Method of Summing Infinite Series," Journal of Research of the Bureau of Standards, Vol. 48 (1952), pp. 228-254.
13.
A. H. Van Tuyl, "Use of Pad~ Fractions in the Calculation of Blunt Body Flows," AIAA Journal, Vol. 9 (1971), pp. 1431-1433.
14.
G. V. R. Rao and B. Jaffe, '~ Numerical Solution of Transonic Flow in a Convergent-Divergent ~ozzle," Final Rept., Contract NAS7-635, March 1969, NASA.
15.
A. H. Van Tuyl, "Calculation of Nozzle Flows Using Pad~ Fractions," AIAA Journal, Vol. II (1973), pp. 537-541.
222
Table i.
Distance of Upstream Limiting Line from Nose of Shock for Paraboloidal Shock Wave r 2 = 2x with Free Stream Maeh Number = 2.
n
34
An = -(I
n+l'an+ 1312~an
-
0.083546753454
e~3)(An)
el(5)e(3)(An)l
0.0835451981166
0.0835451972856
35
.0835466622853
.0835451980201
.0835451972406
36
.0835465789026
.0835451979365
.0835451972874
37
.0835465024447
.0835451978019
.0835451972615
Table 2.
Sequences r
n
and t Corresponding to Axial Point on Upstream Limiting n
Line for Paraboloidal Shock Wave r 2 = 2x with Free Stream Mach Number = 2.
n
r
n
t
n
e~2)(tn )
34
1.5685409
1.5012612
1.5000020
35
1.5665483
1.5012223
1.5000018
36
1.5646684
1.5011858
1.5000017
37
1.5628918
1.5011514
1.5000016
223
1.1 = 0.45
STREAMLINES
0.9
I.
I
0,8
0.7~'~,~
~ = 0.2 LIMITING LINE
0,6
--
0,5
--
0.4--
0.3
~LIMITING CHARACTERISTIC
--
0.2--
0.1
--
0 -0.6
I
1
I
I
I
-0.5
-0.4
-0.3
-0.2
-0.1
0
I
I
0.1
0.2
0.3
X
Figure i.
Limiting line in the nozzle flow with centerline velocity distribution u = 1 +
(/'3/2)x.
THE USE OF PADE FRACTIONS IN THE CALCULATION OF NOZZLE FLOWS* Andrew H. Van Tuyl Naval Surface Weapons Center White Oak Laboratory Silver Spring, Maryland 20910 USA
i.
Introduction In nozzle design, flow calculations are usually carried out by an inverse
procedure in which either the Mach number or velocity is given on the centerline. Calculations in the supersonic region can be carried out accurately by the method of characteristics when the solution in the transonic region is known.
However, the
inverse problem is improperly-posed in the subsonic region, and it is therefore not possible to calculate the subsonic and transonic portions of the flow to arbitrary accuracy by means of finite-difference marching procedures without the use of complex extension.
Methods for calculating the flow in the transonic region
([i], for example) are mainly applicable to the design of nozzles with large throat radius of curvature, such as wind tunnel nozzles.
These methods are not
sufficiently accurate for the design of short nozzles, since rapid changes then occur near the throat.
Improved accuracy in the case of short nozzles is given by
[2], but the accuracy which can be obtained is limited by the fixed number of terms of the power series used.
Short nozzles with a uniform exit flow are of interest
in connection with gas dynamic lasers [3], where two-dimensional nozzles have been used, and with chemical lasers [4], where both two-dimensional and axially symmetric nozzles are applicable.
More accurate subsonic calculations have been carried out
in the axially symmetric case by Armitage [5] and Rao and Jaffe [6] by means of Garabedian's method of complex characteristics
[7, 8].
Calculations by the method
of complex characteristics have also been carried out by Solomon [9] in the twodimensional and axially symmetric cases.
*This work was supported by the Naval Surface Weapons Center Independent Research Fund.
226
When the flow quantity given on the centerline is analytic, the Cauchy-Kowalewski
theorem
([i0], page 39) that the inverse problem can be
solved in terms of power series in the neighborhood eenterline.
it follows from
of a given point of the
While these series are known to converge near the given point, the
region of convergence may be such that the series are either divergent or too slowly convergent at points of physical interest.
It was noted by Van Dyke [ii]
that the former occurs in the case of the inverse blunt body problem, presence of a limiting line (envelope of characteristics) continuation
of the flow.
due to the
in the upstream analytic
This limiting line lies closer to the shock than the
distance between the shock and the body, and hence, the body does not lie within the region of convergence
of the series.
However,
it was found in [12],
[13],
and [14] that Pade" fractions formed from these series give accurate results at the body, and can be used to compute the flow there.
A similar procedure has been
used in [15], where Pade" fractions are formed from certain power series in the neighborhood body problem,
of points of the centerline.
examples indicate that the use of Pade" fractions leads to convergence
when the series diverge, converge.
As in the case of the inverse blunt
and that convergence
is accelerated when the latter
The region of convergence of the series may be restricted by limiting
lines, as in the blunt body problem, calculations
or by complex singularities.
Accurate
can also be carried out in the supersonic region by use of Pade"
fractions when limiting lines do not occur near the centerline.
Near a limiting
line, convergence of both the series and the sequences of Pade" fractions
is found
to be slow. The present paper extends
[15] by investigating
fractions can be used in the calculation
additional ways in which Pade"
of nozzle flows.
In particular,
Pade"
fractions are formed from power series along given curves which intersect the centerline. expansions
By use of these results, more efficient use of the double Taylor obtained in [15] can be made.
227
2.
Power Series Solution of the Inverse Problem The coordinate system used is shown in Figure i, where x and y denote
Cartesian coordinates in the two-dimensional case and cylindrical coordinates in the axially symmetric case, and where the origin is at the sonic point on the centerline.
We will assume irrotational flow of a perfect gas with ratio of
specific heats y.
Let p* = c* = i, where p* is the critical density and c* is the
critical sound speed, and let the stream function satisfy u = (i/py°)~/~y, v = - (i/pyO)~/~x,
(2.1)
where u and v are the x and y components of velocity, respectively, p is the density, and ~ = 0 and i, respectively, in the two-dimensional and axially symmetric cases.
Substituting (2.1) into Bernoulli's equation and the condition
for irrotationality, we obtain (i/y2O)(~x 2 + ~y2) + [2/(y-1)]py+l _ [(y+l)/(y_l)]p2 = 0
(2.2)
and
P(~xx
+ ~yy _ oy-l~y) _ Px~x _ py~y = O,
(2.3)
respectively. Let the density, Mach number, and velocity on the centerline be denoted by Po(X), Mo(X) , and Uo(X) , respectively.
Then Po(X) is given in terms of Mo(X) by
the relation Po = {2/(y+l) + [ (y-i) / (y+l) ]Mo2} -I/(y-i),
(2.4)
and in terms of Uo(X), by Po = { (7+1)/2 - [(Y-I)/2]Uo2}I/(y-I)" Let M 0
(2.5)
(x) and Uo(X) have the expansions Mo(X) = ~ Mi(X-Xl)i-1 i=i
(2.6)
Uo(X ) ~ ui(x-xl )i-I i=l
(2.7)
and
228
in the neighborhood of x = x I. Given either (2.6) or (2.7), we can find the coefficients Pi of the expansion ¢o
Po (x) = .i__~lPi(X-Xl)i-I
(2.8)
by use of (2.4) or (2.5) respectively, and subroutines for power series manipulations.
We can then find the solution of the inverse problem in the
neighborhood of the point (Xl, 0) in the form 4"= ~ ~ ~ij y2i-l+~(X-Xl)J-l' i=l j=l P =
j--~lPij y 2i-2 (X-Xl)J-l" i:1 "=
(2.9)
(2.10)
We see that PlJ = PJ" To summarize the calculation of the remaining coefficients, let the left hand sides of (2.2) and (2.3) be denoted by A and B, respectively.
We have A = ~ ~ y2i-2(X-Xl )j-I i=l j=l ai~
(2.11)
B = ~ ~ bij y2i-l+O(x-xi)J-i i=l j=l
(2.12)
and
For J~2, we find that alj = 2(i+~)241141j + a~j,
(2.13)
411 = [i/(I+~)]{[ (y+l)/(y-1)]Pll2 - [2/(y-l) ]PllY+l}I/2
(2.14)
where
and where alj does not involve ~lj" For i>_2, we have aij = CI 4ij + C2 Pij + a~j,
(2.15) bij = C3 4ij + C4 Pij + b~j,
229
where C I = 2(i+0)(21-i+~)~ii C2 :
,
[2(y+l)/(y-l)](PllY-Pll
), (2.16)
C 3 = 2(2i-l+~)(i-l)Pll , C4 = -2(l+q)(i-l)~ll. and where a~j and b~j do not involve ~ij and Pij" coefficients can now be described as follows: alj = 0 in (2.13) for j = 2, 3, .... and calculating alj.
The calculation of the remaining
First, ~lj is calculated by setting
For each j, a~j is found by setting ~lj = 0
The coefficients ~iJ and Pij are then found for i ~ 2 by
setting aij and bij equal to zero in (2.15) and solving the resulting pair of equations simultaneously.
For given i and j, a~j and b~j are obtained by setting
~ij and PiJ equal to zero and calculating aij and bij.
Assuming that Pi is known
for i ~ i ~ 2K - i, we can find ~ij and Pij for i ~ i ~ K and i ~ J ~ 2K - 2i + i.
3.
Pade" Fractions Let f(z) = ~ c.z i i i=0
be a given power series with co @ O. are defined as follows:
(3.1)
Then Pade" fractions fk,n(Z), k ~ 0, n ~ O,
fk,n(Z) is a rational fraction with numerator and
denominator of degrees less than or equal to n and k, respectively, such that the Taylor expansion of fk,n(Z) in the neighborhood of the origin agrees with (3.1) to more terms than that of any other rational fraction with numerator of degree < n and denominator of degree ~ k ([16], page 377). exists and is unique.
This fraction always
Convenient methods of calculation include the QD algorithm
of Rutishauser [17] and the ~ - algorithm of Wynn [18]. The sequence fn,n(Z), which involves the first 2n + i terms of (4.1), is often found to converge much more rapidly than sequences in which either k or n
230
remains constant.
The convergence of fn,n(Z) has been proved only for special
classes of functions, and none of the available convergence results appears to be directly applicable to the series occurring in either the inverse blunt body problem or the inverse calculation of nozzle flows.
In both [14] and [15],
however, it was found that the sequence fn,n(Z) converges in most cases when the series diverges, and that it accelerates convergence when the latter converges.
4.
Power Series for Constant x Before Pade" fractions can be used, the solution of the inverse problem must
be expressed in terms of power series in one variable.
One approach is to write
(2.9) and (2.10) as single power series in y with coefficients which are functions of x.
We have @
= ~
' , , ~i~x;y
2i-i+~
(4.1)
i=l and oo
y2i-2,
(4.2)
~ij (X-Xl)J-i
(4.3)
Pi (x) = ~ Pij(X-Xl)J-i j=l
(4.4)
p =
Pi(X) i=l
where
~i (x) = ~ j=l and
when Ix - Xll and y are sufficiently small.
The coefficients ~i(x) and Pi(X) can
he calculated by use of Pade" fractions for a given value of x, after which Pade" fractions can be formed from (4.1) and (4.2).
Pade" fractions formed from (4.3)
and (4.4) are exact at x = Xl, but their accuracy decreases in general as Ix - Xll increases.
In a given case, however, it may be possible to find an
interval about x I throughout which acceptable accuracy is obtained.
It would be
more economical to use several such intervals with overlapping regions of validity than to solve the inverse problem for each value of x.
231
In nozzle calculations, expansions in powers of 4 may be more convenient than (4.1) and (4.2).
We know by the Cauchy-Kowalewski theorem that for x sufficiently
near Xl, (4.1) has a nonzero radius of convergence.
Assuming that ~l(X) is not
zero for a given x, we can invert (4.1) to obtain an expansion of the form
y = ~ Yi(X) 4 2i-1 i=l
(4.5)
in the two-dimensional case, and
Yi(x) ~i y2 = ~ i=l in the axially symmetric case for sufficiently small 4.
(4.6) By substituting (4.5) or
(4.6) in (4.2), we can obtain an expansion for 0 in powers of 4.
Similarly,
starting from the Taylor expansion of other flow quantities in the neighborhood of (Xl, 0), we can find their expansions in powers of 4 for constant x.
5.
Power Series Along Given Curves Another procedure for expressing the solution of the inverse problem in terms
of power series in one variable is to calculate ~ and 0 along given curves through (Xl, 0).
Families of such curves can be chosen which sweep out a neighborhood of
(Xl, 0).
Let the equation of a given curve be x - x I = g(y), where
g(y) = ~ gi yi i=l for sufficiently small y.
(5.1)
On substituting x - x 1 = g(y) in (2.9) and (2.10), we
obtain expansions of the forms
(5.2)
4 = ~ aiyi+° i=l and
0 = ~ biY i=l respectively.
i-i
,
(5.3)
With the equation of the given curve written in the alternate form
y = g(x - Xl) , y is replaced by x - x I in (5.1) through (5.3). case when g(y) is a function of y2, we have
In the special
232
= ~
aiy2i-l+a
(5.4)
i=l and 0 = ~ biY i=l along the given curve.
2i-2
(5.5)
We see that the coefficients a. and b. in (5.2) through I
1
(5.5) are known exactly, while the coefficients of (4.1) and (4.2) for x # x I are known only approximately.
In addition, ~i(x) and Pi(x) become less accurate for a
given x as i increases, since the number of terms of their Taylor expansions which are available then decreases.
However, the accuracy obtained in a particular
case is dependent on the choice of Xl, and can be determined only by trial. As in section 4, if a I @ 0, we can invert (5.2) to obtain y =
ci~
(5.6)
i=l in the two-dimensional case and y = ~ di~i/2 i=l in the axially symmetric case for sufficiently small ~.
(5.7) Similarly, if a I ~ 0 in
(5.4), we have y = ~
c
i=l
i*
2i-1
(5.8)
in the two-dimensional case, and y in the axially symmetric case.
2
" = ~ di~l i=l
(5.9)
By use of the preceding expansions for y, we can
obtain expansions for 0 and other flow quantities along the given curve in powers of ~ or ~i/2. A special case of an expansion along a curve of the form x - x I = g(y2) is given by the expansions at constant potential of [15].
The equation of the
equipotential through the point (Xl, 0) is of the form
x - xI =
piy i=l
2i
(5.10)
233
The calculation of the coefficients Pi' starting from the coefficients of (2.9) and (2.10), is described in detail in [15]. We see that the special choices g(y) = ~y and g(y) = ~y2 correspond to Moran's procedure in [13].
In [13], a double power series in x and y is written
so that terms of the same degree in x and y jointly are grouped together.
The
given series is then expressed as a power series in one of the variables with coefficients which are polynomials in the ratio of the variables, is held constant in a given calculation.
and this ratio
The use of the above choices for g(y),
after making the substitution y = z I/2 in the second, is seen to be equivalent to Moran's procedure. In the preceding two cases, it is possible to obtain the coefficients a i and b i explicitly as polynomials in ~.
In the case x - x I = ey (or y = a(x - Xl)),
a more symmetrical approach is to transform (2.9) and (2.10) to polar coordinates r and 8 with origin at (Xl, 0). y
2
On substituting x - x I = r cos e and
2 = r (I - cos28) in (2.9) and (2.10), we obtain power series in r for ~/yl+O
and p with coefficients which are polynomials in cos 8.
6.
The Equation of the Sonic Line The equation of the sonic line has an expansion of the form
x
=
i= I siY
(6.1)
in the neighborhood of the origin both in the two-dimensional and axially symmetric cases.
We can calculate the coefficients s. successively by substituting 1
(6.1) in the left hand side of the equation
oijy2i-2x j - l o 1
(6.2)
i=1 j=l and equating the coefficients of powers of y2 past the constant term to zero. This calculation can be carried out on a computer by use of subroutines for power
234
series manipulations.
Finally, we can obtain power series for flow quantities
along the sonic line by substituting (6.1) or its inversion in their Taylor expansions in the neighborhood of the origin.
7.
The Equation of the Limitin$ Characteristic As shown in Figure i, the limiting characteristic is the right-running
characteristic which passes through the origin.
In both the two-dimensional and
axially symmetric cases, the two families of characteristics satisfy the equations dy/dx = tan (8±~),
(7.1)
where tan8 = - ~x/~y and tans = (M2 - 1) -1/2 . Making these substitutions, we obtain [-~x(M2-1) I/2 ± ~y]dx/dy - ~y(M2-1) I/2 ~ ~x = 0.
(7.2)
The equation of the limiting characteristic has an expansion of the form x = ~
i=l
for sufficiently
s m a l l y.
r i Y 2i
Denoting the left
hand s i d e of ( 7 . 2 ) by C, we f i n d t h a t
C= £ ei y2i-l+e i=l On s e t t i n g
(7.3)
(7.4)
c 1 = 0 and p r o c e e d i n g as i n [ 1 5 ] , we o b t a i n r 1 = (~+1)P12/4
(7.5)
r I = (/5-i) (y+l)P12/8
(7.6)
i n t h e t w o - d i m e n s i o n a l c a s e , and
in the axially symmetric case. P12 is negative.
We note that both values are negative, since
We can verify that c.i = - [2i - (y+l)Pl2/2dl]ri + c~,
where cf does not involve r.. 1 1
(7.7)
Finally, in order to determine r i for i ~ 2, we
start from (7.5) in the two-dimensional case and (7.6) in the axially symmetric case, and calculate
235
r i = c~/[2i - (y+l)Pl2/2dl] , for i = 2, 3, . . . .
For each i, the calculation of cf is carried out by setting l
r° = 0 and calculating 1
c.. 1
After the coefficients
r i have been determined,
flow quantities along the limiting characteristic inversion in their Taylor expansions
8.
(7.8)
we can find power series for
by substituting
in the neighborhood
(7.1) or its
of the origin.
Numerical Results and Discussion As in [15], calculations have been carried out on the CDC 6400 in the axially
symmetric case with y = 1.4, using the velocity on the centerline given by example
(c) of [6].
Coefficients
of the Taylor expansions
computed for i ~ 25 in these calculations, equal to 49.
and hence, with the maximum value of j
The centerline velocity of example
present coordinates
(2.9) and (2.10) are
(c) is given in terms of the
and reference quantities by u = A I + A2/[A 3 + (x-A4)2]2 ,
where A 1 = 0.0657267,
A 2 = 3.1224458,
A 3 = 1.7125400,
centerline velocity is shown in Figure 2.
(8.1) and A 4 = 0.34000641.
In the present units,
This
the value of the
stream function which defines the nozzle contour in [6] becomes ~ = 0.31104. Figure 3 shows the calculated nozzle contour, characteristic
sonic line, and limiting
for the eenterline velocity of example
(c).
Equipotential ~urves
are also shown, as calculated in [15] by means of Pade" fractions formed from expansions
of x - x I and y2 in powers of 4.
calculated by use of Pade" fractions for x = x I.
Points on the nozzle contour were
formed from expansions
The sonic line was calculated
in powers of y2 or
in two ways, by solving the equation
p = 1 by Newton's method with the left hand side replaced by a Pade" fraction, and by forming a Pade" fraction from (6.1).
In the present example,
method was found to be more accurate than the latter for y > 0.6.
the former
Finally,
the
limiting characteristic was calculated in [15] by means of a Pade" fraction formed from the right hand side of (7.3).
Comparison is made with the calculations
[6] and [9] by the method of complex characteristics.
of
236
Similarly, calculations
Figure 4 compares the flow angle calculated in [15] with the
of [6] and [9].
of Pade ~ fractions
The calculations
of [15~ were carried out by means
formed from power series in y2 for x = x I.
In Tables 1 and 2, Pade" fractions of the form fn,n(Z) are compared with the corresponding partial sums of the series y.
(4.1) for x = x I = -I and for two values of
The tables indicate that the sequence of Pade ~ fractions converges
for both
values of y, while the series converges for the smaller value of y and diverges for the larger.
Tables 3 and 4 compare Pade" fractions and partial sums at the points
(-i, I.i) and (-I, 1.6) for expansions form x - x I = ~y2
with x I = -3.
of ~ in powers of y2 along parabolas of the
Finally, Tables 5 and 6 give the same comparison
for expansions of ~ in powers of y along the rays from (-3,0). along rays through hence,
The expansions
(-3, 0) contain all powers of y up to and including y48, and
49 coefficients
are found in the present calculations.
expansions use all 625 of the coefficients
We note that these
~ij obtained in the solution of the
inverse problem, while only 325 are used by the expansions
along parabolas.
The
series converges for both values of y in Tables 3 through 6, but more slowly than the sequence of Pade" fractions. Comparison through
of Tables 4 and 6 shows that the expansion along a straight line
(-3, 0) leads to more rapid convergence of the sequence of Pade" fractions
at (-i, 1.6) than the expansion along a parabola, while both tables show more rapid convergence than Table 2.
We see that the expansion along x = -i used in Tables 1
and 2 is a special case both of an expansion along a straight line through and of an expansion along a parabola. expansion
(xl, 0) in a particular
at which the flow is c~iculated.
(-i, 0)
It follows that the most suitable center of
case is not necessarily Further calculations
the one nearest the point
show that the portion of the
nozzle contour shown in Figure 3 can be calculated to 4 figures or more for x < -0.5 by means of Pade" fractions formed from power series along rays through
(-3, 0).
The remaining portion of the nozzle contour in Figure 3 can be calculated by use of Pade" fractions formed from expansions along rays through the origin and through the point
(-0.25, 0).
This method is found to be more economical than the procedure of
237
section 4, in which ~i(x) and Pi(X) are calculated by means of Pade
fractions,
since the range of x for which the latter are sufficiently accurate becomes small for i ~ 15 in the present calculations.
The time required to find expansions along
a given ray through (Xl, 0) when ~ij and PiJ are known is much less than that for solution of the inverse problem.
9.
References i.
J. R. Baron, "Analytic Design of a Family of Supersonic Nozzles by the Friedrichs Method," WADC Report 54-279, June 1954, Naval Supersonic Lab., MIT, Cambridge, Mass.
2.
D. F. Hopkins and D. E. Hill, "Effect of Small Radius of Curvature on Transonic Flow in Axisymmetric Nozzles," AIAA Journal, Vol. 4 (1966), pp. 1337-1343.
3.
J. D. Anderson, Jr. and E. L. Harris, "Modern Advances in the Physics of Gasdynamic Lasers," AIAA Paper 72-143, San Diego, Calif., 1972.
4.
T. A. Cool, "A Summary of Recent Research on Continuous Wave Chemical Lasers," Modern Optical Methods in Gas Dynamics Research, edited by D. S. Dosanjh, Plenum Press, New York, 1971, pp. 197-220.
5.
J. V. Armitage, "Flow in a deLaval Nozzle by the Garabedian Method," ARL 66-0012, Jan. 1966, Aerospace Research Labs., Dayton, Ohio.
6.
G. V. R. Rao and B. Jaffe, "A Numerical Solution of Transonic Flow in a Convergent-Divergent Nozzle," Final Report, Contract NAS7-635, March 1969, NASA.
7.
P. R. Garabedian, "Numerical Construction of Detached Shock Waves," Journal of Mathematics and Physics, Vol. 36 (1957), pp. 192-205.
8.
P. R. Garabedian, Partial Differential Equations, Wiley, New York, 1964, Chapter 16.
9.
J . M . Solomon, private communication, Jan. 1971, Naval Surface Weapons Center, White Oak, Silver Spring, Md.
i0.
R. Courant and D. Hilbert, Methods of Mathematical Physics, Vol. II, Interscience Publishers, New York, 1962.
ii.
M. D. Van Dyke, "A Model of Supersonic Flow Past Blunt Axisymmetric Bodies with Application to Chester's Solution," Journal of Fluid Mechanics, Vol. 3 (1958), pp. 515-522.
12.
A. H. Van Tuyl, "The Use of Rational Approximations in the Calculation of Flows With Detached Shocks," Journal of the Aero/Space Sciences, Vol. 27 (1960), pp. 559-560.
238
13.
J. P. Moran, "Initial Stages of Axisymmetric Shock-on-Shock Interaction for Blunt Bodies," Physics of Fluids, Vol. 13 (1970), pp. 237-248.
14.
A. H. Van Tuyl, "Use of Pad~Fraction in the Calculation of Blunt Body Flows," AIAA Journal, Vol. 9 (1971), pp. 1431-1433.
15.
A. H. Van Tuyl, "Calculation of Nozzle Flows Using Pad~ Fractions," AIAA Journal, Vol. Ii (1973), pp. 537-541.
16.
H. S. Wall, Analytic Theory of Continued Fractions, D. Van Nostrand, New York, 1948.
17.
P. Henrici, "The Quotient-Difference Algorithm," Further Contributions to the Solution of Simultaneous Linear Equations and the Determination of Eisenvalues, National Bureau of Standards Applied Mathematics Series, No. 49, 1958, pp. 23-46.
18.
P. Wynn, "On a Device for Computing the em(Sn) Transform," Math. Tables and Other Aids to Computation, Vol. I0 (1956), pp. 91-96.
Table I.
Pade" fractions for stream function at (-i, i.i) from power series on the line x = -i.
No. of terms
ii 13 15 17 19 21 23 25
Table 2.
Series
Pade" fractions
0.22136847 0.22139025 0.22137986 0.22138595 0.22138235 0.22138439 0.22138328 0.22138386
0.22138164 0.22138385 0.22138366 0.22138367 0.22138366 0.22138367 0.22138367 0.22138367
Pade" fractions for stream function at (-I, 1.6) from power series on the line x = -i.
No. of terms
ii 13 15 17 19 21 23 25
Series
0.86601815 0.72548523 -0.44164134 1.9872070 -3.3585190 8.5084977 -17.484076 38.076299
Pade" fractions
0.32739899 0.32960800 0.32908139 0.32913508 0.32907155 0.32911728 0.32912434 0.32911631
239
Table 3.
Pade ~ fractions for stream function at (-i, i.i) from power series along parabola through (-3, 0).
No. of terms
ii 13 15 17 19 21 23 25
Table 4.
Pade" fractions
0.22082834 0.22121951 0.22135508 0.22138460 0.22138868 0.22138819 0.22138670 0.22138537
0.22139959 0.22139261 0.22140720 0.22138378 0.22138298 0.22138362 0.22138367 0.22138366
Pade" fractions for stream function at (-i. 1.6) from power series along parabola through (-3, 0).
No. of terms
ii 13 15 17 19 21 23 25
Table 5.
Series
Series
Pade ~ fractions
0.33051685 0.32947504 0.32910476 0.32904116 0.32906254 0.32909083 0.32910815 0.32911519
0.32899479 0.32910273 0.32913379 0.32911671 0.32911675 0.32912158 0.32911528 0.32911678
Pade" fractions for stream function at (-i, i.i) from power series along straight line through (-3, 0).
No. of terms
35 37 39 41 43 45 47 49
Series
Pade" fractions
0.22138349 0.22138358 0.22138358 0.22138363 0.22138366 0.22138366 0.22138366 0.22138367
0.22138363 0.22138366 0.22138367 0.22138367 0.22138367 0.22138367 0.22138367 0.22138367
240 Table 6.
Pade" fractions for stream function at (-i, 1.6) from power series along straight line through (-3, 0).
No. of terms
35 37 39 41 43 45 47 49
"~
Series
Pade" fractions
0.32910043 0.32910691 0.32910959 0.32911685 0.32911775 0.32911594 0.32911575 0.32911506
0.32913355 0.32911442 0.32911469 0.32911464 0.32911467 0.32911474 0.32911480 0.32911470
"~k
/
~1/
SONlCUne~
LIMITING
J
\,{,' 0
Figure I.
\
/ -/~
/ ', I CHARACTERISTIC
~/ X
Schematic diagram of nozzle flow.
241
U 0.5
I
I
-3.0
Figure 2.
-2.0
I
I
-I. 0
X
0
0.5
Centerline velocity distribution of example (c) of [6].
2.5
o,
~ REF. 9
NOZZLE CONTOUR
11"0'I041
Y
-3.0
Figure 3.
-2.5
-2.0
-1.5 x
-1.0
-0.5
Subsonic and transonic portions of nozzle in example (c).
242
/
20
e,-
elm
PRESENTMETHOD I; 15 ......... REF.6 I/ o REF. 9 //~i 10
I
,
,~.~>....7.,I"
0.2 Figure 4.
0.4
y
,
,
,
0.6 Ii 0.8 !
Flow angle e along sonic line in example (c).
A B I B L I O G R A P H Y ON PADE A P P R O X I M A T I O N AND SOME RELATED MATTERS Claude BREZINSKI University
of
Lille
The aim of this paper is to give a bibliography on Pad6 approximants, related matters and applications.
some
For several years, Pad6 approximants had become more and more important in mathematics, numerical analysis and various fields in physics and engineering. They are closely related to many subjects in mathematics as analytic function theory, difference equations, the theory of moments, approximation, analytic continuation, continued fractions, etc. Then a whole bibliography should be a huge one to include the corresponding references of these disciplines. I have divided the references given in this paper into three sections. The first one deals with Pad6 approximation and I hope it is quite complete. The second one is devoted to continued fractions and includes only some historical references and most of the recent papers on this subject. The thrid section contains some applications of Pad6 approximants with a special emphasis on mechanics ; I have also included references on numerical analysis and methods to accelerate the convergence of sequences. Miscellaneous references end the paper. It is obvious that this bibliography is far to be complete because of the limited number of pages of this volume. It is, in fact, less than half of a bigger bibliography on this subject and on all the related matters that I hope to publish in the future. I apologize in advance for any errors and omissions and I thank everybody who would send me any new reference on this subject.
I - PADE approximants i - R.J. ARMS, A. EDREI - The Pad6 tables and continued fractions generated by totally positive sequences - in "Mathematical essays dedicated to A.J. Macintyre" (1970) 1-21. 2 - G.A. BAKER Jr.- The Pad6 approximant method and some related generalizations in "The Pad6 approximant in theoretical physics", G.A. Baker Jr. and J.L. Gammel eds., Academic Press, New York, 1970. 3 - G.A. BAKER Jr.- The theory and application of the Pad~ approximant method J. Adv. Theor. Phys., i (1965) 1-56. 4 - G.A. BAKER Jr.- Reeursive calculation of Pad6 approximants - in "Pad~ approximants and their applications", P.R. Graves - Morris ed., Academic Press, New-York, 1973. 5 - G.A. BAKER Jr.- Best error bounds for Pad~ approximants to convergent series of Stieltjes - J. Math. Phys., i0 (1969) 814-820. 6 - G.A. BAKER Jr.- Certain invariance and convergence properties of Pad~ approximants - Rocky Mountains J. Math., 4 (1974) 141-150. 7 - G.A. BAKER Jr.- The existence and convergence of subsequences of Pad6 approximants - J. Math. Anal. Appl., 43 (1973) 498-528.
246
8
- G.A. 1975.
BAKER Jr.- E s s e n t i a l
of Pads a p p r o x i m a n t s
9 - G.A. B A K E R Jr., J.L. G A M M E L eds. A c a d e m i c Press, New York, 1972.
- Academic
- The Pads a p p r o x i m a n t
Press,
New-York,
in t h e o r e t i c a l
physics-
i0 - G.A. BAKER Jr., J.L. GAMMER, J.G. WILLS - An i n v e s t i g a t i o n of the applicability of the Pads a p p r o x i m a n t m e t h o d - J. Math. Anal. Appl., 2 (1961) 405-418. ii - M. B A R N S L E Y - The b o u n d i n g p r o p e r t i e s of the m u l t i p o i n t Pads a p p r o x i m a n t series of Stieltjes - Rocky M o u n t a i n s J. Math., 4 (1974) 331-334.
to.a
12 - M. B A R N S L E Y - The b o u n d i n g p r o p e r t i e s of the m u l t i p o i n t Pads a p p r o x i m a n t series of Stieltjes on the r e a l line - J. Math. Phys., 14 (1973) 299-313.
to a
13 - M. BARNSLEY, P.D. R O B I N S O N - Dual v a r i a t i o n a l p r i n c i p l e s x i m a n t s - J. Inst. Math. Appl., 14 (1974) 229-250. 14
-
M.
BARNSLEY,
P.D.
tions for Kirkwood (1974) 251-265.
ROBINSON
- Riseman
and P a d S - t y p e
- Pads a p p r o x i m a n t b o u n d s and a p p r o x i m a t e integral e q u a t i o n s - J. Inst. Math. Appl.,
15 - J.L. BASDEVANT - Pads a p p r o x i m a n t s - in "Methods VoI. IV, Gordon and Breach, London, 1970. 16 - J.L. B A S D E V A N T - The Pads a p p r o x i m a t i o n der Physik, 20 (1972) 283-331. 17 - A.F. B E A R D O N - On the c o n v e r g e n c e Appl., 21 (1968) 344-346.
in s u b n u c l e a r
and its p h y s i c a l
of Pads a p p r o x i m a n t s
appro-
solu14
physics",
applications
- J. Math.
- Fort.
Anal.
18 - D° BESSIS - Topics in the theory of Pads a p p r o x i m a n t s - in "Pads approximants", P.R. G r a v e s - M o r r i s ed., The institute of physics, London, 1973. 19 - D. BESSIS, J.D. T A L M A N - V a r i a t i o n a l a p p r o a c h to the t h e o r y of o p e r a t o r Pads a p p r o x i m a n t s - Rocky M o u n t a i n s J. Math., 4 (1974) 151-158. 20 - C. BREZINSKI - C o n v e r g e n c e of Pads a p p r o x i m a n t s for some special sequences Thrid C o l l o q u i u m on a d v a n c e d c o m p u t i n g m e t h o d s in t h e o r e t i c a l physics, Marseille, 1973. 21 - C. BREZINSKI - Rhombus f r a c t i o n s - to appear. 22 - C. BREZINSKI to appear.
algorithms
- Computation
of Pads
23 - C. BREZINSKI - SSries de Stieltjes m e c h 58, Toulon, 12-14 m a i 1975.
connected
-
to the Pads table and c o n t i n u e d
approxlmants
and c o n t i n u e d
et a p p r o x i m a n t s
de Pads
24 - C. BREZINSKI Linear Algebra,
- Some r e s u l t s in the t h e o r y of the v e c t o r 8 (1974) 77-86.
25 - C. BREZINSKI Rocky M o u n t a i n s
- Some r e s u l t s and a p p l i c a t i o n s J. Math., 4 (1974) 335-338.
about
26 ~ C. BREZINSKI - G ~ n S r a l i s a t i o n s de la t r a n s f o r m a t i o n de Pads et de l ' e - a l g o r i t h m e - C a l c o l o (to appear).
fractions
- Colloque
~-algorithm
the vector
de Shanks,
Euro-
-
~-algorlthm
de la table
-
247
27 - J.S.R. ximants",
CHISHOLM - Mathematical P.R. G r a v e s - M o r r i s ed.,
t h e o r y of P a d 6 a p p r o x i m a n t s - in "Pad6 a p p r o The i n s t i t u t e of P h y s i c s , L o n d o n , 1973.
28 - J.S.R. C H I S H O L M - R a t i o n a l a p p r o x i m a n t s Math. Comp., 27 (1973) 841-848.
defined
29 - J.S.R. C H I S H O L M - A p p r o x i m a t i o n by sequences of m e r o m o r p h y - J. Math. Phys., 7 (1966) 39-44,
from
of P a d 6
double
power
approximants
series
-
in r e g i o n s
30 - J.S.R. C H I S H O L M - C o n v e r g e n c e p r o p e r t i e s of P a d 6 a p p r o x i m a n t s - in "Pad6 approximants and their applications", P.R. G r a v e s - M o r r i s ed., A c a d e m i c P r e s s , N e w - Y o r k , 1973. 31 - C.K. CHUI, 0. S H I S H A , P.W. S M I T H - P a d 6 rational approximants - J. A p p r o x . T h e o r y ,
approximants as l i m i t s 12 (1974) 2 0 1 - 2 0 4 .
32 - A.K. C O M M O N - P a d 6 a p p r o x i m a n t s P h y s . , 9 (1968) 32-38.
and bounds
33 - A.K. C O M M O N , P.R. G R A V E S - M O R R I S J. Inst. Math. A p p l i c s . , 13 (1974)
- Some properties 229-232.
34 - J.D.P. D O N N E L L Y - The P a d 6 t a b l e D.C. H a n d s c o m b ed., P e r g a m o n P r e s s ,
to s e r i e s
of b e s t
of S t i e l ~ e s
of C h i s h o l m
- in " M e t h o d s of n u m e r i c a l N e w - Y o r k , 1966.
- J. Math.
approximants
-
approximation",
35 - A. E D R E I - The P a d 6 t a b l e of m e r o m o r p h i c f u n c t i o n s of s m a l l o r d e r w i t h n e g a t i v e z e r o s a n d p o s i t i v e p o l e s - R o c k y M o u n t a i n s J. M a t h . , 4 (1974) 1 7 5 - 1 8 0 . 36 - A. E D R E I - C o n v e r g e n c e de la m 6 t h o d e m o r p h e s - C o l l o q u e E u r o m e c h 58, T o u l o n ,
de P a d 6 a p p l i q u 6 e s 1 2 - 1 4 m a i 1975.
aux fonctions
37 - D. E L L I O T - T r u n c a t i o n e r r o r s in P a d 6 a p p r o x i m a t i o n s to c e r t a i n an a l t e r n a t i v e a p p r o a c h - Math. Comp., 21 (1967) 3 9 8 - 4 0 6 .
functions
38 - S.T. E P S T E I N , M. B A R N S L E Y - A v a r i a t i o n a l a p p r o a c h to the t h e o r y point Pad6 approximants - J. Math. P h y s . , 14 (1973) 3 1 4 - 3 2 5 . 39 - W. F A I R - P a d 6 a p p r o x i m a t i o n Math. Comp., 18 (1964) 627-634.
to the
solution
40 - W. F A I R , Y.L. L U K E - P a d 6 a p p r o x i m a t i o n s Numer. M a t h . , 14 (1970) 3 7 9 - 3 8 2 . 41 - J. F L E I S C H E R - N o n l i n e a r P h y s . , 14 (1973) 246-248.
Pad6
of the R i c c a t i
to t h e o p e r a t o r
approximants
for
Legendre
m6ro-
of m u l t i -
equation
-
exponential
series
:
-
- J. Math.
42 - J. F L E I S C H E R - N o n l i n e a r P a d 6 a p p r o x i m a n t s f o r L e g e n d r e s e r i e s - in " P a d 6 approximants and their applications", P.R. G r a v e s - M o r r i s ed., A c a d e m i c P r e s s , N e w - Y o r k , 1973. 43 - J. F L E I S C H E R - G e n e r a l i z a t i o n s of P a d 6 a p p r o x i m a n t s - in " P a d 6 P.R. G r a v e s - M o r r i s ed., The i n s t i t u t e of p h y s i c s , L o n d o n , 1973. 44 - N.R. Theory,
F R A N Z E N - Some c o n v e r g e n c e 6 (1972) 254-263.
results
for Pad6
approximants
45 - N.R. F R A N Z E N - C o n v e r g e n c e o f P a d 6 a p p r o x i m a n t s for a certain m e r o m o r p h i c f u n c t i o n s - J. A p p r o x . T h e o r y , 6 (1972) 264-271.
approximants",
- J. A p p r o x .
class
of
248
46 - J.L. GAMMEL - Review of two recent g e n e r a l i z a t i o n s of the Pad6 approximant in "Pad6 approximants and their applications", P.R. G r a v e s - M o r r i s ed., Academic Press, New-York, 1973. 47 - J.L. GAMMEL - Effect of r a n d o m errors (noise) in the terms of a power series on the convergence of the Pad~ a p p r o x i m a n t s - in "Pad~ approximants", P.R. G r a v e s - M o r r i s ed., The institute o f physics, London, 1973. 48 - J.L. GAMMEL, J. N U T T A L L - C o n v e r g e n c e of Pad~ approximants to quasi-analytic functions beyond n a t u r a l b o u n d a r i e s - J. Math. Anal. Appl. 49 - J. G I L E W I C Z - N u m e r i c a l detection of the best Pad~ a p p r o x i m a n t and determination of the Fourier coefficients of the i n s u f f i c i e n t l y sampled functions - in "Pad~ approximants and their applications", P.R. G r a v e s - M o r r i s ed., Academic Press, New-York, 1973. 50 - J. GILEWICZ - Totally monotonic and totally positive sequences for the Pad~ approximants m e t h o d - Colloque E u r o m e c h 58, Toulon, 12-14 mai 1975. 51 - J.E. GOLDEN, J.H. McGUIRE, J. N U T T A L L - C a l c u l a t i n g Bessel functions w i t h Pad~ approximants - J. Math. Anal. Appl., 43 (1973) 754-767. 52 - W.B. GRAGG - The Pad~ table and its r e l a t i o n to certain algorithms of numer i c a l analysis - SIAM Rev., 14 (1972) 1-62. 53 - W.B. GRAGG - On Hadamar's theory of polar singularities - in "Pad~ approxim a n t s and their applications", P.R. G r a v e s - M o r r i s ed., A c a d e m i c Press, New-York, 1973. 54 - W.B. GRAGG, G.D. J O H N S O N - The Laurent - Pad~ table - P r o c e e d i n g s IFIP Congress, North-Holland, 1974. 55 ~ P.R. G R A V E S - M O R R I S ed. - Pad~ approximants and their applicationsA c a d e m i c Press, New-York, 1973. 56 - P.R. G R A V E S - M O R R I S ed. - Pad~ approximants - The institute of physics, London, 1973. 57 - T.N.E. GREVILLE - On some conjectures of P. Wynn concerning the E-algorithm MRC Technical summary report 877, Madison, 1968. 58 - A.S. H O U S E H O L D E R - The Pad~ table, the Frobenius identities and the q-d a l g o r i t h m - Linear Algebra, 4 (1971) 161-174. 59 - A.S. HOUSEHOLDER, G.W. STEWART - Bigradients, Hankel determinants and the Pad~ table - in "Constructive aspects of the f u n d a m e n t a l t h e o r e m of algebra", B. D e j o n and P. Henrici eds., A c a d e m i c Press, New-York, 1969. 60 - R.C. JOHNSON - A l t e r n a t i v e a p p r o a c h to Pad~ approximants - in "Pad~ appreximants and their application", P.R. G r a v e s - M o r r i s ed., Academic Press, NewYork, 1973. 61 - W.B. JONES - T r u n c a t i o n error b o u n d for continued fractions and Pad~ approximants - in "Pad~ approximants and their applications", P.R. G r a v e s - M o r r i s ed., A c a d e m i c Press, New-York, 1973. 62 - W.B. JONES - Analysis of t r u n c a t i o n error of a p p r o x i m a t i o n s b a s e d on the Pad~ table and continued fractions - R o c k y Mountains J. Math., 4 (1974) 241-250.
249
63 - W.B. JONES, W.J. THRON - On c o n v e r g e n c e Anal., 6 (1975) 9-16. 64 - I.M. LONGMAN (1971) 53-64.
- Computation
Of Pad6 a p p r o x i m a n t s
of the Pad6 table
65 - Y.L. LUKE - E v a l u a t i o n of the g a m m a f u n c t i o n SIAM J. Math. Anal., i (1970) 266-281. 66 - Y.L. LUKE - The Pad6 table and the T - m e t h o d 110-127. 67 - A. MAGNUS - Certain c o n t i n u e d Math. Z., 78 (1960) 361-374. 68 - A. MAGNUS - P - f r a c t i o n s 4 (1974) 257-260.
fractions
- Intern.
by means
and the Pad6 table
J. comp.
Math.,
3B
of Pad6 a p p r o x i m a t i o n s -
- J. Math.
associated
- SIAM J. Math.
Phys.,
37 (1958)
w i t h the Pad6 table -
- Rocky M o u n t a i n s
J. Math.,
69 - D. M A S S O N - Hilbert space and Pad~ a p p r o x i m a n t - in "The Pad~ a p p r o x i m a n t in t h e o r e t i c a l physics", G.A. Baker Jr. and J.L. G a m m e l eds., A c a d e m i c Press, New-York, 1970. 70 - D. M A S S O N - Pad6 a p p r o x i m a n t s and Hilbert spaces - in "Pad6 a p p r o x i m a n t s and their a p p l i c a t i o n s " , P.R. G r a v e s - M o r r l s ed., A c a d e m i c Press, 1973. 71 - J.H. M c C A B E Pad6 q u o t i e n t s
- A formal e x t e n s i o n of the Pad6 table to include - J. Inst. Math. Applies., 15 (1975) 363-372.
two point
72 - J.B.
- A note on the e - a l g o r i t h m
17-24.
McLEOD
- Computing,
7 (1971)
73 - G. MERZ - Pad6sche N [ h e r u n g s b r [ c k e und I t e r a t i o n s v e r f a h r e n h ~ h e r e n D o c t o r a l thesis, T e c h n i s c h e Hochschule Clausthal, Germany, 1967.
Ordnung
-
74 - J. N U T T A L L - The c o n n e c t i o n of Pad6 a p p r o x i m a n t s w i t h s t a t i o n a r y v a r i a t i o n a l p r i n c i p l e s and the c o n v e r g e n c e of c e r t a i n Pad6 a p p r o x i m a n t s - in "The Pad6 a p p r o x i m a n t in t h e o r e t i c a l physics", G.A. Baker Jr. and J.L. G a m m e l eds., A c a d e m i c Press, New-York, 1970. 75 - J. N U T T A L L - The c o n v e r g e n c e J. Math. Anal. Appl., 31 (1970)
of Pad6 a p p r o x i m a n t s 147-153.
of m e r o m o r p h i c
functions
-
76 - J. N U T T A L L - V a r i a t i o n a l p r i n c i p l e s and Pad6 a p p r o x i m a n t s - in "Pad6 approximants and their applications", P.R. G r a v e s - M o r r i s ed., A c a d e m i c Press, New-York, 1973. 77 - J. N U T T A L L - The c o n v e r g e n c e J. Math., 4 (1974) 269-272.
of c e r t a i n
Pad6 a p p r o x i m a n t s
78 - H. PADE - Sur la r e p r 6 s e n t a t i o n a p p r o c h 6 e r a t i o n n e l l e s - Ann. Ec. Norm. Sup., 9 (1892) 79 - C. P O M M E R E N K E - Pad6 a p p r o x i m a n t s Anal. Appl., 41 (1973) 775-780.
d'une f o n c t i o n 1-93.
and c o n v e r g e n c e
- Rocky Mountains
par des f r a c t i o n s
in c a p a c i t y
- J. Math.
80 - L.D. PYLE - A g e n e r a l i z e d inverse c - a l g o r i t h m for c o n s t r u c t i n g i n t e r s e c t i o n p r o j e c t i o n matrices, with a p p l i c a t i o n s - Numer. Math., i0 (1967) 86-102. 81 - J. R I S S A N E N
- Recursive
evaluation
of Pad6 a p p r o x i m a n t s
for m a t r i x
sequences-
250
IBM J.
Res.
develop.,
(july
1972)
401-406.
82 - E.B. S A F F - An e x t e n s i o n of M o n t e s s u s de B a l l o r e ' s t h e o r e m on t h e c o n v e r g e n c e of i n t e r p o l a t i n g r a t i o n a l f u n c t i o n s - J. Approx. T h e o r y , 6 (1972) 63-67. 83 - W.F. T R E N C H - An a l g o r i t h m f o r the i n v e r s i o n S I A M J. Appl. M a t h . , 13 (1965) 1 1 0 2 - 1 1 0 7 . 84 - R.P. W A N DE R I E T - On c e r t a i n P a d ~ f u n c t i o n - r e p o r t TN 32, M a t h e m a t i c a l 85 - H. V A N R O S S U M - A t h e o r y V a n G o r a i m , A s s e n , 1953.
of f i n i t e
Hankel
polynomials in c o n n e c t i o n c e n t e r , A m s t e r d a m , 1963.
of o r t h o g o n a l
matrices
with
the b l o c k
polynomials
based
on the P a d 6
86 - H.S. W A L L - On the r e l a t i o n s h i p a m o n g the d i a g o n a l Bull. Amer. Math. Soc., 38 (1932) 7 5 2 - 7 6 0 .
files
of a P a d 6
87 - H.S. W A L L - On the P a d 6 a p p r o x i m a n t s associated a n d s e r i e s of S t i e l t j e s - Trans. Amer. Math. Soc., 88 - H.S. W A L L - On the P a d ~ w e r s e r i e s - Trans. Amer. 89 - H.S. ximants
WALL - General - Trans. Amer.
w i t h the c o n t i n u e d 31 (1929) 91-116.
approximants associated with a positive Math. Soc., 33 (1931) 511-532.
t h e o r e m s on t h e c o n v e r g e n c e of s e q u e n c e s Math. Soc., 34 (1932) 4 0 9 - 4 1 6 .
90 - H. W A L L I N - On the c o n v e r g e n c e t h e o r y of P a d 6 a p p r o x i m a n t s the O b e r w o l f a c h c o n f e r e n c e on a p p r o x i m a t i o n t h e o r y , 1971. 91 - H. W A L L I N - The c o n v e r g e n c e of P a d ~ a p p r o x i m a n t s a n d t h e s e r i e s c o e f f i c i e n t s - A p p l i c a b l e A n a l . , 4 (1974) 2 3 5 - 2 5 2 . 92 - J.L~ W A L S H approximation
- P a d ~ a p p r o x i m a n t s as l i m i t of r a t i o n a l - J. Math. M e c h . , 13 ( 1 9 6 4 ) 305-312.
-
table
table
-
fraction
definite
of P a d 6
po-
appro-
- Proceedings
size
-
of
of the p o w e r
functions
of b e s t
93 - J.L. W A L S H - P a d 6 a p p r o x i m a n t s as l i m i t s of r a t i o n a l f u n c t i o n s of b e s t approximation, r e & l d o m a i n - J. A p p r o x . T h e o r y , ii (1974) 2 2 5 - 2 3 0 . 94 - P. W Y N N - U p o n t h e i n v e r s e of f o r m a l p o w e r s e r i e s o v e r c e r t a i n C e n t r e de r e c h e r c h e s m a t h ~ m a t i q u e s , U n i v e r s i t ~ de M o n t r 6 a l , 1970. 95 - P. W Y N N - U p o n t h e g e n e r a l i z e d i n v e r s e of a f o r m a l p o w e r v a l u e d c o e f f i c i e n t s - C o m p o s i t i o M a t h . , 23 (1971) 4 5 3 - 4 6 0 . 96 - P. W Y N N C.R. Acad.
- Sur l ' 6 q u a t i o n a u x d 6 r i v 6 e s p a r t i e l l e s Sc. P a r i s , 278 A (1974) 8 4 7 - 8 5 0 .
99 - P. W Y N N - A g e n e r a l s y s t e m 18 ser. 2 (1967) 69-81. i00 - P. W Y N N
- Some recent
recursions
of orthogonal
developments
series
de la s u r f a c e
97 - P. W Y N N - U b e r f i n e n i n t e r p o l a t i o n s - algorithmus d i e in d e r t h e o r i e d e r i n t e r p o l a t i o n d u t c h r a t i o n a l e Numer. M a t h . , 2 (1961) 1 5 1 - 1 8 2 . 98 - P. W Y N N - D i f f e r e n c e - d i f f e r e n t i a l L o n d o n Math. Soc., 23 (1971) 2 8 3 - 3 0 0 .
algebras
und gewise funktionen
for Pad6
polynomials
in the t h e o r i e s
with vector
de Pad6
-
andere formeln, bestehen -
quotients
~ Quart.
-
- Proc.
J. M a t h . ,
of c o n t i n u e d
fractions
251
a n d the P a d 6 t a b l e
- Rocky Mountains
J. M a t h . ,
4 (1974)
297-324.
i01 - P. W Y N N - U p o n the d i a g o n a l s e q u e n c e s s u m m a r y r e p o r t 660, M a d i s o n , 1966.
of t h e P a d 6 t a b l e
102 - P. W Y N N - E x t r e m a l p r o p e r t i e s H u n g a r i c a e , 25 (1974) 2 9 1 - 2 9 8 .
quotients
of P a d 6
- MRC technical
- A c t a Math.
Acad.
103 - P. W Y N N - U p o n a c o n v e r g e n c e r e s u l t in the t h e o r y Trans. Amer. Math. Soc., 165 (1972) 2 3 9 - 2 4 9 .
of t h e P a d 6
104 - P. W Y N N - Z u r t h e o r i e d e r m i t g e w i s s e n s p e z i e l l e n P a d ~ s c h e n t a f e l n - Math. Z., 109 (1969) 66-70.
funktionen
105 - P. W Y N N - U p o n the P a d 6 t a b l e d e r i v e d J. Numer. A n a l . , 5 (1968) 8 0 5 - 8 3 4 .
from a Stieltjes
106 - P. W Y N N - U p o n s y s t e m s of r e c u r s i o n s w h i c h o b t a i n the P a d 6 t a b l e - N u m e r . M a t h . , 8 (1966) 2 6 4 - 2 6 9 . 107 - P. W Y N N - L ' s - a l g o r i t m o (1961) 4 0 3 - 4 0 8 .
e la t a v o l a
di P a d 6
108 - P. W Y N N - U p o n a c o n j e c t u r e c o n c e r n i n g equations, and certain other matters - MRC M a d i s o n , 1966. 109 - P. W Y N N - G e n e r a l p u r p o s e M a t h . , 6 (1964) 22-36.
vector
for
II - C o n t i n u e d
algorithm
iterated
di Mat.
114 - D. B E R N O U L L I N o v i comm. Acad.
- D i s q u i s i t i o n e s u l t e r i o r e s de i n d o l e Sc. Imp. St P 6 t e r s b o u r g , 20 (1775)
20
problems
related
to t h e m
de f r a c t i o n i b u s
of c o n t i n u e d
fractionum
fractions
continues
fractions
- Nouvelles
CHEBYCHEV
- Sur les f r a c t i o n s
continues
- Jour.
-
continuarum
-
- SIAM Rev.,
Ann.
de M a t h . ,
-
continuis
Irra~ionalit[t unendlieher Kettenbr~che Z qV x v - M a t h . A n n . , 76 ( 1 9 1 5 ) 2 9 5 - 3 0 0 .
evaluation
-
continued
de M a t h . ,
(3) 6 (1887) 118 - P.
of
- Numer.
and matrix
and algorithms
- Adversaria analytica miscellanea Sc. Imp. St P 6 t e r s b o u r g , 20 (1775)
- Sur q u e l q u e s
Roma,
procedures
vector
i13 - D. B E R N O U L L I N o v i comm. Acad.
117 - E. C E S A R O
- SIAM
the q u o t i e n t s
112 : F.L. B A U E R , E. F R A N K - N o t e on f o r m a l p r o p e r t i e s o f c e r t a i n f r a c t i o n s - Proc. Amer. Math. Soc., 9 (1958) 3 4 0 - 3 4 7 .
116 - G. B L A N C H - N u m e r i c a l 7 (1964) 3 8 3 - 4 2 1 .
verkn~pften
fractions
iii - F.L. B A U E R - Use of c o n t i n u e d f r a c t i o n s C I M E s u m m e r s c h o o l l e c t u r e s , P e r i g i a , 1964.
115 - S. B E R N S T E I N , 0. S Z A S Z - U b e r m i t e i n e r A n w e n d u n g a u f die R e i h e
-
a method for solving linear t e c h n i c a l s u m m a r y r e p o r t 626,
epsilon
ii0 - P. W Y N N - A c c e l e r a t i o n t e c h n i q u e s Math. C o m p . , 16 (1962) 3 0 1 - 3 2 2 .
- Rend.
table
series
among
Sci.
set.
11,3
252
(1858)
289-323.
119 - V.F. C O W L I N G , general continued
W. L E I G H T O N , W.J. T H R O N - T w i n c o n v e r g e n c e r e g i o n s f o r t h e f r a c t i o n - Bull. Amer. Math. Soc., 49 (1943) 913-916.
120 - I.V. C Y G A N K O V - S o l u t i o n o f R i c c a t i e q u a t i o n s by c o n t i n u e d Perm. Gos. Univ. Ucen. Zap. M a t . , 17 (1960) 99-107. 121 - I.V. C Y G A N K O V f r a c t i o n s - Perm.
- Solution Gos. Univ.
fractions
of a s p e c i a l R i c c a t i e q u a t i o n b y c o n t i n u e d Ucen. Zap. M a t . , 17 (1960) 109-113.
122 - J.D.P. D O N N E L L Y - C o n t i n u e d f r a c t i o n s - in " M e t h o d s o f n u m e r i c a l m a t i o n " , D.C. H a n d s c o m b ed., P e r g a m o n P r e s s , O x f o r d , 1965. 123 - H.G. E L L I S - C o n t i n u e d f r a c t i o n s s o l u t i o n s o f the g e n e r a l t i a l e q u a t i o n - R o c k y M o u n t a i n s J. M a t h . , 4 (1974) 3 5 3 - 3 5 6 . 124 - L. E U L E R - De f r a c t i o n i b u s St P ~ t e r s b o u r g , ii (1739).
continuis
125 - L. E U L E R - De f o r m a t i o n e St P ~ t e r s b o u r g , 1779.
fractionum
126 - W. (1971)
continued
FAIR - Noncommutative 226-232.
127 - W. F A I R - A c o n v e r g e n c e J. A p p r o x . T h e o r y , 5 (1972)
theorem 74-76.
observationes
continuarum
fractions
Acad.
Acad.
Sc.
- S I A M J. Math.
129 - W. F A I R - C o n t i n u e d f r a c t i o n s o l u t i o n to F r e d h o l m R o c k y M o u n t a i n s J. M a t h . , 4 (1974) 3 5 7 - 3 6 0 .
FRANK - Corresponding 89-108.
133 - E. diques
GALOIS - Ann.
type
continued
fractions
Imp.
2
equations
131 - D.A. F I E L D , W.B. J O N E S - A p r i o r i e s t i m a t e s for t r u n c a t i o n n u e d f r a c t i o n s K ( i / b n) - Numer. M a t h . , 19 (1972) 2 8 3 - 3 0 2 .
Imp.
-
in a B a n a c h
truncation error estimates for continued J. M a t h . , 4 (1974) 3 6 1 - 3 6 2 .
132 - E. (1946)
Sc.
fractions
equation
integral
differen-
Anal.,
continued
to the R i c c a t i 318-323.
approxi-
Riccati
- Comm.
- Acta
for noncommutative
128 - W. F A I R - C o n t i n u e d f r a c t i o n s o l u t i o n a l g e b r a - J. Math. Anal. A p p l . , 39 (1972)
130 - D.A. F I E L D - A p r i o r i K ( i / b n) - R o c k y M o u n t a i n s
-
-
fractions
error
of c o n t i -
- Amer.
J. M a t h . ,
68
- D~monstration d ' u n t h ~ o r ~ m e sur les f r a c t i o n s Math. P u r e s et A p p l . , 19 ( 1 8 2 8 - 1 8 2 9 )
continues
p~rio-
134 - H.L. G A R A B E D I A N , H.S. W A L L - T o p i c s in c o n t i n u e d f r a c t i o n s a n d s u m m a b i l i t y N o r t h w e s t e r n Univ. S t u d i e s in Math. a n d phys. s c i e n c e s , Vol. i, E v a n s t o n a n d C h i c a g o , (1941) 89-132. 135 - I. G A R G A N T I N I , P. H E N R I C I - A c o n t i n u e d f r a c t i o n a l g o r i t h m f o r the c o m p u t a t i o n of h i g h e r t r a n s c e n d e n t a l f u n c t i o n in t h e c o m p l e x p l a n e - Math. C o m p . , 21 (1967) 18-29. 136 - T.F. G L A S S , W. L E I G H T O N - On the c o n v e r g e n c e Bull. Amer. Math. Soc., 49 (1943) 133-135. 137 - W.B.
GRAGG
- Truncation
error
bounds
of a c o n t i n u e d
for ~fraetions
- Bull.
fraction
Amer.
-
Math.
253
Soc.,
76 (1970)
1091-1094.
138 - W.B. G R A G G - M a t r i x i n t e r p r e t a t i o n s and a p p l i c a t i o n s f r a c t i o n s a l g o r i t h m - R o c k y M o u n t a i n s J. M a t h . , 4 (1974)
of the c o n t i n u e d 213-226.
139 - W.B. G R A G G - T r u n c a t i o n ii (1968) 3 7 0 - 3 7 9 .
- Numer.
error
bounds
for g-fractions
140 - M. H A M B U R G E R - U e b e r di K o n v e r g e n z e i n e s K e t t e n b r u c h s - Math. Ann., 81 (1920) 31-45.
mit
141 - H. H A M B U R G E R - B e i t r [ g e K e t t e n b r [ c h e - Math. Z e i t . ,
zur K o n v e r g e n z t h e o r i e 4 (1919) 1 8 6 - 2 2 2 .
142 - T.L. 7 (1965)
fraction
HAYDEN - Continued 292-309.
143 - T.L. H A Y D E N M a t h . , 4 (1974)
- Continued 367-370.
145 - E. (1922)
HELLINGER 18-29.
- Zur
der
approximation
fractions
144 - T.L. H A Y D E N - A c o n v e r g e n c e Math. Soc., 14 (1963) 546-552.
einer
in B a n a c h
problem
Stieltjesschen
for
Potenzreihe
Stieltjes
spaces
147 - P. H E N R I C I , P. P F L U G E R - T r u n c a t i o n e r r o r f r a c t i o n s - Numer. M a t h . , 9 (1966) 120-138.
J.
fractions
- Proe.
Amer.
- Math.
continued 1965. for
continued
Ann.,
- 0ber Math.,
eine besondere Art der 12 (1889) 3 6 7 - 4 0 5 .
152 - T.H. J E F F E R S O N - T r u n c a t i o n Numer. A n a l . , 6 (1969) 3 5 9 - 3 6 4 . 153 - W.B. J O N E S Ph.D. thesis,
error
- Contributions to the Vanderbilt University,
estimates
fractions
- Doctoral
continued
reeller
i n t e g r a l e f.~ e -x2 dx 12 ( 1 8 3 4 ) , A 346-347.
for
T-fractions
- SIAM
J.
t h e o r y of T h r o n c o n t i n u e d f r a c t i o n s N a s h v i l l e , T e n n e s s e e , 1963.
154 - W.B. J O N E S , R.I. S N E L L - T r u n c a t i o n S I A M J. Numer. A n a l . , 6 (1969) 2 1 0 - 2 2 1 . 155 - W.B. J O N E S , R.I. S N E L L f r a c t i o n s K ( a n / l ) - Trans,
-
Stieltjes
Kettenbruchentwicklung
151 - C.G.J. J A C O B I - De f r a c t i o n e c o n t i n u e , in q u a m e v o l d e r e l i c e t - J. f i r die r e i n e u. a n g e w , m a t h . ,
86
fractions
149 - K.L. H I L L A M , W.J. T H R O N - A g e n e r a l c o n v e r g e n c e c r i t e r i o n f o r f r a c t i o n s K ( a n / b n) - Proc. Amer. Math. Soc., 16 (1965) 1 2 5 6 - 1 2 6 2 . 150 - A. H U R W I T Z GrSssen - Acta
Math.,
Mountains
estimates
148 - K.L. H I L L A M - S o m e c o n v e r g e n c e c r i t e r i a f o r t h e s i s , U n i v e r s i t y o f C o l o r a d o , B o u l d e r , 1962.
- Numer.
- Rocky
Kettenbruchtheorie
146 - P. H E N R I C I - E r r o r b o u n d s f o r c o m p u t a t i o n s w i t h in " E r r o r in d i g i t a l c o m p u t a t i o n " , Wiley, New-York,
assoziierten
' sehen
to f u n c t i o n s
continued
Math.,
- Sequences Amer. Math.
156 - W.B. J O N E S , W.J. T H R O N - C o n v e r g e n c e M a t h . , 20 (1968) 1 0 3 7 - 1 0 5 5 .
error
bounds
for
continued
-
fractions
of c o n v e r g e n c e r e g i o n s f o r c o n t i n u e d Soc., 1 7 0 (1972) 4 8 3 ~ 4 9 7 . of c o n t i n u e d
fractions
- Canad.
J.
-
254
157 - W.B. K(an/l)
J O N E S , W.J. T H R O N - T w i n c o n v e r g e n c e r e g i o n s - Trans. Amer. Math. Soc., 150 (1970) 9 3 - 1 1 9
158 - W.B. J O N E S , W.J. continued fractions
T H R O N - A p o s t e r i o r i b o u n d s f o r the t r u n c a t i o n - S I A M J. N u m e r . A n a l . , 8 ( 1 9 7 1 ) 6 9 3 - 7 0 5 .
159 - W.B. J O N E S , W.J. T H R O N - N u m e r i c a l s t a b i l i t y f r a c t i o n s - Math. C o m p . , 28 (1974) 7 9 5 - 8 1 0 . 160 - W.B. J O N E S , W.J. 166 (1966) i 0 6 - i 1 8 . 161 - J.Q. J O R D A N , W. continued fraction (1938) 8 7 2 - 8 8 2 . 162 - A. Ya.
for continued
THRON
- Further
properties
in e v a l u a t i n g
fractions
error
of
continued
of T - f r a c t i o n s
- Math.
Ann.,
L E I G H T O N - On the p e r m u t a t i o n of the c o n v e r g e n t s of a a n d r e l a t e d c o n v e r g e n c e c r i t e r i a - Ann. M a t h . , (2) 39
KHINTCHINE
- Continued
fractions
- P. N o o r d h o f f ,
Groningen,
1963.
163 - A.N. K H O V A N S K Z I - The a p p l i c a t i o n o f c o n t i n u e d f r a c t i o n s a n d t h e i r g e n e r a l i z a t i o n s to p r o b l e m s in a p p r o x i m a t i o n t h e o r y - P. N o o r d h o f f , G r o n i n g e n , 1963. 164 - E. L A G U E R R E - Sur la r @ d u c t i o n & t e n d u e de f o n c t i o n s - C.R. Acad.
en f r a c t i o n s c o n t i n u e s Sc. P a r i s , 87A (1878).
d'une
classe
assez
165 - E. L A G U E R R E - Sur la r @ d u c t i o n en f r a c t i o n s c o n t i n u e s de e F ( x ) , F(x) d @ s i g n a n t u n p o l y n S m e e n t i e r - J. Math. p u r e s et a p p l . , (3) 6 ( 1 8 8 0 ) . 166 - R.E. L A N E - The v a l u e J., 12 (1946) 2 0 7 - 2 1 6 .
for continued
fractions
- D u k e Math.
167 - R.E. L A N E - The c o n v e r g e n c e a n d v a l u e s of p e r i o d i c Bull. Amer. Math. Soc., 51 (1945) 2 4 6 - 2 5 0 .
continued
fractions
168
region
problem
- L.J. L A N G E - On a f a m i l y of t w i n c o n v e r g e n c e f r a c t i o n s - Iii. J. of M a t h . , i0 (1966) 97-108.
169 - L.J. regions 170 - W. Soc.,
regions
-
for c o n t i n u e d
L A N G E , W.J. T H R O N - A t w o - p a r a m e t e r f a m i l y of b e s t t w i n c o n v e r g e n c e f o r c o n t i n u e d f r a c t i o n s - Math. Z e i t . , 73 (1969) 2 7 7 - 2 8 2 .
LEIGHTON - A test-ratio 45 (1939) 97-100.
171 - W. L E I G H T O N - C o n v e r g e n c e 5 (1939) 2 9 8 - 3 0 8 .
test for
theorems
continued
fractions
for continued
- Bull.
fractions
172 - W. L E I G H T O N - S u f f i c i e n t c o n d i t i o n s f o r t h e c o n v e r g e n c e f r a c t i o n - D u k e Math. J., 4 (1938) 7 7 5 - 7 7 8 . 173 - W. L E I G H T O N , W.T. S C O T T - A g e n e r a l Amer. Math. Soc., 45 (1939) 5 9 6 - 6 0 5 . 174 - W. L E I G H T O N , W.J. T H R O N - On v a l u e Amer. Math. Soc., 48 (1942) 9 1 7 - 9 2 0 . 175 - W. L E I G H T O N , W.J. T H R O N - C o n t i n u e d Math. J., 9 (1942) 7 6 3 - 7 7 2 . 176 - W. L E I G H T O N ,
W.J.
THRON
continued
regions
fraction
for
fractions
- On the c o n v e r g e n c e
Math.
- D u k e Math.
J.,
of a c o n t i n u e d
expansion
continued
with
Amer.
fractions
complex
of continued
- Bull.
elements
fractions
- Bull.
- Duke
to
255
meromorphic
functions
- Ann.
Math.,
(2)
44
(1943)
177 - J.S. M A C N E R M E Y - I n v e s t i g a t i o n s c o n c e r n i n g f r a c t i o n s - D u k e Math. J., 26 (1959) 6 6 3 - 6 7 8 . 178 - A. M A G N U S - E x p a n s i o n (1962) 2 0 9 - 2 1 6 .
of power
179 - A. M A G N U S - On P - e x p a n s i o n s S e l s k a b s S k r i f t e r (1964), n°3.
series
of p o w e r
80-89.
positive
into P-fractions
series
- Det.
180 - A. M A G N U S - The c o n n e c t i o n b e t w e e n P - f r a c t i o n s Prec. Amer. Math. $oc., 25 (1970) 6 7 6 - 6 7 9 .
- Math.
Kgl.
Norske
and associated
181 - A. M A R K O V - D e u x d ~ m o n s t r a t i o n s de la c o n v e r g e n c e c o n t i n u e s - A c t a M a t h . , 19 (1895) 93 - 104.
de c e r t a i n e s
182 - A. M A R K O V - N o t e sur les f r a c t i o n s c o n t i n u e s - Bull. Acad. Imp. Sc. St P ~ t e r s b o u r g , 5 (2) (1895) 9-13. 183 - D.F. M A Y E R S - E c o n o m i z a t i o n n u m e r i c a l a p p r o x i m a t i o n " , D.C.
definite
classe
continued
Z.,
Videnskabers
fractions
- On t r u n c a t i o n A n a l . , 3 (1966)
-
fractions
Physico-Math.
of c o n t i n u e d f r a c t i o n s - in " M e t h o d s of H a n d s c o m b ed., P e r g a m o n P r e s s , O x f o r d , 1965.
1 8 4 - J.H. Mc C A B E - A c o n t i n u e d f r a c t i o n e x p a n s i o n , w i t h a t r u n c a t i o n e s t i m a t e , f o r D a w s o n ' s i n t e g r a l - M a t h . C o m p . , 28 (1974) 8 1 1 - 8 1 6 . 185 - E.P. M E R K E S S I A M J. N u m e r .
80
errors for 486-496.
continued
fraction
186 - E.P. M E R K E S , W.T. S C O T T - C o n t i n u e d f r a c t i o n s o l u t i o n s e q u a t i o n - J. Math. Anal. A p p l . , 4 (1962) 3 0 9 - 3 2 7 .
error
computations
of the Riccati
187 - R. DE M O N T E S S U S DE B A L L O R E - Sur les f r a c t i o n s Bull. Soc. Math. de F r a n c e , 30 (1902) 28-36.
continues
alg~briques
-
188 - R. DE M O N T E S S U S DE B A L L O R E - Sur les f r a c t i o n s C.R. Acad. Sc. P a r i s , 134 (1902) 1 4 8 9 - 1 4 9 1 .
continues
alg~briques
-
189 - R. DE M O N T E S S U S DE B A L L O R E - Sur les f r a c t i o n s c o n t i n u e s L a g u e r r e - C.R. Acad. Sc. P a r i s , 140 (1905) 1 4 3 8 - 1 4 4 0 .
alg~briques
de
190 - R. DE M O N T E S S U S DE B A L L O R E - Les f r a c t i o n s A c t a M a t h . , 32 (1909) 2 5 7 - 2 8 1 . 191 - T. M U I R - A t h e o r e m in c o n t i n u a n t s w i t h an i m p o r t a n t a p p l i c a t i o n - Phil.
continues
-
alg~briques
- E x t e n s i o n of a t h e o r e m H a g . , (5) 3 (1877).
-
in c o n t i n u a n t s
192 - T. M U I R - On the p h e n o m e n o n of g r e a t e s t m i d d l e in the c y c l e of a c l a s s of a c l a s s of p e r i o d i c c o n t i n u e d f r a c t i o n - Prec. Roy. Soc. E d i n b u r g h , 12 (1884) 578-592. 193
- H. P A D E - M 6 m o i r e sur les d 6 v e l o p p e m e n t s en f r a c t i o n s c o n t i n u e s de la f o n c t i e n e x p o n e n t i e l l e p o u v a n t s e r v i r d ' i n t r o d u c t i o n ~ la t h 6 o r i e d e s f r a c t i o n s c o n t i n u e s a l g 6 b r i q u e s - Ann. Fac. Sci. de l'Ec. Norm. Sup., 16 ( 1 8 9 9 ) 395-436.
194 - H. P A D E C.R. Acad.
- Sur la d i s t r i b u t i o n des r 6 d u i t e s Sc. P a r i s , 130 (1900).
anormales
d'une
fonction
-
256
195 - H. P A D E - Sur le d 6 v e l o p p e m e n t en f r a c t i o n c o n t i n u e de la f o n c t i o n F ( h , l , h ' , u ) et la g 6 n 6 r a l i s a t i o n de la t h ~ o r i e des f o n c t i o n s s p h 6 r i q u e s C.R. Acad. Sc. P a r i s , 141 ( 1 9 0 5 ) 8 1 9 - 8 2 1 . 196 - O. P E R R O N - 0 h e r z w e i K e t t e n b r [ c h e n Math. - Nat. KI. S. - B., (1957) 1-13. 197 - O. P E R R O N 1950.
- Die
Lehre yon dem
198 - O. P E R R O N - U b e r P a l e r m o , 29 (1910). 199 - S. P I N C H E R L E Norm. Sup., (3)
eine
Kettenbr[chen
spezielle
Klasse
- Sur les f r a c t i o n s 6 (18-89) 1 4 5 - 1 5 2 .
200 - A. P R I N G S H E I M - U b e r G l i e d e r n - Sb. M ~ n c h e n ,
v o n H.S.
von
continues
Wall
- Bayer,
- Chelsea
Pub.
Kettenbr~chen
alg~briques
ein K o n v e r g e n z k r i t e r i u m 29 ( 1 8 9 9 ) .
fir
Akad.
Co.,
-
Wiss.
New-York,
- Rend. Circ. Mat.
- Ann.
Sci.
Kettenbr[che
Ec.
mit positiven
201 - E. R O U C H E - M 6 m o i r e sur le d 6 v e l o p p e m e n t d e s f o n c t i o n s en s ~ r i e s O r d o n n ~ e s s u i v a n t les d ~ n o m i n a t e u r s des r ~ d u i t e s d ' u n e f r a c t i o n c o n t i n u e - J. Ee. P o l y t e c h n i q u e , 37 (1858). 202 - H. R U T I S H A U S E R - 0 b e r e i n e V e r a l l g e m e i n e r u n g Math. M e c h . , 38 (1958) 2 7 8 - 2 7 9 . 203 - H. R U T I S H A U S E R - B e s o h l e u n i g u n g d e r K e t t e n b r [ c h e n - ZAMM, 38 (1958) 187.
der
Konvergenz
Kettenbr[che
einer gewisse
204 - F.T. S C H U B E R T - De t r a n s f o r m a t i o n e s e r i e s in f r a c t i o n e m Acad. So. Imp. St P ~ t e r s b o u r g , 7 ( 1 8 1 5 - 1 8 1 6 ) . 205 - W.T. S C O T T - The c o r r e s p o n d i n g M a t h . , (2) 51 (1950) 56-67.
continued
206 - H. S I E B E C K - 0 b e r p e r i o d i s c h e r m a t h . , 33 (1846) 71-77.
Kettenbr6che
207 - V. S I N G H , W . J . T H R O N n u e d f r a c t i o n s - Proc. 208
- J. S H E R M A N nued fraction
fraction
Klasse
von
continuam
- Mem.
of a J - f r a c t i o n
- Ann.
- J. f ~ r die r e i n e
- A f a m i l y of b e s t t w i n c o n v e r g e n c e Amer. M a t h . Soc., 7 (1956) 2 7 7 - 2 8 2 .
- On t h e n u m e r a t o r s - Tran. Amer. Math.
- Z. A n g e w .
regions
u. a n g e w .
for conti-
of the c o n v e r g e n t s of the S t i e l t j e s S o c . , 35 (1933) 64-87.
conti-
209 - T.J. S T I E L T J E S - Sur la r ~ d u c t i o n en f r a c t i o n c o n t i n u e d ' u n e s ~ r i e p r ~ c ~ d a n t s u i v a n t les p u i s s a n c e s d e s c e n d a n t e s de la v a r i a b l e - Ann. Fac. Sci. T o u l o u s e , 3 (1889) 1-17. 210 - T.J. S T I E L T J E S - N o t e appl. m a t h . , 25 (1891).
sur q u e l q u e s
fractions
211 - T.J. S T I E L T J E S - R e c h e r c h e s sur les f r a c t i o n s U n i v . T o u l o u s e , 8 (1894) 1-122. 212 - T.J. S T I E L T J E S - S u r un d ~ v e l o p p e m e n t Sc. P a r i s , 99 (1884) 5 0 8 - 5 0 9 . 213 - T.J.
STIELTJES
- Recherches
continues
continues
en f r a c t i o n s
sur les f r a c t i o n s
- Quart.
- Ann.
continues
continues
J. p u r e
Fac.
- C.R.
- M6moires
and
Sci.
Acad.
pr6sen-
257
t6s p a r divers savants ~ l'acad6mie des Sciences Sciences et M a t h ~ m a t i q u e s , (2) 32 (2) (1892).
de l ' i n s t i t u t
de France,
214 - W.B. SWEEZY, W.J. THRON - E s t i m a t e s of the speed of c o n v e r g e n c e c o n t i n u e d fractions - SIAM J. Numer. Anal., 4 (1967) 254-270. 215 - W.J. THRON - On p a r a b o l i c c o n v e r g e n c e Math. Zeit., 69 (1958) 172-182.
regions
for c o n t i n u e d
of certain
fractions
-
216 - W.J. THRON - Recent approaches to c o n v e r g e n c e t h e o r y of c o n t i n u e d f r a c t i o n s in "Pad~ a p p r o x i m a n t s and their a p p l i c a t i o n s " , P.R. G r a v e s - M o r r i s ed., A c a d e m i c Press, New-York, 1973. 217 - W.J. THRON - A survey of recent Rocky M o u n t a i n s J. Math., 4 (1974)
convergence 273-282.
results
218 - W.J. THRON - Two families of twin c o n v e r g e n c e f r a c t i o n s - Duke Math. J., i0 (1943) 677-685. 219 - W.J. THRON - Twin convergence Math., 66 (1944) 428-439. 220 - W.J. THRON - A family Duke Math. J., ii (1944)
regions
of simple 779-791.
for c o n t i n u e d
regions
for c o n t i n u e d
convergence
221 - W.J. THRON - C o n v e r g e n c e regions for the g e n e r a l Amer. Math. Soc., 49 (1943) 913-916. 222 - W.J. THRON - Some p r o p e r t i e s of the c o n t i n u e d Bull. Amer. Math. Soc., 54 (1948) 206-218.
for c o n t i n u e d
fractions
regions
fractions-
- Amer.
for c o n t i n u e d
continued
fraction
J.
fractions-
fraction
- Bull.
(l+doz)+K(z/(l+dnZ))-
223 - W.J. THRON - Z w i l l i n ~ s k o n v e r g e n z g e b l e t e f~r K e t t e n b r ~ c h e l+K(a /i), deren eines die K r e i s s c h e i b e |a2n_l 1 < 02 ist. - Math. Zeit., 70 (1959) n 310-344. 224 - W.J. THRON - Convergence regions for c o n t i n u e d f r a c t i o n s processes - Amer. Math. Monthly, 68 (1961) 734-750.
and other infinite
225 - W.J. THRON - Convergence of sequences of linear f r a c t i o n a l t r a n s f o r m a t i o n s and of c o n t i n u e d fractions - J. Indian Math. Soc., 27 (1963) 103-127. 226 - W.J. THRON - Some results and p r o b l e m s in the a n a l y t i c fractions - Math. Student, 32 (1964) 61-73. 227 - W.J. THRON - On the c o n v e r g e n c e fractions - Math. Zeit., 85 (1964)
t h e o r y of c o n t i n u e d
of the even part of c e r t a i n 268-273.
228 - J. T R E M B L E Y - Recherches sur les f r a c t i o n s Belles-Let., Berlin (1794-1795)
continues
229 - K.T. V A H L E N - 0her N [ h e r u n g s w e r t e angew, math,, 115 (1895) 221-233.
und K e t t e n b r [ c h e
230 - E.B. VAN V L E C K - On an e x t e n s i o n Amer. Math. Soc., 4 (1908) 297-332.
Of the 1894 m e m o i r
- Mem.
V A N V L E C K - On the c o n v e r g e n c e
and c h a r a c t e r
Acad.
Roy.
Sc.
- J. fur die reine u.
of Stieltjes
231 - E.B. VAN V L E C K - On the c o n v e r g e n c e of c o n t i n u e d f r a c t i o n s elements - Trans. Amer. Math. Soc., 2 (1901) 215-233. 232 - E.B.
continued
- Trans.
with c o m p l e x
of the c o n t i n u e d
fraction
258
alZ/l+... 233
- Trans.
- E.B. V A N V L E C K coefficients have
Amer.
Math.
Soc.,
476-483.
- On t h e c o n v e r g e n c e of a l g e b r a i c c o n t i n u e d f r a c t i o n s w h o s e l i m i t i n g v a l u e s - Trans. Amer. Math. Soc., 5 (1904) 253-262.
234 - E.B. V A N V L E C K - S e l e c t e d t o p i c s t i n u e d f r a c t i o n s - Amer. Math. Soc. 1903. 235
2 (1901)
in t h e t h e o r y of d i v e r g e n t C o l l o q u i u m Pub., i, B o s t o n
- E.B. V A N V L E C K - On t h e c o n v e r g e n c e of t h e c o n t i n u e d o t h e r c o n t i n u e d f r a c t i o n s - Ann. M a t h . , 3 (1901) 1-18.
s e r i e s and conColloquium,
fraction
236 - H. V O N K O C H - Q w e l q u e s t h & o r ~ m e s c o n c e r n a n t la t h ~ o r i e t i o n s c o n t i n u e s - O f v e r s i g t af. Kongl. V e t e n s k a p s - Akad.
of G a u s s
and
g ~ n ~ r a l e des f r a c F6rhandlingen, 52
(1895). 237 - B. V I S C O V A T O F F - De la m ~ t h o d e g ~ n ~ r a l e p o u r r ~ d u i r e t o u t e s s o r t e s de q u a n t i t ~ s en f r a c t i o n s c o n t i n u e s - Mem. Acad. Sc. Imp. St P & t e r s b o u r g , 1 (1803-1804). 238 - H. W A A D E L A N D - T - f r a c t i o n s J. M a t h . , 4 (1974) 3 9 1 - 3 9 4 .
from
a different
point
of v i e w
- Rocky
Mountains
239 - H. W A A D E L A N D - On T - f r a c t i o n s of f u n c t i o n s h o l o m o r p h i c a n d b o u n d e d c i r c u l a r d i s c - N o r s k e Vid. Selsk. S k r ( T r o n d h e i m ) 1964, n°8. 240 - H. W A A D E L A N D - A c o n v e r g e n c e p r o p e r t y N o r s k e Vid. Selsk. Skr ( T r o n d h e i m ) 1966,
of certain n°9.
T-fraction
241 - H.S, W A L L - On some c r i t e r i a of C a r l e m a n for t h e c o m p l e t e J - f r a c t i o n - Bull. Amer. Math. Soc., 54 (1948) 528-532. 242 - H.S. W A L L - C o n v e r g e n c e of c o n t i n u e d f r a c t i o n s Bull. Amer. Math. Soc., 55 (1949) 3 9 1 - 3 9 4 . 243
- H.S. W A L L - N o t e 56 (1949) 96-97.
244 - H.S. W A L L of S t i e l t j e s 73-84.
on a p e r i o d i c
expansions
convergence
in p a r a b o l i c
fraction
- Amer.
domains
Math.
-
of a
-
Monthly,
- Concerning continuous continued fractions and certain systems i n t e g r a l e q u a t i o n s - Rend. Circ. Mat. di P a l e r m o , 11,2 (1953)
245 - H.S. W A L L - P a r t i a l l y 7 (1956) 1 0 9 0 - 1 0 9 3 .
bounded
246 - H.S. W A L L - S o m e c o n v e r g e n c e M o n t h l y , 54 (1957) 95-103. 247 - H.S. W A L L - N o t e 51 (1945) 930-934. 248 - H.S. W A L L t i o n - Bull.
continued
in a
on a c e r t a i n
continued
problems
continued
fractions
for
- Proc.
continued
fraction
fractions
- Bull.
- N o t e on the e x p a n s i o n of a p o w e r s e r i e s Amer. Math. Soc., 51 (1945) 97-105.
on a r b i t r a r y
J-fractions
Soc.,
- Amer.
Math.
Soc.,
into a continued
frac-
- Bull,
Amer.
Math.
Math.
249 - H.S. W A L L - C o n t i n u e d f r a c t i o n e x p a n s i o n s f o r f u n c t i o n s p a r t s - Bull. Amer. Math. Soc., 52 (1946) 138-143. 250 - H.S. W A L L - T h e o r e m s (1946) 671-679.
Amer.
with positive
Amer.
Math.
real
Soc.,
52
259
251 - H.S.
WALL - Bounded
J-fractions
- Bull.
Amer.
Math.
252 - H.S. W A L L - S o m e r e c e n t d e v e l o p m e n t s in the t h e o r y Bull. Amer. Math. Soc., 47 (1941) 4 0 5 - 4 2 3 . 253
Soc.,
52
of continued
- H.S. W A L L - A c o n t i n u e d f r a c t i o n r e l a t e d to s o m e p a r t i t i o n E u l e r - Amer. Math. M o n t h l y , 48 (1941) 1 0 2 - 1 0 8 .
254 - H.S. W A L L - The b e h a v i o u r of c e r t a i n s i n g u l a r l i n e - Bull. Amer. Math. S o c . , 255 - H.S. W A L L - C o n t i n u e d f r a c t i o n s Math. Soc., 50 (1944) 1 1 0 - 1 1 9 . 256 - H.S. W A L L - C o n v e r g e n c e Soc., 17 (1931) 5 7 5 - 5 7 9 .
criteria
257 - H.S. W A L L - On the c o n t i n u e d t i o n s - Bull. Amer. Math. Soc., 258 - H.S. W A L L - C o n t i n u e d f o r m a t i o n s - Bull. Amer.
Stieltjes 48 (1942)
and bounded
for
fractions 39 (1933)
fractions and cross-ratios groups Math. Soc., 40 (1934) 5 7 8 - 5 9 2 . of the form
260 - H.S. W A L L - On c o n t i n u e d f r a c t i o n s Math. Soc., 44 (1938) 94-99.
representing
261 - H.S. W A L L - C o n t i n u e d f r a c t i o n s a n d t o t a l l y Amer. Math. Soc., 48 (1940) 1 6 5 - 1 8 4 . - The a n a l y t i c
268 - H.S. W A L L , J.J. D E N N I S f r a c t i o n - D u k e Math. J., 264 - H.S. W A L L , H.L. f r a c t i o n s - Trans.
theory
- The l i m i t c i r c l e 12 (1945) 2 5 5 - 2 7 3 .
Amer.
meromorphic
of C r e m o n a
Math.
func-
trans-
Amer.
functions
- Bull.
Amer.
sequences
- Trans.
- Van Nostrand,
for a positive
G A R A B E D I A N - H a u s d o r f f m e t h o d s of s u m m a t i o n Amer. Math. Soc., 48 ( 1 9 4 0 ) 1 8 5 - 2 0 7 .
the
Amer.
- Bull.
fractions
case
near
- Bull.
- Bull.
-
of
l+Kl(bZ/l)
monotone
of continued
fractions
fractions
686-693.
fractions
formulas
functions
which represent 942-952.
259 - H.S. W A L L - On c o n t i n u e d f r a c t i o n s Math. Soc., 41 ( 1 9 3 5 ) 7 9 7 - 7 3 6 .
262 - H.S. W A L L Y o r k , 1948.
continued 427-431.
analytic
continued
(1946)
New-
definite
J-
and continued
265 - H.S. W A L L , E, H E L L I N G E R - C o n t r i b u t i o n s to the a n a l y t i c t h e o r y f r a c t i o n s a n d i n f i n i t e m a t r i c e s - Ann. M a t h . , 44 ( 1 9 4 3 ) 1 0 3 - 1 2 7 .
of c o n t i n u e d
266 - H.S. W A L L , W. L E I G H T O N - On the t r a n s f o r m a t i o n f r a c t i o n s - Amer. J. M a t h . , 58 (1936) 2 6 7 - 2 8 1 .
of c o n t i n u e d
267 - H.S. W A L L , J.F. P A Y D O N - The c o n t i n u e d t r a n s f o r m a t i o n s - D u k e Math. J., 9 (1942) 268 - H.S. W A L L , and odd parts
fraction 360-372.
R.E. L A N E - C o n t i n u e d f r a c t i o n s w i t h - Trans. Amer. Math. S o c . , 67 (1949)
and
convergence
as a s e q u e n c e
absolutely 368-380.
269 - H.S. W A L L , W.T. S C O T T - On the c o n v e r g e n c e a n d d i v e r g e n c e f r a c t i o n s - Amer. J. M a t h . , 69 (1947) 5 5 1 - 5 6 1 . 270 - H,S. W A L L , s e r i e s - Ann.
W.T. S C O T T - C o n t i n u e d f r a c t i o n M a t h . , 41 (1940) 3 2 5 - 3 4 9 .
expansion
of linear
convergent
even
of c o n t i n u e d
for arbitrary
power
260
271 - H.S. W A L L , W.T. S C O T T - V a l u e r e g i o n s Math. Soc., 47 (1941) 5 8 0 - 5 8 5 . 272 - H.S. 1-18.
WALL,
for c o n t i n u e d
W.T.
SCOTT
- Continued
fractions
273 - H.S. W A L L , W.T. Trans. Amer. Math.
SCOTT Soc.,
- A convergence theorem 47 (1940) 1 5 5 - 1 7 2 .
274 - H.S. W A L L , M. W E T Z E L - Q u a d r a t i c f o r m s n u e d f r a c t i o n s - D u k e M a t h . J.~ ii ( 1 9 4 4 )
- Natl.
fractions
Math.
Mag.,
for continued
and convergence 89-102.
275 - H.S. W A L L , M. W E T Z E L - C o n t r i b u t i o n s to the a n a l y t i c T r a n s . Amer. Math. Soc., 55 (1944) 3 7 3 - 3 9 7 .
- Bull.
13
Amer.
(1939)
fractions
regions
theory
for
-
conti-
of J - f r a c t i o n s
276 - R. W I L S O N - D i v e r g e n t c o n t i n u e d f r a c t i o n s a n d p o l a r s i n g u l a r i t i e s - Proc. Lond. Math. Soc., 26 (1927) 1 5 9 - 1 6 8 / 27 (1928) 4 9 7 - 5 1 2 / 28 (1928) 1 2 8 - 1 4 4 . 277 - R. W I L S O N - D i v e r g e n t c o n t i n u e d f r a c t i o n s Proc. Lond. Math. Soc., 30 (1928) 38-57.
and nonpolar
singularities
-
278 - A. W I N T N E R - E i n q u a l i t a t i v e s K r i t e r i u m f i r dis K o n v e r g e n z K e t t e n h r ~ c h e - M a t h . Z e i t . , 30 (1929) 2 8 5 - 2 8 9 .
des a s s o z i i e r t e n
279 - L. W U Y T A C K - E x t r a p o l a t i o n to t h e p o l a t i o n - R o c k y M o u n t a i n s J. M a t h . ,
fraction
limit by using continued 4 (1974) 3 9 5 - 3 9 7 .
inter-
280 - P. W Y N N - C o n t i n u e d f r a c t i o n s w h o s e c o e f f i c i e n t s o b e y a n o n c o m m u t a t i v e of m u l t i p l i c a t i o n - Arch. Rat. M e c h . A n a l . , 12 ( 1 9 6 3 ) 2 7 3 - 3 1 2 . 281 - P. W Y N N - A n o t e on t h e c o n v e r g e n c e of c e r t a i n n o n c o m m u t a t i v e f r a c t i o n s - M R C t e c h n i c a l s u m m a r y r e p o r t 750, M a d i s o n , 1967. 282 - P. W Y N N - V e c t o r
continued
fractions
- Linear
Algehra,
283 - P. W Y N N - U p o n t h e d e f i n i t i o n o f an i n t e g r a l as t h e f r a c t i o n - Arch. Rat. M e c h . A n a l . , 28 (1968) 8 3 - 1 4 8 .
continued
i (1968)
limit
law
357-395.
of a c o n t i n u e d
284 - P. W Y N N - An a r s e n a l of A l g o l p r o c e d u r e s for the e v a l u a t i o n of c o n t i n u e d f r a c t i o n s a n d f o r e f f e c t i n g the e p s i l o n a l g o r i t h m - C h i f f r e s , 9 (1966) 327-362. 285 - P. W Y N N - F o u r l e c t u r e s tions - CIME summer school
on t h e n u m e r i c a l l e c t u r e s , 1965.
286 - P. W Y N N - C o n v e r g i n g f a c t o r s i (1959) 2 7 2 - 3 0 7 a n d 3 0 8 - 3 2 0 .
for
application
continued
fractions
of continued
- Numer.
frac-
Math.,
287 - P. W Y N N - A n o t e on a m e t h o d of B r a d s h a w f o r t r a n s f o r m i n g s l o w l y c o n v e r g e n t s e r i e s a n d c o n t i n u e d f r a c t i o n s - Amer. M a t h . M o n t h l y , 69 (1962) 8 8 3 - 8 8 9 . 288 - P. W Y N N - A n u m e r i c a l 175-196.
s t u d y of a r e s u l t
of S t i e l t j e s
- Chiffres,
6 (1963)
289 - P. W Y N N - On s o m e r e c e n t d e v e l o p m e n t s in the t h e o r y a n d a p p l i c a t i o n c o n t i n u e d f r a c t i o n s - S I A M J. N u m e r . A n a l . , i (1964) 1 7 7 - 1 9 7 , 290 - P. W Y N N - N o t e on a c o n v e r g i n g N u m e r , M a t h . , 5 (1963] 3 3 2 - 3 5 2 .
factor
for a certain
continued
of
fraction
-
-
261
291 - P. W Y N N - The n u m e r i c a l e f f i c i e n c y o f c e r t a i n c o n t i n u e d K o n i n k l . N e d e r l . Akad. Wet., 65 A (1962) 1 2 7 - 1 4 8 . 292 - P. W Y N N - C o m p l e x n u m b e r s on a p p l i c a t i o n to the t h e o r y r e p o r t 646, M a d i s o n , 1966.
fraction
expansions
a n d o t h e r e x t e n s i o n s to the C l i f f o r d a l g e b r a w i t h of c o n t i n u e d f r a c t i o n s , M R C t e c h n i c a l s u m m a r y
llI-
Applications
293 - A.C. A I T K E N - On B e r n o u l l i ' s n u m e r i c a l s o l u t i o n Proc. Roy. Soc. E d i n b u r g , 46 (1926) 2 8 9 - 3 0 5 .
of a l g e b r a i c
294 - A.C. A I T K E N - On i n t e r p o l a t i o n b y p r o p o r t i o n a l parts, d i f f e r e n c e s - Proc. Edin. Math. Soc., 3 (1932) 56-76. 295 - R. A L T - M 6 t h o d e s A - s t a b l e s p o u r l ' i n t 6 g r a t i o n mal conditionn6s - T h ~ s e 3~me c y c l e , P a r i s , 1971.
des
296 - F.L. B A U E R - C o n n e c t i o n s b e t w e e n t h e q - d a l g o r i t h m e - a l g o r i t h m of W y n n - D e u t s c h e F o r s c h u n g s g e m e i n s c h a f t
equations
without
syst~mes
use
of
diff6rentiels
of R u t i s h a u s e r a n d the Teeh. Rep. B a / 1 0 6 , 1957.
297 - F.L. B A U E R - N o n l i n e a r s e q u e n c e t r a n s f o r m a t i o n s - in " A p p r o x i m a t i o n f u n c t i o n s " , G a r a b e d i a n ed., E l s e v i e r , N e w - Y o r k , 1965. 298 - F.L.
BAUER
- The g - a l g o r i t h m
- SIAM
J.,
8 (1960)
-
of
1-17.
300 - M. B A U S S E T - U n e d 6 t e r m i n a t i o n des s u r f a c e s de c h o c p a r a e c 6 1 6 r a t i o n c o n v e r g e n c e - C o l l o q u e E u r o m e e h 58, T o u l o n , 1 2 - 1 4 m a i 1975.
de
301 - L.C. B R E A U X - A n u m e r i c a l s t u d y o f t h e a p p l i c a t i o n of a c c e l e r a t i o n techniq u e s a n d p r e d i c t i o n a l g o r i t h m s to n u m e r i c a l i n t e g r a t i o n - M. Sc. T h e s i s , Louisiana State Univ., New-Orleans, 1971. 302 - C. B R E Z I N S K I - C o n v e r g e n c e d ' u n e f o r m e e o n f l u e n t e C.R. Acad. Sc. Paris, 273 A (1971) 5 8 2 - 5 8 5 .
de l ' E - a l g o r i t h m e
303 - C. B R E Z I N S K I - L ' E - a l g o r i t h m e et les s u i t e s t o t a l e m e n t o s c i l l a n t e s - C.R. Acad. Sc. P a r i s , 276 A (1973) 3 0 5 - 3 0 8 . 304 - C. rique
BREZINSKI - M6thodes d'acc616ration - T h ~ s e , Univ. de G r e n o b l e , 1971.
de la c o n v e r g e n c e
305 - C. B R E Z I N S K I - A p p l i c a t i o n du o - a l g o r i t h m e C.R. Aead. Sc. P a r i s , 270 A (1970) 1 2 5 2 - 1 2 5 3 . 306 - C. B R E Z I N S K I (1971) 1 5 3 - 1 6 2 .
- Etudes
sur les
307 - C. B R E Z I N S K I RIRO, R3 (1970)
- RSsultats 147-153.
get
309 - C. B R E Z I N S K I - R e v i e w of m e t h o d s Rend. Mat. Roma, 7 (1974) 3 0 3 - 3 1 6 .
- Numer.
de s o m m a t i o n
de l ' s - a l g o r i t h m e
to a c c e l e r a t e
the
et
en a n a l y s e
~ la q u a d r a t u r e
p-algorithmes
s u r les p r o c 6 d 6 s
308 - C. B R E Z I N S K I - F o r m e c o n f l u e n t e M a t h . , 23 (1975) 3 6 3 - 3 7 0 .
monotones
-
num6rique
Math.,
num6-
-
17
et l ' c - a l g o r i t h m e
topologique
convergence
-
- Numer.
of s e q u e n c e s -
-
262
310 - C. BREZINSKI - Acc~l~ration de la c o n v e r g e n c e en analyse num~rique P u b l i c a t i o n 41, labo. de Calcul, Univ. de Lille, 1973. 311 - C. BREZINSKI - Conditions d ' a p p l i c a t i o n et de convergence de p r o c ~ d ~ s d ' e x t r a p o l a t i o n - Numer. Math., 20 (1972) 64-79. 312 - C. BREZINSKI - A p p l i c a t i o n de l'e-algorithme ~ la r ~ s o l u t i o n des syst~mes non lin~aires - C.R. Acad. Sc. Paris, 271 A (1970) 1174-1177. 313 - C. BREZINSKI - Integration des syst~mes d i f f ~ r e n t i e l s ~ l'aide du 0-algorithme - C.R. Acad. Sc. Paris, 278 A (1974) 875-878. 314 - C. BREZINSKI - Sur un algorithme de r ~ s o l u t i o n des syst~mes non lin~aires C.R. Acad. Sc. Paris, 272 A (1971) 145-148. 315 - C. BREZINSKI - Numerical stability of a quadratic m e t h o d for solving systems of non linear equations - Computing, 14 (1975) 205-211. 316 - C. BREZINSKI - C o m p u t a t i o n of the eigenelements of a m a t r i x by the e-algorithm - Linear Algebra, ii (1975) 7-20. 317 - C. BREZINSKI, M. CROUZEIX - Remarques sur le proc~d~ A 2 d'Aitken - C.R. Acad. Sc. Paris, 270 A (1970) 896-898. 318 - C. BREZINSKI, A.C. RIEU - The solution of systems of equations using the e-algorithm, and an a p p l i c a t i o n to b o u n d a r y value p r o b l e m s - Math. Comp., 28 (1974) 731-741. 319 - H. CABANNES, M. BAUSSET - A p p l i c a t i o n of the m e t h o d of Pad~ to the determ i n a t i o n of shock wavres - in "Problems of h y d r o d y n a m i c s and continuum mechanics ", in honor of L.I. Sedov, English ed. publ. by SIAM (1968) 95-114. 320 - H.K. CHENG, M. HAFEZ - Convergence a c c e l e r a t i o n of iterative solutions and transonic flow computations - Colloque Euromech 58, Toulon, 12-14 mai 1975. 321 - J.S.R. CHISHOLM - Pad~ a p p r o x i m a n t s and linear integral equations - in "The Pad~ approximant in t h e o r e t i c a l physics", G.A. Baker Jr. and J.L. Gammel eds., Academic Press, New-York, 1970. 322 - J.S.R. CHISHOLM - A p p l i c a t i o n of Pad~ a p p r o x i m a t i o n to n u m e r i c a l integration - Rocky Mountains J. Math., 4 (1974) 159-168. 323 - J.S.R. CHISHOLM - Pad~ a p p r o x i m a t i o n of single v a r i a b l e integrals Colloquium on c o m p u t a t i o n a l methods in t h e o r e t i c a l physics, Marseille, 1970. 324 - J.S.R. CHISHOLM - A c c e l e r a t e d convergence of sequences of quadratur e approximants - second c o l l o q u i u m on c o m p u t a t i o n a l m e t h o d s in t h e o r e t i c a l physics, Marseille, 1971. 325 - J.S.R. CHISHOLM, A.C. GENZ, G.E. ROWLANDS - A c c e l e r a t e d convergence of sequences of quadrature a p p r o x i m a t i o n - J. Comp. Phys., i0 (1972) 284-307. 326 - J. COUNTS, J.E. AKIN - The a p p l i c a t i o n of continued fractions to wave p r o p a g a t i o n problems in a v i s c o e l a s t i c rod - DEMVPI R e s e a r c h Rep. i-i, Dept. of Eng. Mech., V i r g i n i a p o l y t e c h n i c institute, Blacksburg, 1968. 327 - J.D.P. D O N N E L L Y - Applications of the q-d and e-algorithms - in "Methods of numerical approximation", D.C. Handscomb ed., P e r g a m o n Press, Oxford, 1965.
263
328 - B.L. EHLE - A - s t a b l e m e t h o d s and Pad~ a p p r o x i m a t i o n s SIAM J. Math. Anal., 4 (1973) 671-680.
to the e x p o n e n t i a l
-
329 - B.L. EHLE - On Pad~ a p p r o x i m a t i o n s to the e x p o n e n t i a l f u n c t i o n and A - s t a b l e m e t h o d s for the n u m e r i c a l s o l u t i o n of initial value p r o b l e m s - R e s e a r c h rep. CSRR 2010, dept. of AACS, Univ. of Waterloo, Ontario, 1969. 330 - M. F R O I S S A R T - A p p l i c a t i o n s of the Pad~ m e t h o d to n u m e r i c a l analysis C o l l o q u i u m on c o m p u t a t i o n a l m e t h o d s in t h e o r e t i c a l physics, Marseille, 1970. 331 - E. G E K E L E R - U b e r den e - a l g o r i t h m u s 51 (1971) 53-54.
von W y n n - Z. Angew.
332 - E. G E K E L E R - On the s o l u t i o n of systems of e q u a t i o n s r i t h m of Wynn - Math. Comp., 26 (1972) 427-436.
Math.
Mech.,
by the e p s i l o n
algo-
333 - A. GENZ - The e - a l g o r i t h m and some other a p p l i c a t i o n s of Pad~ a p p r o x i m a n t s in n u m e r i c a l a n a l y s i s - in "Pad~ a p p r o x i m a n t s " , P.R. Graves- Morris ed., The institute of physics, London, 1973. 334 - A. G E N Z - A p p l i c a t i o n s of the e - a l g o r i t h m to q u a d r a t u r e p r o b l e m s - in "Pad~ a p p r o x i m a n t s and their a p p l i c a t i o n s ", P.R. G r a v e s - M o r r i s ed., A c a d e m i c Press, New-York, 1973. 335 - B. G E R M A I N
BONNE - T r a n s f o r m a t i o n s
de suites
- RAIRO,
R1
(1973)
84-90.
336 - M. HAFEZ, H.K. CHENG - On a c c e l e r a t i o n of c o n v e r g e n c e and s h o c k - f i t t i n g t r a n s o n i c f l o w c o m p u t a t i o n s - Univ. So. Calif. Memo., 1973. 337 - M, HAFEZ, H.K. CHENG - C o n v e r g e n c e a c c e l e r a t i o n and shock f i t t i n g t r a n s o n i c a e r o d y n a m i c s c o m p u t a t i o n s - AIAA paper, No. 75-51. 338 - P. HENRICI - Some a p p l i c a t i o n s of the q u o t i e n t - d i f f e r e n c e Symp. Appl. Math., 20 (1963) 159-183. algorithm
- NBS appl.
340 - D . C . ' J O Y C E - Survey of e x t r a p o l a t i o n Rev., 13 (1971) 435-490.
processes
in n u m e r i c a l
341 - D.K. K A H A N E R - N u m e r i c a l (1972) 689-694.
quadrature
by the e - a l g o r i t h m
for
algorithm
339 - P. HENRICI - The q u o t i e n t - d i f f e r e n c e 49 (1958) 23-46.
Math.
- Proc.
series,
analysis
- Math.
in
Comp.,
342 - D. LEVIN - Development of n o n - l i n e a r t r a n s f o r m a t i o n s for i m p r o v i n g gence of sequences - Intern. J. Comp. Math., B3 (1973) 371-388.
- SIAM
26
conver-
343 - I.M. LONGMAN - Use of Pad~ table for a p p r o x i m a t e Laplace t r a n s f o r m inversion - in "Pad~ a p p r o x i m a n t s and their a p p l i c a t i o n s " , P.R. G r a v e s - M o r r i s ed., A c a d e m i c Press, New-York, 1973. 344 - E.D. MARTIN, H. LOMAX - Rapid finite d i f f e r e n c e c o m p u t a t i o n and t r a n s o n i c a e r o d y n a m i c flows - AIAA paper, No. 74-11. 345 - I. MARX - Remark c o n c e r n i n g a n o n l i n e a r tion - J. Math. Phys., 42 (1963) 834-335. 346 - K.J.
OVERHOLT
- Extended
sequence
Aitken acceleration
to sequence
- BIT,
5 (1965)
of subsonic
transforma-
122-132.
264
347
- R. P E N N A C C H I - S o m m a di s e r i e n u m e r i c h e t i c a T2, 2 - C a l c o l o , 5 (1968) 51-61.
348
- R. P E N N A C C H I 5 (1968) 37-50.
- La t r a s f o r m a z i o n i
mediante
razionali
la t r a s f o r m a z i o n e
di u n a
349 - A. R O N V E A U X - P a d 6 a p p r o x i m a n t a n d h o m o g r a p h i c p h a s e e q u a t i o n s - in " P a d 6 a p p r o x i m a n t s and thein M o r r i s ed., A c a d e m i c P r e s s , N e w - Y o r k , 1973.
successione
- Calcolo,
transformation applications",
350 - H. R U T I S H A U S E R - On a m o d i f i c a t i o n of the q - d a l g o r i t h m c o n v e r g e n c e - ZAMP, 13 (1962) 4 9 3 - 4 9 6 .
with
quadra-
of Riccati's P.R. G r a v e s -
Graeffe
type
351 - H. R U T I S H A U S E R - B e s t i m m u n g d e r E i g e n w e r t e u n d E i g e n v e k t o r e n einer Matrix m i t H i l f e des Q u o t i e n t e n - D i f f e r e n z e n A l g o r i t h m u s - ZAMP, 6 (1955) 387-401. 352
- H. R U T I S H A U S E R - E i n e F o r m e l y o n W r o n s k i u n d i h r e B e d e n t u n g t i e n t e n - D i f f e r e n z e n A l g o r i t h m u s - ZAMP, 7 (1956) 164-169.
353 - H. R U T I S H A U S E R - S t a b i l e S o n d e r f [ l l e des Q u e t i e n t e n r i t h m u s - Numer. M a t h . , 5 (1963) 95-112. 354 - H. R U T I S H A U S E R 233-251.
- Der Quotienten
355 - H. R U T I S H A U S E R - A n w e n d u n g e n ZAMP, 5 (1954) 496-508. 356 - J.R. S C H M I D T b y an i t e r a t i v e 357
des
Quotienten
- Differenzen
Algorithmus
SCHWARTZ - Series - Phys. Fluids.
solution
structure
- Differenzen
asymmetric
361 - R.E. S H A F E R - On q u a d r a t i c (1974) 447e460.
approximation
-
equations
conical
blunt-body
359 - L.W. S C H W A R T Z - S o l u t i o n s to t h e a s y m m e t r i c b l u n t - b o d y p r o b l e m approximants - C o l l o q u e E u r o m e c h 58, T o u l o n , 1 2 - 1 4 m a i 1975. 360 - L.W. S C H W A R T Z - H y p e r s o n i c f l o w s g e n e r a t e d b y p a r a b o l i c s h o c k w a v e s - Phys. F l u i d s , 17 (1974) 1 8 1 6 - 1 8 2 1 .
5 (1954)
Algorithmus
of t h e T a y l o r - M a c c a l l
for the planar
Algo-
- ZAMP,
- On the n u m e r i c a l s o l u t i o n of l i n e a r s i m u l t a n e o u s m e t h o d - Phil. M a g . , 7 (1951) 3 6 9 - 3 8 3 .
- L.W. S C H W A R T Z - On t h e a n a l y t i c f l o w s o l u t i o n - ZAMP.
358 - L.W. problem
- Differenzen
fir den Quo-
using
Pad6
and paraboloidal
- S I A M J. Numer.
Anal.,
ii
362 - D. S H A N K S - An a n a l o g y b e t w e e n t r a n s i e n t s a n d m a t h e m a t i c a l sequences and s o m e n o n l i n e a r s e q u e n c e to s e q u e n c e t r a n s f o r m s s u g g e s t e d by it - N a v a l O r d n a n c e Lab. Mem. 9994, W h i t e Oak, M d . , 1949. 363 - D. S H A N K S - N o n l i n e a r t r a n s f o r m a t i o n s s e r i e s - J. Math. P h y s . , 34 (1955) 1-42. 364 - M.D. V A N D Y K E N e w - Y o r k , 1964.
- Perturbation
365 - M.D. V A N D y K E - A n a l y s i s Mech. Appl. Math.
and
methods
of d i v e r g e n t
in f l u i d
improvement
and slowly
mechanics
of p e r b u r t a t i o n
convergent
- Academic
series
Press,
- Quart.
J.
265
366 - A.H. V A N T U Y L - C a l c u l a t i o n ii (1973) 537-541.
of n o z z l e
using
Pad~
fractions
- AIAA
J.,
367 - A.H. V A N T U Y L - A p p l i c a t i o n of m e t h o d s f o r a c c e l e r a t i o n of c o n v e r g e n c e to the c a l c u l a t i o n of s i n g u l a r i t i e s of t r a n s o n i c f l o w s - C o l l o q u e E u r o m e c h 58, T o u l o n , 1 2 - 1 4 m a i 1975. 368
- A.H. V A N T U Y L flows - Colloque
- The u s e of P a d ~ f r a c t i o n s in the c a l c u l a t i o n E u r o m e c h 58, T o u l o n , 1 2 - 1 4 m a i 1975.
369 - H.S. WALL, W.T. S C O T T - The t r a n s f o r m a t i o n Amer. Math. Soc., 51 (1942) 255-279.
of s e r i e s
and
of n o z z l e
sequences
- Trans.
370 - P.J.S. W A T S O N - A l g o r i t h m s f o r d i f f e r e n t i a t i o n a n d i n t e g r a t i o n - in " P a d s approx~mants and their applications", P.R. G r a v e s - M o r r i s ed., A c a d e m i c P r e s s , N e w - Y o r k , 1973. 371 - L. W U Y T A C K - The u s e of P a d ~ a p p r o x i m a t i o n C o l l o q u e E u r o m e c h 58, T o u l o n , 1 2 - 1 4 m a l 1975.
in n u m e r i c a l
integration
372 - P. W Y N N - On a c o n n e c t i o n b e t w e e n t h e f i r s t a n d the s e c o n d of the e - a l g o r i t h m - Nieuw. Arch. W i s k . , ii (1963) 19-21. 373 - P. W Y N N - U p o n a s e c o n d c o n f l u e n t Math. Soc., 5 (1962) 160-165.
form
of the
e-algorithm
-
confluent
- Proc.
form
Glasgow
374 - P. W Y N N - A c o m p a r i s o n b e t w e e n t h e n u m e r i c a l p e r f o r m a n c e s of the E u l e r transformation a n d the E - a l g o r i t h m - C h i f f r e s , i (1961) 23-29. 375 - P. W Y N N 19-22.
- On r e p e a t e d
application
o f the
e-algorithm
376 - P. w Y N N - The e p s i l o n a l g o r i t h m a n d o p e r a t i o n a l a n a l y s i s - Math. C o m p . , 15 (1961) 1 5 1 - 1 5 8 . 377
- P. W Y N N - A c c e l e r a t i o n t e c h n i q u e s in n u m e r i c a l r e f e r e n c ~ to p r o b l e m s in one i n d e p e n d a n t v a r i a b l e N o r t h H o l l a n d , (1962) 149-156.
378 - P. W Y N N - A n o t e on p r o g r a m m i n g C h i f f r e s , 8 (1965) 23-62. 379 - P. W Y N N - The r a t i o n a l defined by a power series 380 - P. W Y N N recherches
repeated
formulas
of n u m e r i c a l
analysis with particular - Proc. I F I P C o n g r e s s ,
application
of the
- The a b s t r a c t t h e o r y of t h e e p s i l o n a l g o r i t h m math~matiques n ° 74, Univ. de M o n t r e a l , 1971.
e-algorithm
arrays
- Centre
- Louisiana
State
- P. W Y N N - I n v a r i a n t s a s s o c i a t e d w i t h t h e e p s i l o n a l g o r i t h m a n d c o n f l u e n t f o r m - Rend. Circ. Mat. P a l e r m o , (2) 21 (1972) 31-41.
383 - P. W Y N N 175-195.
- Singular
384 - P. W Y N N
- A sufficient
4 (1961)
-
approximation of f u n c t i o n s w h i c h a r e f o r m a l l y e x p a n s i o n - Math. C o m p . , 14 (1960) 1 4 7 - 1 8 6 .
381 - P. W Y N N - U p o n a h i e r a r c h y of e p s i l o n N e w - O r l e a n s , techn, rep. 46, 1970. 382
- Chiffres,
rules
for certain
condition
nonlinear
for the
algorithms
instability
de
Univ.,
its f i r s t
- BIT,
of the
3 (1963)
e-algorithm
-
266
Nieuw.
Arch.
Wisk.,
3 (1961)
117-119.
385 - P. WYNN - On the p r o p a g a t i o n Numer. Math., i (1959) 142-149,
of error in certain n o n l i n e a r
386 - P. WYNN - On the c o n v e r g e n c e and s t a b i l i t y SIAM J. Numer. Anal., 3 (1966) 91-122.
of the e p s i l o n
algorithms
algorithm
-
-
387 - P. W Y N N - H i e r a r c h i e s of arrays and f u n c t i o n sequences a s s o c i a t e d w i t h the epsilon a l g o r i t h m and its first c o n f l u e n t f o r m - Rend. Mat. Roma, 5 (1972) 819-852. 388 - P. WYNN - A c c 6 1 6 r a t i o n de la c o n v e r g e n c e de s6ries d ' o p 6 r a t e u r s n u m 6 r i q u e - C.R. Acad. Se. Paris, 276 A (1973) 803-806. 389 - P. W Y N N - T r a n s f o r m a t i o n s de s6ries Sc. Paris, 275 A (1972) 1351-1353. 390 - P. WYNN - U p o n some c o n t i n u o u s 197-234 and 235-278. 391 - P. WYNN - Confluent ii (1960) 223-234.
forms
~ l'aide de l ' s - a l g o r i t h m e
prediction
of certain
392 - P. W Y N N - A note on a confluent (1960) 237-240.
en analyse
algorithms
nonlinear
- C.R. Aead.
- Calcolo,
algorithms
f o r m of the s - a l g o r i t h m
9 (1972)
- Arch.
- Arch.
Math.,
Math.,
ii
393 - P. WYNN - On a p r o c r u s t e a n t e c h n i q u e for the n u m e r i c a l t r a n s f o r m a t i o n of slowly convergent sequences and series - Proc. Camb. Phil. Soc., 52 (1956) 663-671. 394 - P. W Y N N - T r a n s f o r m a t i o n s to accelerate the c o n v e r g e n c e of F o u r i e r series G e r t r u d e Blanch a n n i v e r s a r y vol., Wright P a t t e r s o n Air Force Base, (1967) 339-379. 395 - P. WYNN - P a r t i a l d i f f e r e n t i a l equations algorithms - ZAMP, 15 (1964) 273-289.
associated
with certain nonlinear
396 - P. W Y N N - A n u m e r i c a l m e t h o d for e s t i m a t i n g p a r a m e t e r s in m a t h e m a t i c a l m o d e l s - Centre de r e c h e r c h e s m a t h 6 m a t i q u e s , Univ. de Montr6al, rep. CRM 443, 1974. 397 - P. WYNN - On a device i0 (1956) 91-96.
for c o m p u t i n g
the em(S n) t r a n s f o r m a t i o n
398 - P. W Y N N - A c o n v e r g e n c e t h e o r y of some m e t h o d s r e c h e r c h e s m a t h ~ m a t i q u e s , Univ. de Montr6al, rep. 399 - P. W Y N N - A sufficient c o n d i t i o n Numer. Math., i (1959) 203-207.
of i n t e g r a t i o n CRM-193, 1972.
for the i n s t a b i l i t y
400 - P. W Y N N - A c o m p a r i s o n t e c h n i q u e for the n u m e r i c a l slowly c o n v e r g e n t series b a s e d on the use of r a t i o n a l Math., 4 (1962) 8-14.
- MTAC,
- Centre de
of the q-d a l g o r i t h m
t r a n s f o r m a t i o n of f u n c t i o n s - Numer.
-
267
IV - M I S C E L L A N E O U S
401 ~ G.D. A L L E N , C.K. CHUI, W.R. M A D Y C H , Pad~ approximation of S t i e l t j e s s e r i e s
F.J. N A R C O W I C H , P.W. S M I T H - J. Appr. T h e o r y , 14 (1975) 3 0 2 - 3 1 6 .
402 - G.D. A L L E N , C.K. CHUI, W.R. M A D Y C H , F.J. N A R C O W I C H , P.W. S M I T H Pad~ approximant~, N u t t a l l ' s f o r m u l a a n d a m a x i m u m p r i n c i p l e - to a p p e a r . 403 - M.F. B A R N S L E Y - C o r r e c t i o n 16 (1975) 9 1 8 - 9 2 8 4 0 4 - C. B R E Z I N S K I to appear.
- Matrices
terms
for Pad~
semi-d~finies
approximants
positives
et s u i t e s
405 - F. C O R D E L L I E R - U n e n o u v e l l e f o r m e de l ' e - a l g o r i t h m e C o l l o q u e E u r o m e c h 58, T o u l o n , 1 2 - 1 4 m a i 1975. 406 - F. C O R D E L L I E R - R ~ g l e s p a r t i c u l i ~ r e s C o l l o q u e d ' A n a l y s e N u m ~ r i q u e , La G r a n d e
- J.
de m o m e n t s
~tape
vectoriel
407
- F. Publ.
408
~ F. C O R D E L L I E R - L ' e - a l g o r i t h m e vectoriel : interpretation r~gles singuli~res - Colloque d'analyse num~rique, Gourette,
409
- A.C. G E N Z - The a p p r o x i m a t e extrapolation methods - Ph.D.
- S.T. PENt, S I A M J. Math.
412 - J. T O D D
lemniscate
-
g~om~trique m a i 1974.
-
et
c a l c u l a t i o n of m u l t i d i m e n s i o n a l integrals using T h e s i s , Univ. of K e n t at C a n t e r b u r y , 1975.
A. H E S S E L - C o n v e r g e n c e o f n o n c o m m u t a t i v e A n a l . , 6 (1975) 7 2 4 - 7 2 7 .
- The
-
de l ' c - a l g o r i t h m e
410 - P. H I L L I O N - M ~ t h o d e d ' A i t k e n i t ~ r ~ e p o u r les s u i t e s o s c i l l a n t e s x i m a t i o n s - C.R. Acad. Sc. P a r i s , 280 A (1975) 1 7 0 1 - 1 7 0 4 . 411
Phys.,
et ses v a r i a n t e s
pour l'c-algorithme M o t t e , m a i 1975.
CORDELLIER - Interpretation g~om~trique d'une 40, Labo. de C a l c u l , Univ. de L i l l e , 1973.
Math.
constants
- Comm.
ACM,
18
continued
(1975)
d'appro-
fractions
14-19.
UER d'IEEA - Informatique BP 36 59650 - Villeneuve d'Ascq FRANCE
-