Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Symboli
Software
for Lie Symmetry Analysis
W. Hereman1
Department of Mathemati
al and Computer S
ien
es
Colorado S
hool of Mines
Golden, CO 80401, USA
To appear as Chapter 13
CRC Handbook of Lie Group Analysis of Dierential Equations
Volume 3: New Trends in Theoreti
al Developments and Computational Methods
Edited by N.H. Ibragimov
CRC Press, Bo
a Raton, Florida, 1995
November 15, 1999
MCS-94-07
August 1994
Chapter 13
Abstra
t
A survey of te
hniques and symboli
programs for the determination of Lie symmetry groups of
systems of dierential equations is presented. The purpose, methods and algorithms of symmetry
analysis are outlined. An exhaustive review of the literature, in
luding old and modern books and
papers presenting key
on
epts is given. Spe
ial attention is paid to methods for redu
ing the determining equations into standard form, and their subsequent integration. Several examples illustrate
the use of the Lie symmetry software. Throughout the paper, new trends in the development of
symboli
pa
kages for Lie symmetry analysis are indi
ated.
13.1.
The Norwegian mathemati
ian Marius Sophus Lie (1842-1899) and Felix Klein (1849-1925), a
German geometer, pioneered the study of transformation groups that leave systems of equations
invariant. Klein's work fo
used on the role of nite groups in the study of regular bodies and algebrai
equations. Lie founded the theory of
ontinuous transformations groups [179, 180, 181, 182,
183, 184, 185 and Lie groups. Although Lie's starting point had been geometry, the inspirational
sour
e for his group theoreti
investigations was the eld of dierential equations. His goal was to
establish a theory of integration of dierential equations that would mirror Abel's theory for the
solution of algebrai
equations. His work brought many diverse and ad ho
integration methods
for solving spe
ial
lasses of dierential equations under a
ommon
on
eptual umbrella.
Later on, the
on
ept of symmetry evolved into one of the most explosive developments of
mathemati
s and physi
s throughout the twentieth
entury. The theory of Lie groups and Lie
algebras is now applied to diverse elds of mathemati
s in
luding dierential geometry, algebrai
topology, bifur
ation theory, numeri
al analysis, spe
ial fun
tions, to name a few; and to nearly
any area of theoreti
al physi
s, in parti
ular
lassi
al,
ontinuum and quantum me
hani
s,
uid
dynami
s, relativity, and parti
le physi
s.
Lie's innitesimal transformation method provides the most widely appli
able te
hnique to nd
losed form solutions of ordinary dierential equations (ODEs). Standard solution methods for
rst-order or linear ODEs
an be
hara
terized in terms of symmetries. For nonlinear ODEs, Lie's
method, when it su
eeds, provides a means of redu
ing the solution to a series of quadratures and
an be implemented in symboli
programs [41. Under
ertain
onditions, rst- and se
ond-order
ODEs
an be linearized via Lie group transformations [149, 151. Through the group
lassi
ation
of ODEs, Lie also su
eeded in identifying all ODEs that
an either be redu
ed to lower-order ones,
or
ompletely integrated via group theoreti
te
hniques.
Applied to partial dierential equations (PDEs), the method leads to group-invariant solutions,
onservation laws, invariant
enter manifolds in bifur
ation theory [56, 111, 112, 252, 299, et
.
Exploiting the symmetries of PDEs, new solutions
an be derived from old ones, and PDEs
an
be
lassied into equivalen
e
lasses. Spe
ial, physi
ally signi
ant solutions arising from symmetry methods allow one to investigate the asymptoti
or physi
al behavior of general solutions.
The group-invariant solutions obtained via Lie's approa
h provide insight into the physi
al models themselves. Expli
it solutions also serve as ben
hmarks in the design, a
ura
y testing, and
omparison of numeri
al algorithms.
Lie's original ideas greatly in
uen
ed the study of physi
ally important systems of dierential
equations in
lassi
al me
hani
s,
uid dynami
s, elasti
ity, and many other applied areas. Currently, Lie symmetry methods are applied to dieren
e, dierential-dieren
e equations [177, 301
and integro-dierential equations. For a good review of the present state of aairs see [60.
The appli
ation of Lie groups methods to
on
rete physi
al systems involves tedious, me
hani
al
omputations. Even the
al
ulation of the
ontinuous symmetry group of a modest system
of dierential equations is prone to fail, if done with pen
il and paper. Programmable
omputer
algebra systems (CAS) su
h as Mathemati
a, MACSYMA, Maple, REDUCE, AXIOM and MuPAD, are extremely useful aides in su
h
omputations. Symboli
pa
kages, written in the language
of CAS,
an nd the dening (or determining) equations of the Lie symmetry group. The most
sophisti
ated pa
kages then redu
e the determining system into an equivalent but more suitable
system, subsequently solve that system in
losed form, and go on to
al
ulate the innitesimal
generators that span the Lie algebra of symmetries.
A large body of literature exists on the topi
of Lie symmetries. We list some older books
[8, 9, 10, 26, 29, 35, 48, 51, 55 [73, 84, 130, 140, 161, 174, 196, 197, 198, 220, 222, 272, more
modern books [12, 13, 18, 21, 36, 45, 78 [80, 90, 103, 105, 106, 107, 108 [110, 112, 144, 145, 147,
148, 149, 153, 162, 178 [175, 193, 214, 221, 224, 227, 229, 244, 245, 247 [248, 252, 253, 275, 281, 285,
305, 310, 316, 326, some re
ent
onferen
e pro
eedings [7, 43, 102, 155, 166, 301, published le
ture
notes [230, re
ent Ph.D. theses [54, 120, 141, 146, 164, 219, 276, 278, 311, and papers presenting
key
on
epts [6, 11, 14, 15, 28, 109 [150, 171, 251, 264, 290, 291, 292, 303, 304, 306, 312, 313, 315.
Also of interest are the le
ture notes and preprints published by The International Sophus Lie
Center in Oslo (for example, [169), and the notes of the
ourses organized by ERCIM on \Partial
Dierential Equations and Group Theory" [230, 270.
Extensive tables of symmetry group generators for well-known equations from mathemati
al
physi
s have been gathered by Rogers and Ames [244 and Ibragimov [150, 153, 154.
Many
onferen
es addressed issues related to group analysis of dierential equations. We single
out the latest volumes of \Group Theoreti
al Methods in Physi
s" [75, 76, wherein a list of
previous
olloquia may be found. As of 1994, a new journal [155
overs topi
s in Lie groups and
their appli
ations.
We highly re
ommend the spe
ial issues of A
ta Appli
andae Mathemati
ae on \Symmetries of
Partial Dierential Equations" [304, now available in book form [305, and the other two volumes
in this series edited by Ibragimov [153, 154.
Biographies of Lie and his
ontemporaries are found in [19, 134, 150, 152, 226, 326. Delightful
histori
al notes about symmetry resear
h in general, and Lie's work in parti
ular
an be found in
the new, updated edition of Olver's
lassi
[214. Complete referen
es to Lie's original work are
also given in this book, and in [153. Translations of some of Lie's fundamental papers are available
in [154, 179, 180.
Readers interested in the dierential geometri
al foundation of Lie symmetry analysis, in
luding
topi
s like Lie-Ba
klund mappings, Cartan forms, supersymmetry, graded dierential geometry, and
gauge theories, may want to
onsult [83, 96, 111, 171, 172, 193, 214, 216 [224, 279, 307, 308, 310,
328, and [123, 124 for related REDUCE algorithms.
In this review, we will sparsely address the
omputation of
onservation laws, whi
h through
Noether's famous theorem [200 and its extensions [36, 249, is intrinsi
ally
onne
ted with the
investigation of variational symmetries. Indeed, re
all that variational symmetries admitted by
a Lagrangian system|symmetries that leave the a
tion integral invariant|
an be obtained by
omputing the Lie-Ba
klund symmetries of the
orresponding Euler-Lagrange equations. On
e
the Lie-Ba
klund symmetries are obtained and the variational symmetries are singled out, Boyer's
formulation of Noether's theorem
an be used to
al
ulate the
onservation laws.
In Se
tion 13.2 we dis
uss the purpose, methods and algorithms used in the
omputation of
Lie symmetries. Apart from a detailed des
ription of the methods for
omputing and solving the
determining equations, we address the redu
tion of systems of PDEs into standard and passive
forms. This topi
, in turn, ties in with the
omputation of the size of the symmetry group. In this
se
tion we also address some of the newest trends in the development of symboli
software for Lie
symmetries.
In Se
tion 13.3 we look beyond Lie-point symmetries, addressing
onta
t and generalized symmetries, as well as non
lassi
al or
onditional symmetries.
Se
tion 13.4 is devoted to a detailed review of modern symboli
pa
kages that aid in the investigation of Lie symmetries for systems of dierential equations. The software is grouped a
ording
to the underlying CAS. The review of the available
ode intentionally fo
uses on pa
kages written
after 1985. More details about pioneering eorts and software written prior to 1985
an be found
elsewhere [138. In Se
tion 13.5, three examples illustrate the results that
an be obtained with
the software pa
kages. Finally, in Se
tion 13.6 we draw some
on
lusions.
Although no survey
an be
onsidered exhaustive, our study intended to
over all the Lie
symmetry software, with the ex
lusion of software for Lie group
omputations, su
h as LiE [300,
whi
h is outside the s
ope of this review.
13.2.
13.2.1.
PURPOSE
The
lassi
al \Lie symmetry group of a system of dierential equations" is a lo
al group of point
transformations, meaning dieomorphisms on the spa
e of independent and dependent variables,
whi
h map solutions of the system to solutions. Various other types of lo
al symmetries [14 and
nonlo
al symmetries [6, 31, 32, 35, 36, 305 have been studied, as well as approximate symmetries [20. The software reviewed in this paper attempts to
ompute the
lassi
al Lie-point symmetries. Several pa
kages go beyond that in
omputing
onta
t (or dynami
al) and generalized
(or Lie-Ba
klund) symmetries, non
lassi
al (or
onditional) symmetries. Several programs
ould
be modied to
ompute Cartan's dynami
al symmetries [51, 285 and hidden symmetries. Two
programs that utilize point symmetries to generalize spe
ial solutions or nd similarity variables
re
ently be
ame available [295, 296, 297, 322. The latter programs a
tually try to integrate the
hara
teristi
system of rst-order dierential equations.
Loosely speaking,
onta
t symmetries are generalized symmetries of order one; i.e., the
oe
ients in the ve
tor eld in
lude rst derivatives of the dependent variables.
The name Lie-Ba
klund symmetries,
ommonly used for generalized symmetries, is somewhat
misleading for its
onne
tion with Ba
klund transformations [199. Ba
klund transformations
onstitute a set of
ase-spe
i
equations that allow one to transform one solution into another, whereas
generalized symmetries are innitesimal symmetries involving higher-order derivatives of the dependent variables. Also, note that Ba
klund transformations do not have group properties. Among
the several ways of
omputing Lie-Ba
klund symmetries, this paper fo
uses on an extension of the
original methods due to Lie. For a dis
ussion of the dieren
e with other symmetry approa
hes,
in
luding other
omputer algebra methods, we refer the reader to [97, 99, 117, 118, 327, 329.
Conditional symmetries are found by the \non
lassi
al method of group-invariant solutions,"
as introdu
ed in 1969 by Bluman and Cole [34. Further generalizations of the non
lassi
al method
lead to the less pra
ti
al \weak symmetries" [217, 218, and the method of dierential
onstraints
[213, 214.
Nonlo
al (potential) symmetries [31, 32, 33, 36, 173, 221, 305 for a system of PDEs are
omputed as follows. First, one repla
es (one or more of) the PDEs in the given system by equivalent
onserved forms; se
ond, one introdu
es auxiliary potential variables; nally, one determines the
point symmetries of the resulting auxiliary system of PDEs. The form of an innitesimal generator then determines whether or not it denes a nonlo
al symmetry. The te
hnique of potential
symmetries leads to interesting linearizations involving non-invertible mappings.
\Hidden" symmetries [2, 3, 4, 128 show up when, for example, the order of an ODE is in
reased
by one and the number of symmetries (of the now higher-order ODE) is in
reased at least by two.
Upon subsequent redu
tion, a symmetry may be lost if the transformation is done in non-normal
subgroup variables. Hidden symmetries also o
ur when an ODE is redu
ed in order, and when
this indu
es an additional symmetry that was not a symmetry of the original ODE.
Somewhat related to this is the following: If a higher-order ODE is rewritten as an equivalent
system of rst-order equations, and one analyzes the system of rst-order equations, rather than
the single equation of higher order, the
lass of symmetries (and
onsequently solutions)
an be
substantially enlarged [220, 221, 251. A similar situation arises with PDEs. When a PDE is
redu
ed to a lower-dimensional PDE one
an lose or gain symmetries. Clarkson [60 provides some
examples.
Dis
rete symmetries of dierential equations and dis
onne
ted Lie-point groups have been
onsidered by Reid, Weih and Wittkopf [240.
Pre
ise denitions and a dis
ussion of internal symmetries (also
alled dynami
al symmetries)
and external symmetries are given in [14, 15. It is beyond the s
ope of this survey.
13.2.2.
TIONS
Although Lie's theory is geometri
, one has to resort to dierential-algebrai
methods to
ompute
Lie symmetries. Indeed, the
riterion for a ve
tor eld to be a generator is purely geometri
,
namely,
ertain fun
tions must vanish on a submanifold. But dierential geometry gives us no
tools to implement su
h
riteria in
on
rete appli
ations. The use of dierential algebra imposes
some restri
tions on the type of problems that
an be handled. For instan
e, the dierential
equations must be polynomial in their variables (so that one
an work in the ring of dierential
polynomials); the dierential equations must be solvable for some derivatives, et
.
There are three major methods to
ompute Lie symmetries. The rst one uses prolonged ve
tor
elds, the se
ond utilizes dierential forms (wedge produ
ts) due to Cartan [51. The third one
uses the notion of \formal symmetry" [37, 38, 195. Although restri
ted to evolution systems with
two independent variables, the latter method provides a very qui
k way to
ompute
anoni
al
generalized symmetries. Due to its limited s
ope we will not elaborate on that te
hnique.
13.2.2.1.
The rst method is used in the algorithm or our program SYMMGRP.MAX [52 and in most of the
other Lie symmetry pa
kages. Instead of looking for the Lie group G, one looks for its Lie algebra
L, realized by the ve
tor eld. From the Lie algebra of symmetry generators, one
an obtain the Lie
group of point transformations upon integration of a system of rst-order
hara
teristi
equations
(see Se
tion 13.5.1 for a simple example).
For notational simpli
ity, let us
onsider the
ase of Lie-point symmetries [35, 36, 214. We
follow the method, notations and terminology used in [214.
We start with a system of m dierential equations,
i(x; u(k)) = 0; i = 1; 2; :::; m;
(13.1)
of order
k, with p independent and q dependent (real) variables, denoted by x = (x1 ; x2 ; :::; xp )
p
2 IR ; u = (u1; u2; :::; uq ) 2 IRq : We stress that m; k; p and q are arbitrary positive integers. The
partial derivatives of ul are represented using a multi-index notation, for J = (j1 ; j2; :::; jp) 2 INp;
we denote
jJ j ul
;
(13.2)
x1 j1 x2 j2 :::xp j
where jJ j = j1 + j2 + ::: + jp. Finally, let u(k) denote a ve
tor whose
omponents are all the partial
derivatives of order 0 up to k of all the ul .
The group transformations have the form x~ = g (x; u); u~ =
g (x; u); where the fun
tions g
and
g are to be determined. Note that the subs
ript g refers to the group parameters. Instead of
looking for a Lie group G, we look for its Lie algebra L, realized by ve
tor elds of the form
ulJ
p
X
i
q
X
(x; u) x + 'l (x; u) u l :
=
i
i=1
l=1
(13.3)
The problem is now redu
ed to nding the
oe
ients i(x; u) and 'l (x; u): In essen
e, the
omputer
onstru
ts the kth prolongation pr(k) of the ve
tor eld , applies it to the system (13.1), and
requests that the resulting expression vanishes on the solution set of (13.1).
This sounds straightforward, but the method involves tedious
al
ulations be
ause the length
and
omplexity of the expressions in
rease rapidly as p; q; m, and espe
ially k, in
rease. The number
of determining equations then also rises dramati
ally as is shown in [244 on page 346. Here are
the details and the steps to be performed:
1. Constru
t the kth prolongation of the ve
tor eld in (13.3) by means of the formula
pr(k) = +
where the
oe
ients
J
l
q X
X
J
l
l=1 J
(13.4)
= Di 'l (x; u)
p
X
ul
j =1
Jj Di
(x; u);
(13.5)
where Ji is a p tuple with 1 on the ith position and zeros elsewhere, and Di is the total
derivative operator
q X
X
(13.6)
+
ulJ +J l ; 0 jJ j k:
Di =
xi l=1 J
uJ
The higher-order prolongations are dened re
ursively as
i
J +Ji
l
= Di
J
l
p
X
ul
j =1
J +Jj Di
(x; u); jJ j 1:
(13.7)
2. Apply the prolonged operator pr(k) to ea
h equation i (x; u(k)) and require that
pr(k) i j =0 = 0 i; j = 1; :::; m:
(13.8)
The meaning of
ondition (13.8) is that pr(k) vanishes on the solution set of the originally
given system (13.1). Pre
isely, this
ondition assures that is an innitesimal symmetry
generator of the group transformation; x~ = g (x; u); u~ =
g (x; u); i.e., that u(x) is a
solution of (13.1) whenever u~(~x) is one.
j
3. Choose, if possible, m
omponents of the ve
tor u(k), say v1; :::; vm, su
h that:
(a) Ea
h vi is equal to a derivative of a ul (l = 1; :::; q) with respe
t to at least one variable
xi (i = 1; :::; p):
(b) None of the vi is the derivative of another one in the set.
(
) The system (13.1)
an be solved algebrai
ally for the vi in terms of the remaining
omponents of u(k), whi
h we denote by w. Hen
e, vi = S i(x; w); i = 1; :::; m:
(d) The derivatives of vi, vJi = DJ S i(x; w); where DJ D1j D2j :::Dpj ,
an all be expressed
in terms of the
omponents of w and their derivatives, without ever reintrodu
ing the
v i or their derivatives.
The requirements in step 3 put some restri
tions on the system (13.1), but for many systems
the
hoi
e of the appropriate vi is quite obvious. For example, for a system of evolution
equations
ui
(x ; :::; xp 1; t) = F i(x1 ; :::; xp 1; t; u(k)); i = 1; :::; m;
(13.9)
t 1
where u(k) involves
derivatives with respe
t to the variables xi but not t, an appropriate
i
u
hoi
e is vi = t :
4. Use vi = S i(x; w) to eliminate all vi and their derivatives from the expression (13.8), so that
all the remaining variables are now independent of ea
h other. It is ta
itly assumed that the
resulting expression is now a polynomial in the ulJ :
5. Obtain the determining equations for i(x; u) and 'l (x; u) by equating to zero the
oe
ients
of all fun
tionally independent expressions (monomials) in the remaining derivatives ulJ .
In the above algorithm the variables xi ; ul, and ulJ are treated as independent; the dependent ones
are i and 'l :
In summary, the result of implementing (13.4) is a system of linear homogeneous PDEs for i
and 'l ; in whi
h x and u are independent variables. These are the so-
alled determining or dening
equations for the symmetries of the system. Solving these by hand, intera
tively or automati
ally
with a symboli
pa
kage, will give the expli
it forms of the i(x; u) and 'l (x; u):
The pro
edure, whi
h is thoroughly dis
ussed in [214,
onsists of two major steps: deriving the
determining equations, and solving them expli
itly.
We note that in [30, Bluman proves theorems about the forms of admitted innitesimal transformations, whi
h
over a large
lass of s
alar ODEs and PDEs. In essen
e, the theorems in [36
imply that the
oe
ients of the ve
tor eld are either free or depend linearly on the dependent
variable. Use of these theorems
an signi
antly simplify the tedious work involved in setting up
and solving the determining equations.
1
13.2.2.2.
DIFFERENTIAL FORMS
A dierential geometri
approa
h to invarian
e groups and solutions of PDEs was presented by
Harrison and Estabrook [131. They showed how to derive innitesimal symmetries using Cartan's
exterior dierential
al
ulus. The isove
tor elds, whi
h are the generators of geometri
transformations with suitable algebrai
invarian
e properties, are then used to obtain invariant solutions
of several PDEs, in
luding the heat equation, the Korteweg-de Vries equation, and the va
uum
Maxwell equations.
Brie
y, this method to nd innitesimal symmetries pro
eeds as follows. The system of dierential equations is rewritten as a Pfaan system, that is a system of one-forms. The
ondition for
an isove
tor is then that the
ontra
tion of every form in the
losed ideal generated by the exterior
system with the isove
tor remains in the ideal. That is
ompletely equivalent to the
ondition in
the jet bundle approa
h. From a
omputational point of view there is a disadvantage: the exterior
system usually
onsists of more forms than the PDE system of equations. To redu
e the number of
forms, one
an opt to formulate the system of dierential equations as a
losed exterior dierential
system of two-forms. The underlying manifold is prolonged by new
oordinates, the so-
alled prolonged variables. The exterior dierential system is then prolonged by spe
ial one-forms in these
prolonged variables. The
ondition on these added one-forms is that the prolonged exterior dierential systems remain
losed. This
ondition leads to the determining equations for the
oe
ients
of the manifold.
For the mathemati
al formulation and several worked examples, the reader is referred to [49, 86,
125, 131, 165, 279. Olver's forth
oming book [216 addresses the appli
ation of various approa
hes,
parti
ularly those of Lie and Cartan, to equivalen
e and symmetry problems. That book also
dis
usses the determination of invariants, the
lassi
ation of dierential equations, variational
problems, and so on.
It should also be noted that routines [77, 254, 255, 309 for Cartan's exterior
al
ulus [132 are
in
luded as standard pa
kages of CAS. For instan
e, CARTAN in the MACSYMA Share Library,
EXCALC in REDUCE, and DierentialForms [330, 331 in Mathemati
a. For a list of other
symboli
pa
kages that manipulate obje
ts of dierential geometry we refer the reader to [255.
13.2.3.
To design a reliable and powerful integration algorithm for a system of determining equations the
system needs to be brought into a
anoni
al form. Various related
on
epts appear in the literature,
were authors refer to the normal, orthonomi
, involutive and passive forms [45, 174. All of these
have slightly dierent denitions, for whi
h the reader should
onsult the later
ited literature.
Involution is a geometri
on
ept, independent of
oordinate systems. Passive systems are dened
with respe
t to a given
oordinate system and a given ordering of terms. The same holds for
dierential Grobner bases (DGBs).
The original theory of involutive systems goes ba
k to Cartan [51, Janet [159, 160, Kahler
[161, Kol
hin [167, 168, Riquier [242, Ritt [243, Thomas [286, and Tresse [288, 289. In the
introdu
tion to his le
ture notes [230, Pommaret dis
usses the basi
ideas of the theory of dierential elimination. His des
ription, whi
h
ontains interesting histori
al remarks (see also [228),
relates the work of the Fren
h to the formal
ohomology approa
h advo
ated by D.C. Spen
er and
ollaborators in the USA. A
on
ise survey of the theory of involutive systems with
omplete proofs
is also given by Finikov [95.
Roughly speaking, the methods that redu
e systems of linear homogeneous PDEs into an equivalent, but mu
h simpler standard form,
an be viewed as generalizations of triangulation by GaussJordan elimination, but are applied here to systems of linear PDEs. First, the original system is
appended by all its dierential
onsequen
es. Se
ond, highest derivatives are eliminated, and, if
they o
ur, integrability
onditions are added to the system. The pro
edure is then repeated until
the new system is in involution.
Although S
hwarz did valuable pioneer work on solving systems of linear homogeneous PDEs
[259, he is by far not the only one working on innovative ways of
lassifying, subsequently redu
ing,
and nally solving overdetermined systems of linear homogeneous PDEs. Early eorts to implement
Cartan's exterior-forms approa
h to involutive systems, and the method of Riquier and Janet, are
due to Russian teams [16, 113, 114, 287.
Kornyak, Fedorova and Fush
hi
h [92, 170, Bo
harov [40, Bo
harov and Bronstein [42, Ganzha
and Yanenko [114, Gerdt and Zharkov [119, Pankrat'ev [223 and Topunov [287, amongst others,
partially implemented algorithms to redu
e systems of PDEs.
The GRBASES algorithm of Pankrat'ev and the implementation of the Riquier-Janet method
by Topunov are both in REFAL. Both programs
an only redu
e linear systems of PDEs, this in
ontrast with the program DIFFGROB2 of Manseld dis
ussed below.
A good des
ription of the relation between the Riquier-Janet theory and the modern implementations (Grobner bases) is given by Topunov [287 (see also [266). For an introdu
tion of Grobner
bases, in
luding Bu
hberger's algorithm [46, 47, we re
ommend the new books by Cox, Little and
O'Shea [74, and Be
ker and Weispfenning [25. Most of this work dire
tly relates to Lie symmetry
omputations, with the ex
eption of Pankrat'ev, who oers general algorithms for the
omputation
of Grobner bases in dierential and dieren
e modules.
In the West, Reid and
ollaborators [234, 235, 237, 238, 241, S
hwarz [259, 266, 267, and
Wolf and Brand [320, 321, 324, 323, 325, partially implemented algorithms to redu
e systems of
PDEs. Their work led to sophisti
ated symboli
ode in MACSYMA, Maple, and REDUCE for
that purpose.
In [266, 267, S
hwarz des
ribes the algorithm InvolutionSystem, based on the theory of dierential equations due to Riquier [242 and Janet [159, 160, to transform a linear system of PDEs
into involutive form. In modern language: the involutive form is a DGB with respe
t to the sele
ted term ordering. If all
onsistent orderings for the terms in the system of PDEs are known,
the algorithm InvolutionSystem may be applied repeatedly to determine a universal Grobner base
[267. S
hwarz designed his algorithm InvolutionSystem for a spe
i
purpose, namely to determine the size of the Lie symmetry group of a given system of PDEs without having to integrate
the determining equations. We devote Se
tion 13.2.5 to this important topi
, where details about
S
hwarz's program SYMSIZE
an be found.
As dened by Reid [235, a standard form of a system is obtained by repeating the following
steps: (i) write ea
h equation in solved form with respe
t to its highest order derivative, (ii) repla
e
these highest order derivatives throughout the rest of the system, (iii) add any new equations arising
from integrability
onditions.
As said, standard-form algorithms have their origins in the work of Riquier and Janet for passive forms. Note that the Riquier-Janet form is not a standard form, but
an be fairly simply
transformed into a standard form. Also, in method, design of algorithm, and in a
tual implementation, modern standard form algorithms are quite dierent from the original ine
ient methods
and algorithms proposed by Riquier and Janet.
A rst, but brief a
ount of Reid's algorithm standard form [235, whi
h also has it roots in the
lassi
al Riquier-Janet theory, appeared in [233 and [234. The algorithm was rst implemented
in MACSYMA, and later on a MAPLE version be
ame available [241.
The algorithm standard form redu
es systems of PDEs to a simplied standard form. Again,
the pro
edure
an be viewed as a generalization to linear dierential equations of the Gaussian
redu
tion method for matri
es or linear systems. The algorithm now takes as input the system of
PDEs and a matrix whi
h spe
ies a
omplete ordering on the derivatives appearing in the system.
It then redu
es the system of PDEs to an equivalent simplied ordered triangular system with
all integrability
onditions in
luded and all redundan
ies (dierential and algebrai
) eliminated.
Reid's algorithm implements an equivalen
e
lass approa
h to the problem of bringing a system of
PDEs into a standard form. For that purpose, Reid developed a new
ompletion method based on
a free dire
tion index (rather than the monomials of the Riquier-Janet theory).
Within standard form, Reid uses an \update strategy" based on updating lists of equations: one
term equations, easy equations, hard (or yet-to-be
lassied) equations; with spe
ial user-dened
tuning knobs (parameters), so that the user
an
ontrol the
ow of equations between the various
lists. Thus standard form works on easier parts of the system rst, a strategy that be
omes
ru
ial
when dealing with large systems.
Further details about Reid's algorithm and examples of its use
an be found in [237, where it
is shown how dire
ted graphs representing the dependen
ies amongst the system's variables
an be
used to simplify or numeri
ally integrate the system. On
e the system is in standard form, one
an
ontinue with the automati
determination of a Taylor series solution of the system to a spe
ied
nite degree.
Reid and Wittkopf's pa
kage [241 fa
ilitates automated interfa
ing with major symmetry pa
kages su
h as DIMSYM [277, LIESYMM [49, and SYMMGRP.MAX [52, and also with the differential Grobner basis pa
kage DIFFGROB2 [189. A TEX interfa
e between standard form and
Hi
kman's program [142, 143 that uses physi
al variable notation has been provided by Lisle. Full
details and many illustrative examples of the pa
kage, whi
h, besides the fun
tion standard form,
in
ludes other powerful algorithms for symmetry analysis of PDEs, are given in [241.
Reid and M
Kinnon developed a re
ursive algorithm
alled Rsolve Pdesys [239 that builds on
Reid's standard form [235 and on algorithms of Abramov and Kvashenko [5 and Bronstein [44.
Reid and M
Kinnon's algorithm RSolve Pdesys now nds parti
ular solutions of linear systems of
PDEs using only ODE solution te
hniques. Applied to symmetry problems, their algorithm will
nd all polynomial/rational solutions of the determining equations provided the symmetry group
is nite-dimensional.
Several other approa
hes, and
onsequently implementations, are possible to
omplete a given
system of PDEs to an involutive system. S
hu, Seiler and Calmet [256 present an algorithm
in AXIOM to perform this task. Their method is highly geometri
al and their implementation is
based in part on the Cartan-Kuranishi theorem [45, whi
h assures that the integrability
onditions
for the determining equations
an be found in a nite number of steps. A detailed des
ription of
their programs,
alled JET, is given in Seiler's thesis [274. JET
an be viewed as an environment
for
omputations within the geometri
theory of PDEs based on the jet bundle formalism. Some
standard tasks are put into AXIOM pa
kages. One su
h pa
kage is
alled CartanKuranishi, whi
h
ompletes a given system to an involutive one. Another pa
kage, in development at the time
of writing,
ontains a pro
edure to set up determining equations for
lassi
al and non-
lassi
al
symmetries.
Hartley and Tu
ker [132 implemented an algorithm (in REDUCE) to analyse involutive systems
of exterior forms, based on the Cartan-Kahler theory. Later they extended their program [133,
originally in REDUCE, now partly rewritten in Maple, to non-involutive systems. For su
h systems
their
ompletion pro
edure
onstru
ts the needed integrability
onditions. Their work
orresponds
that of S
hu, Seiler and Calmet on involutive systems, however, Hartley and Tu
ker use exterior
systems.
A
ording to my sour
es, a student of Fa
kerell (Sydney) has implemented the Vessiot approa
h
to involution. Vessiot's method
an be viewed as a dual version to the exterior system approa
h.
GROB2 a pa
kage
alled DIRMETH is available to
ompute the determining equations related to
the symmetry redu
tion of PDEs via the dire
t method of Clarkson and Kruskal [65.
The program CRACK by Wolf and Brand [323, 324, 325 also
arries out a Grobner Basis
analysis but in slightly modied form. First, the algorithm is enri
hed by the integration of PDEs
whenever possible, but only in su
h a way, that the new integrated PDEs are still polynomial
in the Grobner basis. In other words, the `
riti
al pair
ompletion steps' of the Grobner basis
algorithm and the integrations used within CRACK are
onsistent. Sele
tive integration
an redu
e
the
omplexity and aid in solving the determining equations, in parti
ular for systems for whi
h
pure Grobner basis methods would be unfeasible. Se
ond, for e
ien
y reasons, only a restri
ted
ompletion algorithm is used, although it is the authors intention to extend it to a
omplete Grobner
basis algorithm in the future.
A
ording to a re
ent paper [240, Wittkopf is also developing an algorithm,
alled di redu
e,
whi
h attempts to redu
e polynomially nonlinear systems of PDEs to the form of a DGB. In
essen
e, the algorithm is a dierential analogue of Bu
hberger's elimination algorithm for polynomial equations. Wittkopf's algorithm uses redu
tion rather than pseudo-redu
tion, and in
orporates strategies for e
ient memory management.
Finally, Oaku [211 is designing and implementing software in the
omputer algebra system
Risa/asir to automati
ally nd the stru
ture of the solution spa
e of systems of linear PDEs.
Oaku's method is based on the notion of Grobner basis and the Bu
hberger algorithm applied to
rings of dierential operators with polynomial
oe
ients (Weyl algebra).
13.2.4.
The most
hallenging part of Lie symmetry analysis by
omputer, involves the design of an \integrator" for the overdetermined systems of linear homogeneous PDEs. This topi
is also of importan
e
in the study of so-
alled adjoint symmetries of dierential equations [250, and in many other areas
where determining equations of the same type o
urs. Ideally, good integration
ode should be
appli
able to generi
systems of linear dierential equations, whi
h do not ne
essarily
ome from
symmetry analysis.
In the
ontext of Lie symmetry analysis, one
an aim at the design of faster and more powerful
algorithms that work for large systems of determining equations, typi
ally a few hundred, and that
automati
ally redu
e systems to where they
an be handled intera
tively with the
omputer, or by
hand.
Sin
e the early developments [135, 257, 260, 283, 284, 285 of semi-heuristi
methods to solve
determining equations, substantial progress has been made in understanding the mathemati
s of
this problem and a new breed of algorithms is now available. These algorithms attempt to
lose
the gap between solution te
hniques for ODEs and PDEs (
onsult [280 for an impressive review
and large bibliography).
Two other important topi
s tie in with the integration of the determining equations: (i) the
transformation of the determining equations into standard and passive forms; and (ii) the
omputation of the size of the symmetry group dis
ussed in Se
tions 13.2.3 and 13.2.5, respe
tively.
The design of algorithms and programs to bring the determining equations into standard form
were a major step forward. On
e systems are redu
ed into standard involutive form or de
oupled,
subsequent integration is more tra
table and reliable. One
ould use separation of variables, standard te
hniques for linear dierential equations, and spe
i
heuristi
rules given below. The only
determining equations left for manual handling should be the \
onstraint" equations or any other
equations whose general solutions
annot be written expli
itly in
losed form.
In order to be able to make the determination of
ertain types of Lie generators into a de
ision
pro
edure, one needs an algorithm for solving linear homogeneous ODEs. Su
h equations are
always obtained as the lowest equation of the redu
ed determining system, with redu
tion based
on lexi
ographi
al term ordering. An important step towards this goal is the fa
torization as it is
applied in SPDE [264, 265. An in-depth review of issues related to the implementation of this and
other algorithms is given in [269.
After sear
hing the relevant literature [42, 233, 234, 235, 239, 259, 266, 267, 287, 320, 321
[323, 324, 325, it is
lear that many mathemati
al questions remain open. Despite the innovative
eorts of Reid, S
hwarz, Wolf and Brand, and many others, there is no general algorithm available
to integrate an arbitrary (overdetermined) system of determining equations that
onsists of linear
homogeneous PDEs for the 's and the 's.
Most integration algorithms are based on a set of heuristi
rules [135, 165, 257, 260, 277, 284.
In the
omputer programs reviewed in Se
tion 13.4, the following rules are used.
1. Integrate single term equations of the form
jI j f (x1 ; x2 ; :::; xn )
= 0;
(13.10)
x1 i x2 i :::xn i
where jI j = i1 + i2 + ::: + in ; to obtain the solution
1
f (x1 ; x2 ; :::; xn ) =
n iX
k 1
X
k=1 j =0
(13.11)
(13.12)
by fj = 0 (j = 0; 1; :::; n): More generally, this method of splitting equations (via polynomial
de
omposition) into a set of smaller equations is also allowed when fj are dierential equations
themselves, if the variable xk is missing.
3. Integrate linear dierential equations of rst and se
ond order with
onstant
oe
ients.
Integrate rst-order equations with variable
oe
ients via the integrating fa
tor te
hnique,
provided the resulting integrals
an be
omputed in
losed form.
4. Integrate higher-order equations of type
n f (x1 ; x2 ; :::; xn )
= g(x1; x2; :::; xk 1; xk+1; :::; xn);
(13.13)
xk n
n su
essive times to obtain
(xk )n g(x ; x ; :::; x ; x ; :::; x )
f (x1 ; x2 ; :::; xn ) =
(13.14)
1 2
k 1 k+1
n
n!
n
5. Solve any simple equation (without derivatives) for a fun
tion (or a derivative of a fun
tion)
provided that it both (i) o
urs linearly and only on
e, and (ii) depends on all the variables
that o
ur as arguments in the remaining terms.
6. Expli
itly integrate exa
t equations.
7. Substitute the solutions obtained above in all the equations.
8. Add dieren
es, sums or other linear
ombinations of equations (with similar terms) to the
system, provided these
ombinations are shorter than the original equations.
With these simple rules, and perhaps a few more, the determining system
an often be drasti
ally
simplied. Amazingly, in many
ases nothing beyond the above heuristi
rules is needed to solve
the determining equations
ompletely. If that is not possible, after simpli
ation, the programs
return the remaining unsolved dierential equations for further inspe
tion. In most programs, the
user
an then intera
tively simplify and fully solve the determining equations on the
omputer,
thereby minimizing human errors.
At least for Lie-point symmetries, solving the determining equations
an usually be done by
hand, using elementary results from the theory of linear PDEs. Solving them on a
omputer may
be time
onsuming, sin
e the simplest approa
h varies greatly from
ase to
ase. Furthermore, a
omputer program may a
identally not
at
h the most general result and therefore may return
an in
omplete symmetry group. The author is aware of this problem, whi
h o
urred when testing
some of the reviewed symmetry programs. Even worse, the
omputer algorithm may not be able
to determine the solution of the determining equations in a nite number of steps.
13.2.5.
GROUP
S
hwarz [266, 267 and Reid [234, 235, 236 independently developed algorithms for determining
the size of the Lie symmetry groups of dierential equations without integrating the determining
equations expli
itly.
S
hwarz's algorithm SYMSIZE [266, 267 is available with the
omputer algebra system REDUCE, as part of the pa
kage SPDE (see Se
tion 13.4.2). S
hwarz also translated SYMSIZE into
the language of S
rat
hpad II, the prede
essor of AXIOM. Use of SYMSIZE
ir
umvents some of
the short
omings mentioned at the end of the previous se
tion. Indeed, if a dierential equation
has no other than obvious symmetries or if the symmetry group is small (be
ause all generators
are algebrai
and of low degree), SYMSIZE will greatly help in
ompletely solving the symmetry
problem.
In
ontrast to the heuristi
algorithms for the expli
it
omputation of the symmetry generators,
the size of the symmetry group
an always be determined with SYMSIZE in a nite number of steps.
SYMSIZE a
epts a system of PDEs as input, and allows one to
ompute a priori the number of
free parameters if the group is nite and the number of unspe
ied fun
tions of the group is innite.
In turn, SYMSIZE allows one to test a posterior if the solution of the determining equations is
omplete. In
ases where some, perhaps all, symmetries are known by inspe
tion or from the
physi
s of the problem, the knowledge of the size of the symmetry group
an evade an expensive
sear
h for more, perhaps nonexistent, symmetries.
At the heart of SYMSIZE is the pro
edure InvolutionSystem, whi
h transforms the determining system into an involutive system by means of a
riti
al pair/
ompletion algorithm. Similar
algorithms are applied in
omputing Grobner bases in polynomial ideal theory (see Se
tion 13.2.3).
Con
urrently, yet independent of S
hwarz, Reid [234, 235, 236 realized that triangularization
algorithms may allow bypassing the expli
it solution of the determining equations and
ompute
the size of the symmetry group and the
ommutators immediately. Reid developed the program
SYMCAL [235, written originally in MACSYMA, but now
onverted by Reid and Wittkopf into
Maple [241.
In [233, 236, a non-heuristi
algorithm stru
ture
onstant is presented, based on Taylor and
standard form, whi
h always determines (in a nite number of steps) the dimension and the stru
ture
onstants of the nite part of the Lie symmetry algebra. An extension of the algorithm
[233 also allows one to
lassify dierential equations (with variable
oe
ients) a
ording to the
stru
ture of their symmetry groups. Furthermore, the approa
h advo
ated by Reid applies to the
determination of symmetries of Lie-
onta
t and Lie-Ba
klund types, as well as potential symmetries.
The algorithm des
ribed in [236, provides information about the dimension and
ommutators
of the Lie symmetry algebra. It is based on expli
it Taylor expansions of the symmetry generators,
and therefore is
omputationally expensive and restri
ted to nite dimensional Lie algebras. The
newest Maple algorithm [238 allows one to
ompute the dimension and the
ommutation relations
without Taylor expansions; hen
e, it is appli
able to innite-dimensional Lie algebras.
Readers interested in the problem of determining the \size" of the solution spa
e for arbitrary
involutive systems should
onsult a re
ent paper by Seiler [273. Seiler's results
an be applied to
linear and nonlinear determining systems.
Finally, we should mention that skillful use of the tools for redu
ing systems of linear homogeneous PDEs, available within the pa
kage CRACK [323, 324, 325,
an also greatly assist in the
investigation of the size of the symmetry group.
13.3.
The dis
ussion of symmetries other than point symmetries is limited here to those for whi
h
symmetry software is already available. For a general review of various types of symmetries we
refer to the new edition of Olver's book [214 and to Clarkson [60.
13.3.1.
For the
omputation of generalized symmetries or Lie-Ba
klund symmetries [13, 214 the use of
symboli
programs is even more appropriate. The pro
edure to determine symmetries is essentially
the same as that for point symmetries, although the
al
ulations are lengthier and more time
onsuming. In a generalized ve
tor eld, whi
h still takes the form of (13.3), the fun
tions i and
l may now depend on a nite number of derivatives of u, i.e.,
=
p
X
i
i=1
q
X
'l (x; u(k) ) l :
u
l=1
(13.15)
If k = 1 the generalized symmetry determines a
lassi
al
onta
t symmetry and vi
e versa, at least
in the
ase of one dependent variable. The even simpler
ase k = 0; with u(0) = u; leads to point
symmetries. Olver [214 dis
usses various possibilities to simplify the
al
ulations, for example by
putting the symmetries in evolutionary form, or by xing the order of derivation on whi
h the 's
and 's may depend.
13.3.2.
Re
ently it was shown that the \non
lassi
al method of group-invariant solutions," originally introdu
ed in [34,
an determine new solutions of various physi
ally signi
ant nonlinear PDEs.
Examples in
lude the nonlinear S
hrodinger (NLS) equation [58, and its
ylindri
al version
[62; the Boussinesq equation [176; the Kadomtsev-Petviashvili equation [72 and other members
of its hierar
hy [314; the Burgers equation [88, 232; the telegraph equation [207; the FitzhughNagumo equation [87, 88, 210, and other rea
tion-diusion equations [66; the Helmholtz [209 and
shallow water wave equations [69, 70; and a
lass of nonlinear heat equations [67, 68. Levi and
Winternitz re
ently showed how
onditional symmetries
an be determined for the 2D-Toda latti
e,
a dierential-dieren
e equation [177. For a well-do
umented perspe
tive on the
omputation of
non
lassi
al symmetries we re
ommend [71 and [59, 60.
An example is the Boussinesq equation where the new redu
tions that follow from appli
ation
of the non
lassi
al method were dis
overed earlier with a dire
t method [58, 57, 65. The dire
t
method was also applied to the Zabolotskaya-Khokhlov equation [61 and the Davey Stewartson
system [63, 64. Re
ently, Olver [215 proved that both methods are equivalent in the
ase of ber
preserving transformations, whi
h means that the new independent variables depend only on the
old independent variables, not on the original dependent variables. Arrigo, Broadbridge and Hill
[17 expli
itly derive
riteria for whi
h the dire
t method and the non
lassi
al symmetry method
lead to the same results; they use the Burgers and Boussinesq equations as illustrative examples.
For shallow water wave equations, Clarkson and Manseld [69 have shown that the non
lassi
al
symmetry method
an lead to parti
ular solutions whi
h
an not be obtained via the singular
manifold method. Estevez [87, 88 and Pu
i [232 pointed out some interesting
onne
tions between
the dire
t method of Clarkson and Kruskal [65. Estevez [87, 88 and Estevez and Gordoa [89 also
ompare the non
lassi
al symmetry method with the singular manifold approa
h from Painleve
analysis. Nu
i [206 shows how Ba
klund transformations
an be obtained via the investigation
of non
lassi
al symmetries.
In
ontrast to Lie-point symmetries, for example, the transformations
orresponding to non
lassi
al (or
onditional) symmetries neither leave the dierential equation invariant, nor transform all
the solutions into other solutions. They merely transform a subset of solutions into other solutions.
A
ounting for \non
lassi
al symmetries," the program should automati
ally add the q invariant
surfa
e
onditions [34, 213,
p
X
u
Ql (x; u(1) ) = i (x; u)
xi
i=1
l
(13.16)
and their dierential
onsequen
es, to the system (13.1). However, the in
lusion of non
lassi
al
symmetries, and perhaps other types of symmetries as dis
ussed in [14, 15, 213, requires solving systems of determining equations whi
h are no longer linear. Consequently, new integration
algorithms must be designed.
It should be noted that various other types of
onditional symmetries
ould be
onsidered. For
instan
e, one
ould ask under what extra
onditions a
lass of PDEs would admit a symmetry
hosen beforehand. Extensive work on this problem, whi
h we will not address here, was done by
Fush
hi
h [100.
13.4.
In this se
tion we review the most modern Lie symmetry programs,
lassied a
ording to the
underlying
omputer algebra system. Fo
using on new trends, pa
kages written before 1985 are
only brie
y mentioned. Resear
hers interested in more details about some of the pioneering work
ould
onsult [138.
13.4.1.
COMPUTER IMPLEMENTATION
Ideally, a fully automated software pa
kage for Lie symmetries should
onsist of ee
tive, powerful
algorithms and fast pro
edures for the following tasks:
1. derivation of the determining equations for large or
ompli
ated systems of equations;
2. redu
tion of determining equations into so-
alled standard form;
3. nding the size of the symmetry group;
4. determining any obvious symmetry generators;
5. simplifying and integrating the determining equations to
ompute the generators, if not all
the generators have been found yet.
Then the program should be able to exe
ute the following steps in the order relevant to the spe
i
appli
ation:
(a)
al
ulation of
ommutator tables, based on the results of 1, 2 and 3;
(b)
al
ulation of group invariant solutions.
The program should be able to handle:
al
ulation of nonlo
al (potential) symmetries [31,
32, 36, 221, 305;
al
ulation of non
lassi
al redu
tions (
onditional symmetries) and resulting
solutions;
al
ulation of generalized symmetries; and
al
ulation of equivalent
onservation laws.
Furthermore, it should be able to a
ept systems with free unknown (
lassi
ation) fun
tions
[186, 236, 277 as input. If so, questions su
h as \for whi
h values of these parameters or parametri
fun
tions does a given ODE or PDE have pres
ribed symmetries or spe
ial solutions"
ould be
answered.
Other ideas
ould be in
orporated in the design of faster and more e
ient symboli
software
for Lie symmetry analysis. Let me give a
ouple of examples. Lengthy
al
ulations should be
broken up into smaller pie
es by
onsistently taking advantage of the \linear algebra" stru
ture
of the Lie symmetry problem. For instan
e, prolongations should be applied to ve
tor elds of
single equations or subsets of equations, and not to the whole system at on
e. Furthermore, full
expansions of the prolonged ve
tor elds should be halted until the expli
it forms are a
tually
needed. Avoiding lengthy, redundant expansions will make the generation of the determining
equations mu
h more e
ient, parti
ularly for non
lassi
al and Lie-Ba
klund symmetries.
Currently, no software handles this entire ambitious program. Many Lie symmetry programs
arry out parts of the listed tasks. Making matters worse, the available Lie symmetry programs,
with the ex
eption of Lie by Head [135, work with spe
i
CAS, whi
h has to be bought separately.
In Table 1 we list the most modern software pa
kages, along with information about developers
and distributors. In Table 2 we summarize the s
ope of these pa
kages.
With the ex
eption of DELiA and DIMSYM, all the programs listed in Table 1 are publi
domain software. Potential users
an obtain the software from the developers or through the
referen
ed sour
es. The main
ost in using these pa
kages is related to the
ost of the underlying
CAS. As a rule of thumb, individual
opies of Mathemati
a, MACSYMA, Maple, REDUCE, and
the like,
ost about 10% of the pri
e of the platform you buy them for.
13.4.2.
REDUCE PROGRAMS
In the early '80s, S
hwarz developed his well-do
umented program SPDE [257, 258, 260, 261,
262, 263, 264, The program automati
ally derives and often su
essfully solves the determining
equations for Lie-point symmetries with minimal intervention by the user. Sin
e 1986 SPDE is
distributed together with REDUCE for various types of
omputers, ranging from PCs to CRAYs.
In 1994 version 1.0 of SPDE be
ame available. Although S
hwarz de
ided to keep the old name,
the new program is drasti
ally dierent. A
ording to the do
umentation [268, SPDE 1.0 guarantees that all innitesimal symmetry generators with algebrai
oe
ients will be obtained if
the equations are nonlinear and of order higher than one. Con
erning the input, the equations
must be algebrai
in their arguments. There is no restri
tion on the number of independent and
dependent variables, and the equations
an have any number of
onstant parameters (no arbitrary
fun
tions). The program
omputes the determining equations, then generates a Grobner basis for
the determining system in a term ordering spe
ied by the user (total degree, lexi
ographi
, et
.).
The integration of the redu
ed system is
arried out automati
ally, the symmetry generators and
their
ommutator table are displayed in LATEX (if so desired).
Based on Cartan's exterior
al
ulus [51, 95, 140, Edelen [79, 80, 81, 82, Gragert [120, and
Gragert, Kersten and Martini [125 used
omputer algebra systems to
al
ulate the
lassi
al Lie
symmetries of dierential equations. More re
ently, Gragert [121, 122 added a pa
kage for more
general Lie algebra
omputations, in
luding
ode for higher-order and super symmetries and super
prolongations. Kersten [164, 165 further perfe
ted the software pa
kage for the
al
ulation of the
Lie algebra of innitesimal symmetries (in
luding Lie-Ba
klund symmetries) of exterior dierential
systems.
Eliseev, Fedorova and Kornyak [85, wrote
ode in REDUCE-2 to generate (but not solve)
the system of determining equations for point and
onta
t symmetries. Their paper dis
usses
the algorithm and shows three worked examples. Fedorova and Kornyak [91, 93 generalized the
algorithm to in
lude the
ase of Lie-Ba
klund symmetries.
The intera
tive REDUCE program NUSY by Nu
i [202, 203, 204, 205, in
luded with this book
[208, generates determining equations for Lie-point, non-
lassi
al, Lie-Ba
klund and approximate
symmetries and provides intera
tive tools to solve them. The manual [202, 204, 208 gives a
lear des
ription of the various routines with their s
ope and limitations, and has several worked
examples.
The pa
kage CRACK by Wolf and Brand [323, 324, 325 solves overdetermined systems of dierential equations with polynomial terms. To do this, it uses
ode for de
oupling, separating and simplifying PDEs. Integration of exa
t PDEs and dierential fa
torization are also possible. CRACK
has many appli
ations that are fa
ilitated via spe
ial tools. For instan
e, the fun
tion LIEORD
an aid in the investigation of Lie symmetries of ODEs. With CRACK one
an also
onstru
t
the Lagrangian for a given se
ond-order ODE, and nd rst integrals via the integrating fa
tor
method. In attempting to solve standard ODEs, the program makes use of the REDUCE pa
kage
ODESOLVE written by Ma
Callum [187. The fun
tions and tools available within CRACK allow
simpli
ation and integration of linear homogeneous PDEs, beyond those derivable via symmetry
analysis.
Upon
ompletion of CRACK, Wolf went on to develop three new REDUCE programs,
alled
LIEPDE, QUASILINPDE and APPLYSYM, whi
h all make use of the tools of CRACK.
LIEPDE [321 nds Lie-point and
onta
t symmetries of PDEs by deriving and solving a few
simple determining equations, before
ontinuing with the
omputation of the more
ompli
ated
determining equations. This idea, whi
h makes the program highly e
ient, was used in Wolf's
FORMAC program [318, 319, 320, and is also implemented in the design of the feedba
k me
hanism
of SYMMGRP.MAX [52. For solving the determining equations, LIEPDE makes use of modules of
the pa
kage CRACK dis
ussed above. The dieren
e is that within LIEPDE the steps are
arried
out automati
ally, without intervention by the user. This approa
h is parti
ularly useful when
applied to large systems of PDEs, or in the
omputation of higher-order symmetries, where spa
e
and memory limitations
ome into play.
The aim of QUASILINPDE [322 is to nd the solutions of quasi-linear PDEs. These solutions
are then used by APPLYSYM [322, whi
h applies the symmetry to lower the order of ODEs, to
al
ulate similarity variables for PDEs, to ee
tively redu
e the number of independent variables of
a system of PDEs, and to generalize spe
ial known solutions of ODEs and PDEs. To our knowledge,
APPLYSYM is one of the rst symboli
programs that truly applies point symmetries that
an
be
al
ulated with the program LIEPDE. The program APPLYSYM is automati
but
an also be
used intera
tively. Thus far, APPLYSYM is only appli
able to point symmetries for whi
h the
generators are at worst rational. The a
tual problem solving is done in all these programs through
a
all to the pa
kage CRACK for solving overdetermined systems of PDEs.
In [115, 116, Gerdt introdu
ed the program HSYM for the expli
it
omputation of higher-order
symmetries for PDEs. If the given system of equations has arbitrary parameters, the ne
essary
onditions for the existen
e of higher order symmetries will lead to a system of algebrai
equations
in the parameters. Via the program ASYS, that algebrai
system is redu
ed into standard form
via a Grobner basis algorithm. The fo
us in Gerdt's work is on the investigation of the integrability of polynomial type nonlinear evolution equations, by verifying the existen
e of higher order
symmetries and their asso
iated
onservation laws.
Sarlet and Vanden Bonne [250 oer spe
i
pro
edures to assist in the
omputation of adjoint
symmetries of se
ond-order ODEs. This assistan
e, however, is limited to the
onstru
tion of
determining equations for
ertain
lasses of adjoint symmetries, whi
h are of the same nature as
determining equations for (generalized) symmetries, and relies on other pa
kages su
h as DIMSYM
below for solving these determining equations. In addition, pro
edures have been written for testing
whether a given adjoint symmetry
an give rise to a Lagrangian or a rst integral for the original
se
ond-order equations.
The program DIMSYM by Sherring [277, in
ollaboration with Prin
e, was inspired by Head's
symmetry program LIE [135, in turn in
uen
ed by SPDE [260, 264, but is mu
h larger and grew
independently of it during development. It is
apable of nding various types of symmetries,
urrently, point symmetries, Lie-Ba
klund, and
onditional symmetries. DIMSYM
an isolate spe
ial
ases, bring the determining equations in standard form for example, and aid in the solution of
group
lassi
ation problems. It attempts to determine the generators and allows one to
he
k
whether or not the generators are
orre
t. It allows the user to spe
ify the dependen
e of the symmetry ve
tor eld
oe
ients, whi
h is parti
ularly pra
ti
al if one wants to
ompute Lie-Ba
klund
symmetries. DIMSYM provides the user with a lot of
exibility: ansatze
an be made, simpli
ation routines
an be
alled separately, manual intervention is possible, et
. Quite often su
h
interventions indeed allow the user to
omplete the desired
omputations, whereas the DIMSYM
in auto-pilot mode may not.
DIMSYM has routines that
onvert the input equations into standard form. Another attra
tive
feature of the pa
kage is that the integrator for the determining equations also works for systems
of linear homogeneous dierential equations not ne
essarily obtained from symmetry analysis. The
overall strategy of the solver is to put the system of determining equations into standard form
based on Reid's algorithm (see Se
tion 13.2.3), while solving expli
itly any equations in the system
that the algorithm is
apable of solving.
Finally, we mention the programs by Ito [156, 157, 158 for the determination of symmetries
and
onservation laws of systems of evolution equations. Ito's program does not use any of the
algorithms dis
ussed in Se
tion 13.2.2, but uses innitesimal symmetries to determine the form of
onservation laws.
13.4.3.
MACSYMA PROGRAMS
One of the rst programs was written by S
hwarzmeier and Rosenau [246, 271. Their program
al
ulates the determining equations, simplies them a bit, but does not solve them automati
ally.
The MACSYMA version of the program SYMCON [294, whi
h was originally written in muMATH, tries to
ompute Lie-point and Bessel-Haagen generalized symmetries (of any order) and
their
onservation laws. Vafeades later produ
ed PDELIE [295, 296, 297, 298, whi
h is a drasti
ally improved version of SYMCON [294. The pa
kage PDELIE attempts to produ
e similarity
solutions of ODEs, analyze PDEs with a multipli
ative or additive s
alar parameter, and
ompute
the
ommutator table and the stru
ture
onstants of the Lie algebra. PDELIE also allows one to
ompute the Noether
onservation laws of variational systems.
PDELIE
onsists of several subroutines. Let's dis
uss the main ones. The fun
tion PL SYM,
produ
es the determining equations and the generators of the Lie group. It uses a standard
form algorithm by Reid and a set of heuristi
rules to fa
ilitate the integration. The fun
tion
PL SOLVE tries to nd the invariants of the symmetry group. Using these invariants, it then
dimensionally redu
es the given dierential equation. In
ases where the redu
ed equation is an
ODE, it tries to integrate expli
itly, thus arriving at spe
ial similarity solutions of the original
equation. The fun
tions PL COMTAB performs
omputations with elements in the
ommutator
table of stru
ture
onstants of the Lie algebra. The fun
tion PL CON
omputes the densities of
the Noether
onservation laws of systems of variational and divergen
e type.
Just as PDELIE, the program SYM DE by Steinberg [282, 283, 284 was re
ently added to the
out-of-
ore library of MACSYMA. Steinberg's program
omputes innitesimal symmetry operators
and the expli
it form of the innitesimal transformations for simple systems. In
ases where the
program
annot automati
ally nish the
omputation, the user
an intervene and, for instan
e, ask
for innitesimal symmetries of polynomial form. The program solves some (or all) of the determining equations automati
ally and, if needed, the user
an (intera
tively) add extra information.
Steinberg intends to extend his program so that it would in
lude the
al
ulation of generalized
symmetries.
The program SYMMGRP.MAX written by Champagne, Hereman and Winternitz [52, 137, is
a modi
ation of an earlier pa
kage [53 that has been extensively used over the last de
ade at the
University of Montreal and in many pla
es elsewhere. It has been tested on hundreds of systems
of equations and has thus been solidly debugged.
The
exibility within SYMMGRP.MAX and the possibility of using it intera
tively, allows the
user to nd the symmetry group of arbitrarily large and
ompli
ated systems of equations on
relatively small
omputers. For example, whenever the prolongation
an be applied su
essfully to
the
omplete system, or a subset thereof, it produ
es a list of determining equations. This list is
free of trivial fa
tors, dupli
ation and dierential redundan
ies.
To make SYMMGRP.MAX work for large systems of dierential equations, the designers followed the path that would be taken in manual
al
ulations. That is, obtain in as simple a manner
as possible the simplest determining equations, solve them and feed the information ba
k to the
omputer. Partial information
an be extra
ted very rapidly. For instan
e, one
an derive a subset
of the determining equations, su
h as those that o
ur as
oe
ients in the highest derivatives in
the independent variables. These are usually single-term equations, whi
h express that the
oe
ients of the ve
toreld are independent of some variables or depend linearly on some of the other
variables.
A feedba
k me
hanism fa
ilitates the solution of the determining equations step by step on the
omputer; hen
e, avoiding human error in the algebrai
simpli
ations. Typi
ally, users will provide
information about the 's and ''s, as it be
omes available from solving the determining equations
step by step. The amount of intera
tion by the user will depend on the
omplexity of the system
of dierential equations and on the
apa
ity of the
omputer used. A worked example showing
the use of the feedba
k me
hanism is given in [52. Needless to say, with the feedba
k me
hanism,
the program SYMMGRP.MAX
an also be used to verify previously
al
ulated solutions of the
determining equations and, hen
e, dete
t errors in the literature on the subje
t.
Although not designed for that purpose, the program SYMMGRP.MAX
an be easily adapted
to
ompute the determining equations
orresponding to non
lassi
al symmetries [60, 67, 68, 69, 71.
In [71, Clarkson and Manseld give a detailed explanation of su
h an adaptation. Their proof of
orre
tness of the proposed adaptation is based on the theory of Grobner bases.
13.4.4.
MAPLE PROGRAMS
In [49, Carminati, Devitt and Fee present LIESYMM for
reating the determining equations via
the Harrison-Estabrook pro
edure. Within LIESYMM various intera
tive tools are available for
integrating the determining equations, and for working with Cartan's dierential forms. Their
program is independent of Donsig's dierential forms pa
kage diorms, also available in Maple.
Khai T. Vu (Department of Mathemati
s, Monash University, Clayton, Vi
toria, Australia) has
translated Head's muMATH program LIE [135, dis
ussed in Se
tion 13.4.7, into Maple syntax.
In 1994, the Maple version of LIE, whi
h
omputes Lie-point symmetries, was still being tested
and, therefore, was not yet released. The -version of the program
omputes the determining
equations, solves them, and gives the expli
it forms of the (ve
toreld)
oe
ients together with
the generators.
Hi
kman [142, 143 oers a
olle
tion of Maple routines that aid in the
omputation of Liepoint symmetries, non-lo
al symmetries, and Wahlquist-Estabrook-type prolongations. The tools
for symmetry analysis in
lude user-friendly pro
edures to enter names of variables, to
reate total
derivatives, to generate and prolong ve
tor elds, and to derive and partially solve determining
equations. Program and do
umentation are available via anonymous FTP from math.
anterbury.a
.nz.
Manseld has developed the pa
kage DIRMETH for the
omputation of symmetries via the
dire
t method proposed by Clarkson and Kruskal [65. The program DIRMETH is part of DIFFGROB2, dis
ussed in detail in Se
tion 13.2.3. Other eorts in the design of pa
kages for the dire
t
method are given in the do
toral thesis of Williams [311.
13.4.5.
MATHEMATICA PROGRAMS
Herod [141 developed MathSym for deriving the determining equations
orresponding to Lie-point
symmetries, in
luding non
lassi
al (or
onditional) symmetries. Upon derivation of the determining
equations, the program redu
es these equations via an algorithm based on the method of Riquier
and Janet. Herod's do
toral thesis
ontains the well-do
umented
ode of MathSym and appli
ations
to various equations from
uid dynami
s.
Re
ently, the pa
kages Lie.m and Bae
klund.m have been added by Baumann [22, 23, 24
to MathSour
e, the Mathemati
a Program Library. Baumann's program Lie.m [22 follows the
stru
ture of our MACSYMA program SYMMGRP.MAX [52 very
losely. Users familiar with
SYMMGRP.MAX will have a short learning
urve with Lie.m. In
ontrast to SYMMGRP.MAX,
the program Lie.m
an handle trans
endental fun
tions in the input equations. The newest version
of Lie.m
an be used to
ompute point symmetries,
onta
t symmetries and non
lassi
al symmetries. Lie.m brings the determining equations in
anoni
al form via the pro
edure of Janet and
Riquier, and goes on to solve the determining equations automati
ally. A nite set of integration
rules, similar to the ones des
ribed in Se
tion 13.2.4, are implemented.
On
e the solution of the determining equations is obtained, the program
an
ontinue with the
omputation of the ve
tor basis, ideals, and
ommutator table of the Lie algebra, its stru
ture
onstants, Casimir operators, and its metri
tensor.
Baumann's pa
kage Bae
klund.m [23, 24
ontains fun
tions that attempt to
ompute generalized symmetries for PDEs and ODEs and invariants of ODEs only. When applied to se
ond-order
ODEs, the program attempts to verify if the
omputed symmetries are of variational type. If so,
the program
al
ulates the
orresponding invariants (integrals of motion). For the expli
it
al
ulations to be su
essful, quite often one has to spe
ify that the
oe
ients in the ve
toreld are
polynomials in the
oordinates and momenta. With this \ansatz" one may not be able to obtain
all the generalized symmetries, but one may su
essfully obtain expli
it forms of invariants.
Berube and de Montigny [27 produ
ed Lie-symmetry
ode in Mathemati
a. Their program
symmgroup.
omputes the determining equations for Lie-point symmetries. In its syntax and
format symmgroup.
losely follows the stru
ture of SYMMGRP.MAX. The data for the program
may
onsist of DEs with arbitrary fun
tions. Trans
endental fun
tions in both dependent and
independent variables are also permitted. In [27, three well-
hosen examples are given to illustrate
the
apabilities of the program.
Finally, Coult (while at Carleton College, Northeld, Minnesota) developed a Mathemati
a
program, temporarily
alled symmgroup.m, for the
omputation of the determining equations
orresponding to Lie-point symmetries of a large
lass of dierential equations (with polynomial terms).
13.4.6.
S hwarz [263 rewrote SPDE [262, 264 for use with version 1 of S rat hpad II, a symboli manipulation program developed by IBM. S rat hpad II is now superseded by AXIOM.
Seiler and
o-workers [256, 274 are designing a pa
kage that will
ompute determining equations
for
lassi
al and non-
lassi
al symmetries. See Se
tion 13.2.3 for a des
ription of their program
JET for geometry
omputations based on the jet bundle formalism.
13.4.7.
MUMATH PROGRAMS
The program LIE by Head [135 is based on version 4.12 of muMATH, but is self
ontained and
runs on IBM
ompatible PCs. As a matter of fa
t, the program
omes bundled with a limited
version of muMATH. Head's program
al
ulates and solves the determining equations (for Liepoint symmetries) automati
ally for single equations and systems of dierential equations. LIE
also
omputes the Lie ve
tors and their
ommutators. Interventions by the user are possible but
are rarely needed. The sour
e
ode of the program is available, in
luding the heuristi
routines
that attempt to solve the determining equations. Due to the limitations of muMATH, the program
LIE is bounded by the 256 KB of memory for program and workspa
e. For a program of limited
size, LIE is remarkable in its a
hievements.
Version 4.2 of LIE is freely available by FTP from various publi
domain software ar
hives su
h
as SIMTEL and asso
iated ar
hives. By this printing, version 4.3 of LIE also will be available. In
addition to Lie-point symmetries, the new version will be able to
ompute
onta
t and generalized
(Lie Ba
klund) symmetries.
The SYMCON pa
kage written by Vafeades [294 also uses muMATH to
al
ulate the determining equations (without solving them). The program is restri
ted to point symmetries. Furthermore,
the program veries whether the symmetry group is of variational or divergen
e type and
omputes
the
onservation laws asso
iated with the symmetries. Unfortunately, these programs are
onned
to the 256 KB memory a
essible by muMATH and,
annot presently handle very large systems of
equations. This limitation motivated Vafeades to rewrite his SYMCON program in MACSYMA
syntax [295, 296, 297, 298. The MACSYMA version of SYMCON
an handle generalized symmetries and their
onservation laws.
For
ompleteness, Mikhailov developed software in muMATH to verify the integrability of
systems of PDEs by testing for the existen
e of higher symmetries. The program
omputes spe
ial
symmetries,
anoni
al
onservation laws, and
arries out
onformal transformations to bring PDEs
into
anoni
al form. With their PC program, Mikhailov, Shabat and Sokolov [195 produ
ed an
exhaustive list of integrable nonlinear S
hrodinger-type equations. In this
ontext, integrable means
that the equations have innitely many
onserved quantities and innitely many lo
al symmetries.
13.4.8.
Kornyak and Fush
hi
h [101, 104, 170 developed programs in Turbo C and AMP for the
omputation of Lie-Ba
klund symmetries. Their programs also
lassify equations with arbitrary parameters
and fun
tions with respe
t to su
h symmetries. It is important to note that their programs redu
e
the determining equations into passive form (see Se
tion 13.2.3). All integrability
onditions are
then expli
it and, therefore, the resulting system is in involution.
We should mention their two FORMAC programs. The rst program,
alled LB, was written
in the PL/1 language by Fedorova and Kornyak [91, 92. The su
essor,
alled LBF, was developed
by Fush
hi
h and Kornyak [104. Both programs
reate the system of determining equations for
Lie-Ba
klund symmetries and attempt to solve these equations. The program LBF, with its 1362
lines of PL/1-FORMAC
ode, is
ompletely automati
and
onsists of 37 subroutines, one of whi
h
brings the determining equations in passive (Riquier-Janet) form. The program LB [92 is available
from the Computer Physi
s Communi
ations Program Library in Belfast. The above programs were
designed for low-memory requirements so that they
ould run on PCs.
The PL/1-based FORMAC pa
kage CRACKSTAR developed by Wolf [318, 319, 320 allows
one to investigate Lie symmetries of systems of PDEs, besides dealing with dynami
al symmetries
of ODEs [225, and the like. A good overview of the
apabilities of CRACKSTAR is given in
[319; a des
ription of the routines and worked examples are in [225. For e
ien
y, CRACKSTAR
generates and solves rst-order determining equations early on, and then
ontinues with the higherorder determining equations. The su
essor of CRACKSTAR is the REDUCE pa
kage CRACK
dis
ussed already in Se
tion 13.4.2.
Gerdt [115, 116, Gerdt and Zharkov [119 and Gerdt, Shva
hka and Zharkov [117, 118 used
REDUCE and PL/1-FORMAC to investigate the integrability of nonlinear evolution equations.
Their program FORMINT
ontains algorithms to
al
ulate Lie-Ba
klund symmetries and
onserved
densities, but does not use the jet bundle formalism.
The
al
ulation of the Lie group by
omputer was also proposed by Popov [231, who used the
program SOPHUS for the
al
ulation of
onservation laws of evolution equations.
In [42, Bo
harov and Bronstein present SCoLAr, a pa
kage written in standard PASCAL that
nds innitesimal symmetries and
onservation laws of arbitrary systems of dierential equations.
An appli
ation of SCoLAr to the Kadomtsev-Pogutse equations is given in [129.
The PC pa
kage DELiA, standing for \Dierential Equations with Lie Approa
h," is an outgrowth of the SCoLar proje
t [42. DELiA, written in Turbo PASCAL by Bo
harov and his
ollaborators [37, 38, 39, 40, is a stand-alone
omputer algebra system for investigating dierential equations. It performs various tasks based on Lie's approa
h, su
h as the
omputation of
Lie-point and Lie-Ba
klund symmetries,
anoni
al
onserved densities and generalized
onservation
laws, simpli
ation and partial integration of overdetermined systems of dierential equations, et
.
The methods used in DELiA and many examples are well des
ribed in the user guides [37, 40.
In order to be able to handle large problems, DELiA rst generates and solves rst-order
determining equations, and then
ontinues to generate and solve the higher-order determining
equations. The analyzer/integrator, whi
h is available as a separate tool at the user level, in
ludes
a general algorithm for passivization [42, together with a set of integration rules for linear and
quasi-linear systems of PDEs. Currently, a MS Windows version of DEliA,
alled MS Win DELiA,
is under development.
Using the algorithmi
language REFAL, Topunov [287 developed a software pa
kage for symmetry analysis that
ontains subroutines to redu
e determining systems in passive form.
13.5.
EXAMPLES
In this se
tion we give three examples that illustrate the
omputation of Lie-point symmetries
with symboli
software. The rst and simplest example involves a single s
alar nonlinear equation.
The se
ond example illustrates how symmetries of a nonlinear
omplex equation are
omputed by
splitting the equation into a system of nonlinearly
oupled equations for the real and imaginary
parts of the original dependent variable. The last and most
ompli
ated example involves a system
of ve
tor equations that needs to be split into equations for its s
alar
omponents in order to
ompute its Lie symmetries.
13.5.1.
u3 uxxx = 0:
(13.17)
Clearly, this is one equation with two independent variables and one dependent variable. The
assignments of the variables are as follows:
x 7 ! x[1 ; t 7 ! x[2 ; u 7 ! u[1:
(13.18)
This permits us to rewrite the equation (13.17) in a form a
epted by the program SYMMGRP.MAX; i.e.,
ut
e1 : u[1,[0,1-u[1^3*u[1,[3,0.
+
t + 'u :
x
t
u
(13.19)
eta2
u[1
= 0;
eta2
x[1
= 0;
eta1
u[1
= 0;
2 phi1
u[12
= 0;
(13.20)
2 eta1
x[12
= 0;
3 phi1
x[13
= 0;
3 eta1
u[13
x[13
= 0;
2 phi1
u[1x[1
phi1
x[2
3 phi1
u[13
(13.21)
3u[1 xeta1
[1 + 3 phi1 = 0:
These determining equations are easily solved expli
itly, either automati
ally with SPDE, LIE and
PDELIE, or with the feedba
k me
hanism within SYM DE and SYMMGRP.MAX. The general
solution, rewritten in the original variables, is
x = k1 + k3 x + k5 x2 ;
(13.22)
t
= k2 3k4 t ;
(13.23)
u
' = (k3 + k4 + 2k5 x) u ;
(13.24)
where k1; :::; k5 are arbitrary
onstants. The ve innitesimal generators then are
G1 = x ;
(13.25)
G2 = t ;
(13.26)
G3 = xx + uu ;
(13.27)
G4 = 3tt + uu ;
(13.28)
2
G5 = x x + 2xuu :
(13.29)
Clearly, (13.17) is invariant under translations (G1 and G2 ) and s
aling (G3 and G4): The
ow
orresponding to ea
h of the innitesimal generators
an be obtained via simple integration. As an
example, let us
ompute the
ow
orresponding to G5. This requires integration of the rst-order
system
dx~
= x~2 ;
x~(0) = x ;
u[1
eta2
x[2
d
dt~
d
t~(0)
= 0;
t;
(13.30)
du~
d
= 2~xu~ ;
u~(0) = u ;
where is the parameter of the transformation group. One readily obtains
x
x~() =
(1 x) ;
t~()
u~()
= (1
(13.31)
(13.32)
t;
u
:
x)2
(13.33)
Therefore, we
on
lude that for any solution u = f (x; t) of equation (13.17), the transformed
solution
x~ ~
u~(~x; t~) = (1 + x~)2 f (
(13.34)
1 + x~ ; t)
will solve u~t~ u~3u~x~x~x~ = 0:
13.5.2.
THE NONLINEAR SCHRODINGER
EQUATION
In order to
ompute the Lie-point symmetries of the
elebrated nonlinear S
hrodinger equation [1,
iut + uxx + ujuj2 = 0;
(13.35)
one needs to repla
e the single
omplex equation by a
oupled system. One way of doing that is
by introdu
ing the real and imaginary parts v; w of the
omplex variable u via u(x; t) = v(x; t) +
iw(x; t): This yields
vt + wxx + w(v 2 + w2 ) = 0;
(13.36)
2
2
wt vxx v (v + w ) = 0:
One alternative is to repla
e (13.35) by a system
onsisting of the equation itself and its
omplex
onjugate [58 and to interpret the variables u and v = u as real,
k ut + uxx + v u2 = 0;
(13.37)
2
k vt + vxx + u v = 0:
In order to work with real quantities throughout, the imaginary unit i was temporarily repla
ed
by the
onstant k during the
omputations. On
e the determining equations are obtained, k = i
should be reintrodu
ed.
Another alternative is to write u(x; t) = R(x; t) exp(i
(x; t)); thus repla
ing (13.35) by a
oupled
system
Rt + 2Rx
x + R
xx = 0;
(13.38)
2
3
Rxx R
t R
x + R = 0;
for the real modulus R(x; t) and real phase
(x; t).
Adhering to (13.36), SYMMGRP.MAX (or for that matter any other symmetry program)
qui
kly generates the twenty determining equations for the
oe
ients of the ve
tor eld
= x + t + 'v + 'w :
(13.39)
x
t
v
w
The rst eleven single-term determining equations are similar to (13.20), and provide information
about the dependen
ies of the 0s and the 0s on x; t; v and w; and their linearity in the latter two
independent variables. The remaining nine determining equations are a bit more
ompli
ated, but
the entire system is readily solved.
In the original variables, the solution reads
x = k1 + 2k4 t + k5 x ;
(13.40)
t = k2 + 2k5 t ;
(13.41)
v
' = k3 w k4 xw k5 v ;
(13.42)
w
' = k3 v + k4 xv k5 w ;
(13.43)
where k1 ; :::; k5 are arbitrary
onstants. As in the previous examples, the
omplete symmetry
algebra is spanned by ve ve
tor elds (generators):
G1
G2
G3
G4
G5
= x ;
= t ;
= wv vw ;
= 2tx x(wv vw ) ;
= xx + 2tt vv ww :
(13.44)
(13.45)
(13.46)
(13.47)
(13.48)
If we had
arried out the
omputations with (13.38), where u(x; t) = R(x; t) exp(i
(x; t)); we
would have found:
G1
G2
G3
G4
G5
= x ;
= t ;
=
;
= 2tx x
;
= xx + 2tt RR :
(13.49)
(13.50)
(13.51)
(13.52)
(13.53)
Either way, (13.35) is invariant under translations in spa
e and time (G1 and G2 ): Generator
G3
orresponds to adding an arbitrary
onstant to the phase of u: The Galilean boost is generated
by G4 : Finally, G5 indi
ates invarian
e of the equation under s
aling (or dilation). Similarity
redu
tions
an then be obtained by solving the
hara
teristi
equations,
dx
x
dv dw
= dt
=
= ;
t 'v 'w
(13.54)
= 0;
= 0:
(13.55)
The a
tual redu
tions
an be found in [58, where a quite general nonlinear S
hrodinger equation
is treated. It is well known [176 that all the redu
tions of the NLS
an be obtained from G1
through G5 ; in other words, non
lassi
al symmetries would not lead to new symmetry redu
tions.
To
ompute non
lassi
al symmetries of (13.36), it su
es to repla
e vt and wt from (13.55). If
t 6= 0; we set t = 1 for simpli
ity. Thus,
vt
wt
=
=
x vx + 'v ;
x wx + 'w :
(13.56)
(13.57)
The
ase t = 0 has to be
onsidered separately. Sin
e SYMMGRP.MAX allows the user to give
information about the
oe
ients in the ve
tor eld, the
omputation
an now pro
eed as in the
lassi
al
ase. For worked examples, we refer the reader to [60, 71.
13.5.3.
As an example of a large system of dierential equations, we take the equations for Magneto-HydroDynami
s (MHD) [98 and
arry out the sear
h for Lie-point symmetries with SYMMGRP.MAX.
The MHD equations, with or without dissipative terms, have be
ome a ben
hmark for developers of Lie symmetry pa
kages. Nu
i [201
omputed the
lassi
al symmetries in 1984 and
ommented that the
al
ulations by hand took her a year. Mi
hel Grundland undertook the same
formidable task and also nished the job in about a year. He
onveyed to the author that the long
winter in Newfoundland (where he was at the time) helped.
If we negle
t dissipative ee
ts, and thus restri
t the analysis to the ideal
ase, the MHD
equations
an be redu
ed to
+ (~v r) + r ~v
t
~v
1 ~ 2) (H~ r)H~
+
(
~v r)~v + r(p + H
t
2
~
H
+ (~v r)H~ + H~ r ~v (H~ r)~v
t
r H~
p
p
(
)
+
(
~v r)( )
t
= 0;
= ~0 ;
(13.58)
(13.59)
= ~0 ;
= 0;
= 0;
(13.60)
(13.61)
(13.62)
~
with pressure p, mass density ,
oe
ient of vis
osity ,
uid velo
ity ~v and magneti
eld H:
Using the rst equation, we eliminate from the last equation, hen
e repla
ing it by
p
+ p(r ~v) + (~v r)p = 0:
t
(13.63)
If we split the ve
tor equations in s
alar equations for the ve
tor
omponents, we have a system of
nine equations, with four independent variables and eight dependent variables. For
onvenien
e, we
denote the
omponents of the ve
tor ~v by vx; vy and vz ; not to be
onfused with partial derivatives
of v:
The variables to be eliminated are sele
ted as follows: for the rst seven variables and the ninth
variable we pi
k the partial derivatives with respe
t to t of ; vx; vy ; vz ; Hx; Hy ; Hz and p. From the
eighth equation, we sele
t Hx =x for elimination.
We will only
onsider the
ase where 6= 0: We ran this
ase on a Digital VAX 4500 with 64
MB of RAM, and on an IBM Ris
6000 workstation with 32 MB of RAM. On the VAX it took 50
minutes of CPU time, on the IBM workstation 1 hour and 50 minutes, for SYMMGRP.MAX to
reate the 222 determining equations for the
oe
ients of the ve
tor eld
=
+
+
y + z + t + ' + 'p
x
y
z
t
p
'v
+
'v
+
'v
+
'H
+
'H
vx
vy
vz
Hx
Hy
x
+ 'H
:
Hz
Using SYMMGRP.MAX intera
tively, we then integrated the determining system and obtained
the solution expressed in the original variables,
x = k2 + k5 t k8 y k9 z + k11 x ;
(13.64)
= k3 + k6 t + k8 x k10 z + k11 y ;
= k4 + k7 t + k9 x + k10 y + k11 z ;
= k1 + k12 t ;
= 2 (k11 k12 k13 ) ;
= 2 k13 p ;
= k5 k8 vy k9 vz + (k11 k12 )vx ;
= k6 + k8 vx k10 vz + (k11 k12 )vy ;
= k7 + k9 vx + k10 vy + (k11 k12 )vz ;
= k13 Hx k8Hy k9Hz ;
= k13 Hy + k8Hx k10 Hz ;
= k13 Hz + k9Hx + k10Hy :
It is
lear that there is a thirteen-dimensional Lie algebra spanned by the generators:
G1 = t ;
G2 = x ;
G3 = y ;
G4 = z ;
G5 = tx + v ;
G6 = ty + v ;
G7 = tz + v ;
G8 = xy yx + vx v vy v + Hx H Hy H ;
G9 = yz zy + vy v vz v + Hy H Hz H ;
G10 = zx xz + vz v vx v + Hz H Hx H ;
G11 = xx + yy + zz 2 + vx v + vy v + vz v ;
G12 = tt + 2 (vx v + vy v + vz v ) ;
G13 = 2 + 2pp + HxH + Hy H + Hz H :
y
z
t
'
'p
'v
'v
'v
'H
'H
'H
x
y
z
x
y
z
(13.65)
(13.66)
(13.67)
(13.68)
(13.69)
(13.70)
(13.71)
(13.72)
(13.73)
(13.74)
(13.75)
(13.76)
(13.77)
(13.78)
(13.79)
(13.80)
(13.81)
(13.82)
(13.83)
(13.84)
(13.85)
(13.86)
(13.87)
(13.88)
Thus, the MHD equations (13.58)-(13.62) are invariant under translations G2 through G4, Galilean
boosts G5 through G7, rotations G8 through G10 , and dilations G11 through G13: In
ontrast to the
results obtained for the 1+1 and the 2+1 dimensional versions of the MHD problem, the dimension
of the Lie algebra for (13.58)-(13.62) in the full 3+1 dimensions (x; y; z and time t) is independent
of the value of the
oe
ient of vis
osity : Our results
onrm those in [98, and of those of
Grundland and Lalague [126, 127, who
omputed the
lassi
al and some non
lassi
al symmetries
of the MHD, and also
lassied all the subalgebras in
onjuga
y
lasses. The MHD system and
our results have been used by other investigators [49, 241, 277 to test their symmetry programs.
13.5.4.
Champagne and Winternitz [53 used SYMMGRP.MAX to
ompute the Lie-point symmetries of
the Korteweg de Vries equation with variable
oe
ients,
ut + f (x; t)uux + g (x; t)uxxx = 0;
(13.89)
illustrating that SYMMGRP.MAX
an easily handle equations involving arbitrary fun
tions.
Also in [53, the point symmetries of a modied Kadomtsev-Petviashvili equation
(uxxx 2u3x 4ut)x 6uxxuy + 3uyy = 0;
(13.90)
in 2+1 dimensions are
omputed with SYMMGRP.MAX. This example was
hosen be
ause it leads
to an innite-dimensional Lie algebra involving four arbitrary fun
tions of t:
A
ompletely worked example of the
al
ulation of Lie-point symmetries of a system of PDEs is
given in [52. This example shows the use of the feedba
k me
hanism within SYMMGRP.MAX to
ompletely solve the determining equations. It involves the Karpman equations [163, for whi
h our
symmetries were independently veried with REDUCE programs by Kersten and Gragert (private
ommuni
ation), Sherring and Prin
e [278, and Wolf [321.
Finally, SYMMGRP.MAX was re
ently used to
ompute Lie-point symmetries of two large
systems of equations representing
lassi
al eld theories [139. Currently, the author is adapting
SYMMGRP.MAX for the
al
ulation of Lie-point symmetries of dieren
e-dierential equations.
13.6.
CONCLUSION
The various programs that we reviewed need very little data and are straightforward to use provided
the user has a
ess to and knows the basi
s of the underlying CAS, su
h as MACSYMA, Maple,
Mathemati
a and REDUCE.
Apart from the theoreti
al study of the underlying mathemati
s, there is a need for further
development and implementation of ee
tive algorithms for generating, redu
ing, simplifying and
fully solving the determining equations for (
lassi
al and non
lassi
al) Lie-point symmetries and
generalized or Lie-Ba
klund symmetries.
The availability of sophisti
ated symboli
programs
ertainly will a
elerate the study of symmetries of physi
ally important systems of dierential equations in
lassi
al me
hani
s,
uid dynami
s,
elasti
ity, and other applied areas.
ACKNOWLEDGEMENTS
While writing this survey I was able to
onsult with many experts in this eld. I am grateful for their
suggestions,
omments,
orre
tions and
riti
isms whi
h greatly helped improve the manus
ript.
Sin
ere thanks goes to M. Ablowitz, B. Abraham-Shrauner, G. Baumann, D. Berube, G.
Bluman, A. Bo
harov, F. Cantrijn, B. Champagne, P. Clarkson, R. Conte, N. Coult, B. Fu
hssteiner,
V. Gerdt, K. Govinder, P. Gragert, M. Grundland, F. Gungor, A. Head, B. Herbst, S. Herod, M.
Hi
kman, D. Holm, M. Kruskal, D. Levi, M. Ma
Callum, F. Mahomed, E. Manseld, L. Mar
hildon,
A. Mikhailov, M.C. Nu
i, P. Olver, J. Ondi
h, G. Prin
e, D. Rand, G. Reid, M. Roelofs, W. Sarlet,
W. Seiler, F. S
hwarz, J. Sherring, W.-H. Steeb, S. Steinberg, Z. Thomova, M. Torrisi, P. Vafeades,
K. Vu, T. Wil
ox, P. Winternitz, T. Wolf, and C. Wulfman.
Spe
ial thanks goes to S. Collart for inviting me to write a preliminary version of this survey paper for Euromath Bulletin. It is my pleasure to thank N.H. Ibragimov for giving me the
opportunity to
ompletely rework that review paper and to publish it as a
hapter in this book.
I am grateful to T. Hearn and H. Melenk for providing me with a free
opy of REDUCE.
Resear
h for this survey paper was supported in part by Grant # CCR-9300978 of the National
S
ien
e Foundation of the United States of Ameri
a.
Table 1
Distributor
Developer's Address
Refs.
CRACK
LIEPDE
& APPLYSYM
(REDUCE)
REDUCE
Network Library
[321
[323
T.Wolfmaths.qmw.a .uk
DELiA
(Pas
al)
Beaver Soft
715 O
ean View Ave
Brooklyn
NY 11235, USA
A. Bo
harov et al.
A. Bo
harov
Wolfram Resear
h
100 Trade Center Dr.
Urbana-Champaign
IL 61820-7237, USA
[37
E. Mansfield
Dept. of Maths.
Univ. of Exeter
Exeter EX4 4QE
United Kingdom
[189
J. Sherring
G. Prin
e
S
hool of Maths.
Latrobe University
Bundoora, VI 3083
Australia
[277
V. Eliseev et al.
V. Eliseev
Lab. Comp. Te
h. Aut.
JINR, Dubna
Mos
ow Region
141980 Russia
[85
Cost:
$ 300
DIFFGROB2
(Maple)
DIMSYM
(REDUCE)
LaTrobe University
S
hool of Maths.
Cost:
LIE
(REDUCE)
$ 225
CPC
Program Library
Belfast
N. Ireland
Cat. No. AABS
galois.maths.qmw.a
.uk
/ftp/pub/
ra
k
alexeiwri. om
lizmaths.exeter.a .uk
eu
lid.exeter.a
.uk
pub/liz
matjslure.latrobe.edu.au
G.Prin
elatrobe.edu.au
ftp.latrobe.edu.au
/ftp/pub/dimsym
Table 1 ont.
Distributor
Developer's Address
Refs.
LIE
(muMath)
(independent)
SIMTEL
A. Head
CSIRO
Div. Mat. S
i. & Te
h.
Clayton, Vi
toria
3168 Australia
[135
headrivett.mst. siro.au
Lie
& LieBae
klund
(Mathemati
a)
Wolfram
Resear
h
MathSour
e
0202-622
0204-680
G. Baumann
Abt. Math. Phys.
Universitat Ulm
D-7900 Ulm
Germany
[22
[23
[24
[120
[123
J. Carminati et al.
G. Fee
Dept. Comp. S
i.
University of Waterloo
Waterloo, Canada
[49
S. Herod
Program Appl. Math.
University of Colorado
Boulder, CO 80309, USA
[141
M.C. Nu
i
Dept. di Mathemati
a
Universita di Perugia
06100 Perugia, Italy
[202
LIEDF/INFSYM
& others
(REDUCE)
Liesymm
(Maple)
MathSym
(Mathemati
a)
NUSY
(REDUCE)
Waterloo
Maple
Software
(Pa
kages)
wuar
hive.wustl.edu
/edu/math/msdos/..
../adv.diff.equations/lie42
bautheophys.physik.uni-ulm.de
mathsour
e.wri.
om
/pub/PureMath/Cal
ulus
gragertmath.utwente.nl
kerstenmath.utwente.nl
wmsidaisy.uwaterloo.
a
wmsidaisy.waterloo.edu
sherodnewton.
olorado.edu
newton.
olorado.edu
pub/mathsym
nu igauss.dipmat.unipg.it
Table 1 ont.
Distributor
Developer's Address
Refs.
PDELIE
(MACSYMA)
MACSYMA
Out-of-Core
Library
P. Vafeades
Dept. of Eng. S
i.
Trinity University
San Antonio, TX 78212, USA
[295
[296
[297
peterengr.trinity.edu
REDUCE
Program Lib.
Rand Corp.
F. S
hwarz
GMD, Inst. SCAI
D-53731 Sankt Augustin
Germany
[264
[266
fritz.s hwarzgmd.de
[241
S. Steinberg
Dept. Math. & Stat.
Univ. New Mexi
o
Albuquerque, NM 87131, USA
[284
D. B
erub
e & M. de Montigny
M. de Montigny
D. Berub
e
Centre Traitement Inform.
Univ. Laval, St.-Froy
Canada G1K 7P4
[27
B. Champagne et al.
W. Hereman
Dept. Math. Comp. S
i.
Colorado S
h. of Mines
Golden, CO 80401, USA
[52
SPDE
& SYMSIZE
(REDUCE)
SYMCAL
(Maple
& MACSYMA)
SYM DE
(MACSYMA)
MACSYMA
Out-of-Core
Library
symmgroup.
(Mathemati
a)
SYMMGRP.MAX
(MACSYMA)
CPC
Program Lib.
Belfast
N. Ireland
Cat. No. ACBI
gumbo.engr.trinity.edu
redu e-netlibrand.org
reidmath.ub
.
a
math.ub
.
a
pub/reid
stanlymath.unm.edu
wheremanlie.mines.
olorado.edu
ftp.mines.
olorado.edu
pub/papers/math
s dept/symmetry
or
onta
t:
p
v1.am.qub.a
.uk
Table 2
Name
System
Developer(s)
Point
General.
Non l.
CRACK
REDUCE
Yes
DELiA
Pas al
Bo harov et al.
Yes
Yes
No
Yes
DIFFGROB2
Maple
Mansfield
Redu tion
DIMSYM
REDUCE
Sherring
Yes
Yes
No
Yes
LIE
REDUCE
Eliseev et al.
Yes
Yes
No
No
LIE
muMath
Head
Yes
Yes
No
Yes
Lie
Mathemati a
Baumann
Yes
No
Yes
Yes
LieBae klund
Mathemati a
Baumann
No
Yes
No
Intera tive
LIEDF/INFSYM
REDUCE
Yes
Yes
No
Intera tive
LIEPDE
REDUCE
Yes
Yes
No
Yes
Table 2 ont.
Name
System
Developer(s)
Point
General.
Non l.
Liesymm
Maple
Carminati et al.
Yes
No
No
Intera tive
MathSym
Mathemati a
Herod
Yes
No
Yes
Redu tion
NUSY
REDUCE
Nu i
Yes
Yes
Yes
Intera tive
PDELIE
MACSYMA
Vafeades
Yes
Yes
No
Yes
SPDE
REDUCE
S hwarz
Yes
No
No
Yes
SYMCAL
Maple/MACSYMA
Redu tion
SYM DE
MACSYMA
Steinberg
Yes
No
No
Partially
symgroup.
Mathemati a
Berub
e & de Montigny
Yes
No
No
No
SYMMGRP.MAX
MACSYMA
Champagne et al.
Yes
No
Yes
Intera tive
SYMSIZE
REDUCE
S hwarz
Redu tion
Bibliography
[1
[2
[3
[4
[5
[6
[7
[8
[9
[10
[11
and
, Solitons, Nonlinear Evolution Equations and Inverse
London Mathemati
al So
iety Le
ture Notes on Mathemati
s 149 (Cambridge
University Press, Cambridge, UK, 1991).
, Hidden symmetries and linearization of the modied Painleve-In
e
equation, J. Math. Phys. 34 (1993) 4809-4816.
and
, Hidden symmetries of dierential equations, Contemp. Math. 160 (1994) 1-13.
and
, Hidden symmetries of nonlinear ordinary
dierential equations, in: Exploiting Symmetry in Applied and Numeri
al Analysis, Pro
.
AMS-SIAM Summer Seminar, Fort Collins, Colorado, 1992, Eds.: E. Allgower, K. Georg and
R. Miranda. Le
tures in Applied Mathemati
s 29 (Ameri
an Mathemati
al So
iety, Providen
e, Rhode Island, 1993) 1-10.
and
, Fast algorithms to sear
h for the rational solutions of linear dierential equations with polynomial
oe
ients, in: Pro
. ISSAC '91, Bonn,
Germany, 1991, Ed.: S.M. Watt (ACM Press, New York, 1991) 267-270.
and
, Nonlo
al symmetries. Heuristi
approa
h, J. Sov. Math. 55 (1991) 1401-1450.
and
, Eds., Exploiting Symmetry in Applied and
Numeri
al Analysis, Pro
. AMS-SIAM Summer Seminar, Fort Collins, Colorado, 1992, Le
tures in Applied Mathemati
s 29 (Ameri
an Mathemati
al So
iety, Providen
e, Rhode Island,
1993).
, Nonlinear Ordinary Dierential Equations in Transport Pro
esses (A
ademi
Press, New York, 1968).
, Nonlinear Partial Dierential Equations in Engineering. Vol. I (A
ademi
Press, New York, 1965).
, Nonlinear Partial Dierential Equations in Engineering. Vol. II (A
ademi
Press, New York, 1972).
, Symmetry in nonlinear me
hani
s, in: Nonlinear Equations in the Applied
S
ien
es, Eds.: W.F. Ames and C. Rogers (A
ademi
Press, Boston, 1992) 33-78.
M.J. Ablowitz
S attering,
P.A. Clarkson
B. Abraham-Shrauner
B. Abraham-Shrauner
B. Abraham-Shrauner
S.A. Abramov
A. Guo
P.G.L. Lea h
K.Yu. Kvashenko
E. Allgower, K. Georg
W.F. Ames
W.F. Ames
W.F. Ames
W.F. Ames
N.Kh. Ibragimov
R. Miranda
[12
[13
[14
[15
[16
[17
[18
[19
[20
[21
[22
[23
[24
[25
[26
and
, Eds., Nonlinear Equations in the Applied S
ien
es (A
ademi
Press, Boston, 1992).
and
, Lie-Ba
klund Transformations in Appli
ations,
Studies in Applied Mathemati
s 1 (SIAM, Philadelphia, 1979).
and
, Internal, external and generalized symmetries, Adv. in Math. 100 (1993) 53-100.
and
, Internal symmetries of dierential equations, in: Modern Group Analysis: Advan
ed Analyti
al and Computational Methods in
Mathemati
al Physi
s. Pro
. Int. Workshop A
ireale, Catania, Italy, 1992, Eds.: N.H. Ibragimov, M. Torrisi and A. Valenti (Kluwer A
ademi
Publishers, Dordre
ht, The Netherlands,
1993) 7-21.
and
, Dokl. Akad. Nauk S.S.S.R. 214 (1974)
p. 296, in Russian.
and
, Non
lassi
al symmetry solutions and the
methods of Bluman-Cole and Clarkson-Kruskal, J. Math. Phys. 34 (1993) 4692-4703.
, Le
tures on Lie Groups and Systems of Ordinary and Partial Dierential Equations (Los Alamos, New Mexi
o, 1983).
, Sophus Lie, The Mathemati
al Intelligen
er 16, no. 1 (1994) 16-19.
and
, Perturbation methods in group analysis, J. Sov. Math. 55 (1991) 1450-1490.
, Similarity, Self-similarity and Intermediate Asymptoti
s (Consultants
Bureau, New York, 1979).
, Lie Symmetries of Dierential Equations: A Mathemati
a program to determine Lie symmetries. Wolfram Resear
h In
., Champaign, Illinois, MathSour
e 0202-622,
1992.
, Generalized Symmetries: A Mathemati
a Program to Determine Lie-Ba
klund
Symmetries. Wolfram Resear
h In
., Champaign, Illinois, MathSour
e 0204-680, 1993.
, Appli
ations of the generalized symmetry method, in: Modern Group Analysis: Advan
ed Analyti
al and Computational Methods in Mathemati
al Physi
s. Pro
. Int.
Workshop A
ireale, Catania, Italy, 1992, Eds.: N.H. Ibragimov, M. Torrisi and A. Valenti
(Kluwer A
ademi
Publishers, Dordre
ht, The Netherlands, 1993) 43-53.
and
, Grobner Bases (Springer Verlag, Berlin, 1993).
and
, A Survey of Lie Groups and Lie Algebras with Appli
ations and Computational Methods, Classi
s in Applied Mathemati
s 2 (SIAM, Philadelphia,
1972 & 1989).
W.F. Ames
C. Rogers
R.L. Anderson
N.H. Ibragimov
P.J. Olver
P.J. Olver
N.N. Janenko
J.M. Hill
R. Axford
N.A. Baas
N.Kh. Ibragimov
G.I. Barenblatt
G. Baumann
G. Baumann
G. Baumann
T. Be ker
V. Weispfenning
J.G. Belinfante
B. Kolman
[27
[28
[29
[30
[31
[32
[33
[34
[35
[36
[37
[38
[39
[40
and
, A Mathemati
a Program for the Cal
ulation of Lie Point
Symmetries of Systems of Dierential Equations. Preprint CRM-1822 (Centre de Re
her
hes
Mathematiques, Montreal, Canada, 1992).
, Lie-group theory for symboli
integration of rst-order dierential equations,
in: Pro
. 1979 MACSYMA User's Conferen
e, Washington, D.C., 1979, Ed.: V.E. Lewis
(MIT Press, Boston, 1979) 362-384.
, Lezioni sulla Teoria dei Gruppi Continui Finiti di Transformatizioni (Enri
i
Spoerri, Pisa, Italy, 1918).
, Simplifying the form of Lie groups admitted by a given dierential equation,
J. Math. Anal. Appl. 145 (1990) 52-62.
, Potential symmetries and linearization, in: Appli
ations of Analyti
and
Geometri
Methods to Nonlinear Dierential Equations. NATO ASI Series C: Math. and
Phys. S
s., Ed.: P.A. Clarkson (Kluwer A
ademi
Publishers, Dordre
ht, The Netherlands,
1993) 363-373.
, Potential symmetries and equivalent
onservation laws, in: Modern Group
Analysis: Advan
ed Analyti
al and Computational Methods in Mathemati
al Physi
s. Pro
.
Int. Workshop A
ireale, Catania, Italy, 1992, Eds.: N.H. Ibragimov, M. Torrisi and A. Valenti
(Kluwer A
ademi
Publishers, Dordre
ht, The Netherlands, 1993) 71-84.
, An overview of potential symmetries, in: Exploiting Symmetry in Applied
and Numeri
al Analysis, Pro
. AMS-SIAM Summer Seminar, Fort Collins, Colorado, 1992,
Eds.: E. Allgower, K. Georg and R. Miranda. Le
tures in Applied Mathemati
s 29 (Ameri
an
Mathemati
al So
iety, Providen
e, Rhode Island, 1993) 97-109.
and
, The general similarity solution of the heat equation, J. Math.
Me
h. 18 (1969) 1025-1042.
and
, Similarity Methods for Dierential Equations, Applied
Mathemati
al S
ien
es 13 (Springer Verlag, New York, 1974).
and
, Symmetries and Dierential Equations, Applied Mathemati
al S
ien
e 81 (Springer Verlag, New York, 1989).
, DEliA: A System of Exa
t Analysis of Dierential Equations using S. Lie
Approa
h. Report by Joint Venture OWIMEX Program Systems Institute of the U.S.S.R.
(A
ademy of S
ien
es, Pereslavl-Zalessky, U.S.S.R., 1989).
, DEliA: proje
t presentation, SIGSAM Bulletin 24 (1990) 37-38.
, Will DEliA grow into an expert system?, in: Design and Implementation
of Symboli
Computation Systems, Pro
. Int. Symposium, Capri, Italy, 1990, Ed.: A. Miola,
Le
ture Notes in Computer S
ien
e 429 (Springer Verlag, Berlin, 1990) 266-267.
, DEliA: A System of Exa
t Analysis of Dierential Equations using S. Lie
Approa
h. DELiA 1.5.1 User Guide (Beaver Soft Programming Team, Brooklyn, New York,
1991).
D. B
erub
e
M. de Montigny
W.A. Beyer
L. Bian hi
G.W. Bluman
G.W. Bluman
G.W. Bluman
G.W. Bluman
G.W. Bluman
G.W. Bluman
G.W. Bluman
A.V. Bo harov
A.V. Bo harov
A.V. Bo harov
A.V. Bo harov
J.D. Cole
J.D. Cole
S. Kumei
[41
[42
[43
[44
[45
[46
[47
[48
[49
[50
[51
[52
[53
[54
[55
A.V. Bo harov
M.L. Bronstein
P. Broadbridge
J.M. Hill
M. Bronstein
B. Bu hberger
B. Bu hberger
J.E. Campbell
G.J. Fee
-Ferro
G. Carra
E. Cartan
B. Champagne, W. Hereman
B. Champagne
B.W. Char
C.C. Chevalley
P. Winternitz
P. Winternitz
P.A. Griffiths
[56
[57
[58
[59
[60
[61
[62
[63
[64
[65
[66
[67
[68
[69
[70
and
, Lie-point symmetries in bifur
ation problems, Ann. Inst. Henri
Poin
are, Physique Theorique 56 (1992) 375-414.
, New exa
t solutions of the Boussinesq Equation. Euro. J. Appl. Math. 1
(1990) 279-300.
, Dimensional redu
tions and exa
t solutions of a generalized nonlinear
S
hrodinger equation, Nonlinearity 5 (1992) 453-472.
, Non
lassi
al symmetry redu
tions of nonlinear partial dierential equations,
Mathl. Comput. Modelling 18 (1993) 45-68.
, Non
lassi
al symmetry redu
tions of the Boussinesq equation. Chaos, Solitons and Fra
tals 4 (1994), in press.
and
, Non
lassi
al symmetry redu
tions and exa
t solutions of the
Zabolotskaya-Khokhlov equation, Euro. J. Appl. Math. 3 (1992) 381-415.
and
, Symmetry redu
tions of a generalized,
ylindri
al nonlinear
S
hrodinger equation, J. Phys. A: Math. Gen. 26 (1993) 133-150.
and
, New symmetry redu
tions and exa
t solutions of the Davey
Stewartson system. II. Redu
tions to partial dierential equations, Preprint No. 170 Program
in Applied Mathemati
s (University of Colorado at Boulder, Boulder, Colorado, USA, 1993).
and
, New symmetry redu
tions and exa
t solutions of the Davey
Stewartson system. I. Redu
tions to ordinary dierential equations, J. Math. Phys. 35 (1994)
255-283.
and
, New similarity solutions of the Boussinesq equation,
J. Math. Phys. 30 (1989) 2201-2213.
and
, Non
lassi
al symmetry redu
tions and exa
t solutions of nonlinear rea
tion-diusion equations, in: Appli
ations of Analyti
and Geometri
Methods to Nonlinear Dierential Equations. NATO ASI Series C: Math. and Phys. S
s., Ed.:
P.A. Clarkson (Kluwer A
ademi
Publishers, Dordre
ht, The Netherlands, 1993) 375-389.
and
, Symmetries of the nonlinear heat equation, in:
Modern Group Analysis: Advan
ed Analyti
al and Computational Methods in Mathemati
al
Physi
s. Pro
. Int. Workshop A
ireale, Catania, Italy, 1992, Eds.: N.H. Ibragimov, M. Torrisi
and A. Valenti (Kluwer A
ademi
Publishers, Dordre
ht, The Netherlands, 1993) 155-171.
and
, Symmetry redu
tions and exa
t solutions of a
lass
of nonlinear heat equations, Physi
a D 70 (1993) 250-288.
and
, On a shallow water wave equation, Nonlinearity 7
(1994) 975-1000.
and
, Redu
tions and exa
t solutions of a shallow water
wave equation, in: Pro
. 14th IMACS World Congress on Computational and Applied Mathemati
s, Atlanta, Georgia, 1994, Ed.: W.F. Ames, vol. 1 (IMACS, New Brunswi
k, New Jersey,
1994) 97-101.
G. Ci
ogna
G. Gaeta
P.A. Clarkson
P.A. Clarkson
P.A. Clarkson
P.A. Clarkson
P.A. Clarkson
S. Hood
P.A. Clarkson
S. Hood
P.A. Clarkson
S. Hood
P.A. Clarkson
S. Hood
P.A. Clarkson
M.D. Kruskal
P.A. Clarkson
E.L. Mansfield
P.A. Clarkson
E.L. Mansfield
P.A. Clarkson
E.L. Mansfield
P.A. Clarkson
E.L. Mansfield
P.A. Clarkson
E.L. Mansfield
[71
[72
[73
[74
[75
[76
[77
[78
[79
[80
[81
[82
[83
[84
[85
and
, Algorithms for the non
lassi
al method of symmetry
redu
tions, SIAM J. Appl. Math (1994) in press.
and
, Non
lassi
al symmetry redu
tions for the KadomtsevPetviashvili equation, Physi
a D 49 (1991) 257-272.
, An Introdu
tion to the Lie Theory of One-Parameter Groups (Ste
hert and Co.,
New York, 1931).
and
, Ideals, Varieties and Algorithms. An Introdu
tion
to Computational Algebrai
Geometry and Commutative Algebra. Undergraduate Texts in
Mathemati
s (Springer Verlag, Berlin, 1992).
and
, Eds. Group Theoreti
al
Methods in Physi
s. Pro
. XIX International Colloquium, Salaman
a, Spain, 1992. Anales
de Fsi
a. Monografas 1 (EDITORIAL CIEMAT, Real So
iedad Expa~nola de Fsi
a, Spain,
1993).
and
, Eds. Group Theoreti
al Methods in Physi
s. Pro
. XVIII
International Colloquium Group Theoreti
al Methods in Physi
s, Mos
ow, U.S.S.R., 1990.
Le
ture Notes in Physi
s 382 (Springer Verlag, Berlin, 1991).
, The Dierential Forms Pa
kage, available in Maple 4.2. (The Symboli
Computation Group, University of Waterloo, Ontario, Canada, 1990).
, Similarity Solutions of Nonlinear Partial Dierential Equations. Resear
h
Notes in Mathemati
s 88 (Pitman, Boston, Massa
husetts, 1983).
, Programs for
omputer implementation of the exterior
al
ulus with appli
ations to isove
tor
al
ulations, Comp. & Maths. with Appls. 6 (1980) 415-424.
, Isove
tor Methods for Equations of Balan
e (Sijtho & Noordho, Alphen
aan de Rijn, The Netherlands, 1980).
, Applied Exterior Cal
ulus (Wiley-Inters
ien
e, New York, 1985).
, Exterior
al
ulus for ve
tors and forms, The Mathemati
a Journal 1, no.
3 (1991) 92-95.
and
, Transformation Methods for Nonlinear Partial Differential Equations (World S
ienti
, Singapore, 1992).
, Continuous Groups of Transformations (Prin
eton University Press,
Prin
eton, New Jersey, 1933).
and
, A REDUCE program for determining
point and
onta
t Lie symmetries of dierential equations, Comp. Phys. Comm. 36 (1985)
383-389.
P.A. Clarkson
E.L. Mansfield
P.A. Clarkson
P. Winternitz
A. Cohen
D. Cox, J. Little
D. O'Shea
V.V. Dodonov
J. Mateos Guilarte
V.I. Man'ko
A. Donsig
L. Dresner
D.G.B. Edelen
D.G.B. Edelen
D.G.B. Edelen
D.G.B. Edelen
D.G.B. Edelen
L.P.
Jian-hua Wang
Eisenhart
V.V. Kornyak
[86
[87
[88
[89
[90
[91
[92
[93
[94
[95
[96
[97
[98
[99
P.G. Est
evez
P.G. Est
evez
P.G. Est
evez
N. Euler
P.R. Gordoa
W.-H. Steeb
R.N. Fedorova
V.V. Kornyak
R.N. Fedorova
V.V. Kornyak
R.N. Fedorova
A. Ferro
V.V. Kornyak
G. Gallo
S.P. Finikov
J.K. M Iver
A.S. Fokas
J.C. Fu hs
B. Fu hssteiner, W. Oevel
W. Wiwianka
[100
, Conditional symmetries of the equations of mathemati
al physi
s, in: Modern Group Analysis: Advan
ed Analyti
al and Computational Methods in Mathemati
al
Physi
s. Pro
. Int. Workshop A
ireale, Catania, Italy, 1992, Eds.: N.H. Ibragimov, M. Torrisi
and A. Valenti (Kluwer A
ademi
Publishers, Dordre
ht, The Netherlands, 1993) 231-239.
[101
, Symmetry in the problems of mathemati
al physi
s, in: Algebrai
Theoreti
al Studies in Mathemati
al Physi
s (Kiev Inst. of Math., Kiev, U.S.S.R., 1981)
6-28, in Russian.
[102
, Ed., Symmetry Analysis of Equations of Mathemati
al Physi
s (Inst.
Math. A
ad. Ukraine, Kiev, Ukraine, 1992).
[103
and
, Subgroup Analysis of the Galilei
and Poin
are Groups and Redu
tion of Nonlinear Equations (Naukova Dumka, Kiev, Ukraine,
1991) in Russian.
[104
and
, Computer algebra appli
ation for determining Lie
and Lie-Ba
klund symmetries of dierential equations, J. Symb. Comp. 7 (1989) 611-619.
[105
and
, Symmetries of Maxwell's Equations (D. Reidel, Dordre
ht, The Netherlands, 1987).
[106
and
, Symmetry in Equations of Quantum Me
hani
s
(Nauka, Mos
ow, U.S.S.R., 1990) in Russian.
[107
and
, Symmetry Analysis and Exa
t Solutions of Nonlinear Equations of Mathemati
al Physi
s (Naukova Dumka, Kiev, U.S.S.R, 1989)
in Russian.
[108
and
, Symmetry Analysis and Exa
t Solutions of Equations of Nonlinear Mathemati
al Physi
s, Mathemati
s and Its Appli
ations 246
(Kluwer, Dordre
ht, The Netherlands, 1993).
[109
and
, Symmetry and exa
t solutions of nonlinear spinor
equations, Phys. Rep. 172 (1989) 123-174.
[110
and
, Nonlinear Spinor Equations: Symmetry and Exa
t
Solutions, (Naukova Dumka, Kiev, Ukraine, 1992) in Russian.
[111
, Redu
tion and equivariant bran
hing lemma: dynami
al systems, evolution
PDEs, and Gauge theories, A
ta Appl. Math. 28 (1992) 43-68.
[112
, Geometri
al Symmetries of Nonlinear Equations and Physi
s (1994) in preparation.
[113
and
,
Daukl. Akad. Nauk S.S.S.R. 261 (1981) p. 1044.
[114
and
, Compatibility Analysis for Systems of Dierential
Equations on a Computer. ITPM Preprint # 20-82 (ITPM, Novosibirsk, U.S.S.R., 1982) in
Russian.
W. Fush
hi
h
W.I.
Fush hi h
W.I. Fush hi h
W.I. Fush hi h
V.V. Kornyak
W.I. Fush hi h
A.G. Nikitin
W.I.
Fush hi h
A.G.
A.F. Barannik
Nikitin
N.I. Serov
N.I. Serov
W.I. Fush hi h
R.Z. Zhdanov
W.I. Fush hi h
R.Z. Zhdanov
G. Gaeta
G. Gaeta
V.G. Ganzha
N.N. Yanenko
N.N. Janenko
[115
, Computer algebra tools for higher symmetry analysis of nonlinear evolution equations, in: Programming Environments for High-Level Problem Solving, Eds.: P.W.
Ganey and E.N. Houstis (Elsevier S
ien
e Publishers, Amsterdam, The Netherlands, 1993)
107-115.
[116
, Computer algebra, symmetry analysis and integrability of nonlinear evolution
equations, Int. J. Mod. Phys. C 4 (1993) 279-286.
[117
and
, FORMINT - A Program for Classi
ation of Integrable Nonlinear Evolution Equations, Comp. Phys. Comm. 34 (1985) 303-311.
[118
and
, Computer Algebra Appli
ation for
Classi
ation of Integrable Non-linear Evolution Equations, J. Symb. Comp. 1 (1985) 101-107.
[119
and
, Computer generation of ne
essary integrability
onditions for polynomial-nonlinear evolution systems, in: Pro
. ISSAC '90, Tokyo, Japan, 1990,
Eds.: S. Watanabe and M. Nagata (ACM Press, New York, 1990) 250-254.
[120
, Symboli
Computations in Prolongation Theory. Ph.D. Thesis, Department of Mathemati
s (Twente University of Te
hnology, Ens
hede, The Netherlands, 1981).
[121
, Lie algebra
omputations, A
ta Math. Appl. 16 (1989) 231-242.
[122
, Prolongation algebras of nonlinear PDEs, in: Pro
. ERCIM Advan
ed
Course on Partial Dierential Equations and Group Theory, Part II, Bonn, 1992, Ed.: J.F.
Pommaret (Gesells
haft fur Mathematik und Datenverarbeitung, Sankt Augustin, Germany,
1992) 1-20.
[123
and
, Implementation of dierential geometry obje
ts
and fun
tions with an appli
ation to extended Maxwell equations, in: Pro
. EUROCAM '82,
Marseille, Fran
e, 1982, Ed.: J. Calmet. Le
ture Notes in Computer S
ien
e 144 (Springer
Verlag, New York, 1982) 181-187.
[124
and
, Graded dierential geometry in REDUCE: Supersymmetri
stru
tures of the modied KdV equation, A
ta Appl. Math. 24 (1991) 211-231.
[125
and
, Symboli
omputations in applied
dierential geometry, A
ta Appl. Math. 1 (1983) 43-77.
[126
and
, Lie subgroups of the symmetry group of the equations
des
ribing a nonstationary and isotropi
ow I. Subgroupes de Lie des groupes de symmetry
pour les equations de la dynamique de
uid I, Can. J. Phys. (1994) in press.
[127
and
, Lie subgroups of symmetry groups of the
uid dynami
s and the magnetohydrodynami
s equations II, Can. J. Phys. (1994) in press.
[128
and
, Hidden symmetries of energy-
onserving dierential
equations, IMA J. Appl. Math. 51 (1993) 147-153.
V.P. Gerdt
V.P. Gerdt
A.Yu. Zharkov
V.P. Gerdt
A.Yu. Zharkov
A.Yu. Zharkov
P.K.H. Gragert
P.K.H. Gragert
P.K.H. Gragert
P.K.H. Gragert
P.H.M. Kersten
P.K.H. Gragert
P.H.M. Kersten
A.M. Grundland
A.M. Grundland
A. Guo
L. Lalague
L. Lalague
B. Abraham-Shrauner
A. Martini
[129
and
, Symmetries and Conservation Laws of Kadomtsev-Pogutse Equations, A
ta
Appl. Math. 15 (1989) 23-64.
[130
, Similarity Analyses of Boundary Value Problems in Engineering (Prenti
e
Hall, Englewood Clis, 1964).
[131
and
, Geometri
approa
h to invarian
e groups and solution of partial dierential equations, J. Math. Phys. 12 (1971) 653-666.
[132
and
, A
onstru
tive implementation of the Cartan-Kahler
theory of exterior dierential systems, J. Symb. Comp. 12 (1991) 655-667.
[133
and
, Constrained dynami
s and exterior differential systems, J. Phys. A 24 (1991) 5253-5265.
[134
, The birth of Lie's theory of groups, The Mathemati
al Intelligen
er 16, no. 2
(1994) 6-17.
[135
, LIE: A PC Program for Lie Analysis of Dierential Equations, Comp. Phys.
Comm. 77 (1993) 241-248.
[136
Grobner Basis Pa
kage, Wolfram Resear
h In
., Champaign, Illinois, MathSour
e 0206-794, 1994.
[137
, SYMMGRP.MAX: A symboli
program for symmetry analysis of partial
dierential equations, in: Exploiting Symmetry in Applied and Numeri
al Analysis, Pro
.
AMS-SIAM Summer Seminar, Fort Collins, Colorado, 1992, Eds.: E. Allgower, K. Georg and
R. Miranda. Le
tures in Applied Mathemati
s 29 (Ameri
an Mathemati
al So
iety, Providen
e, Rhode Island, 1993) 241-257.
[138
, Review of symboli
software for the
omputation of Lie symmetries of differential equations, Euromath Bulletin 1, no. 2 (1994) 45-82.
[139
and
, Lie point symmetries of
lassi
al
eld theories, in: Group Theoreti
al Methods in Physi
s, Pro
. XIX International Colloquium,
Salaman
a, Spain, 1992, Eds.: M.A del Olmo, M. Santander, and J. Mateos Guilarte. Anales
de Fsi
a. Monografas 1 (CIEMAT, Real So
iedad Expa~nola de Fsi
a, Spain, 1993) 402-405.
[140
, Dierential Geometry and the Cal
ulus of Variations (A
ademi
Press, New
York, 1968).
[141
, Computer Assisted Determination of Lie Point Symmetries with Appli
ation to
Fluid Dynami
s. Ph.D Thesis, Program in Applied Mathemati
s (The University of Colorado,
Boulder, Colorado, 1994).
[142
, The Use of Maple in the Sear
h for Symmetries, Resear
h Report no. 77,
Department of Mathemati
s (University of Canterbury, Christ
hur
h, New Zealand, 1993).
V.N. Gusyatnikova, A.V. Samokhin, V.S. Titov, A.M. Vinogradov
Yumaguzhin
A.G. Hansen
B.K. Harrison
D. Hartley
F.B. Estabrook
R.W. Tu ker
P.A. Tu key
T. Hawkins
A.K. Head
G. Helzer
W. Hereman
W. Hereman
R. Hermann
S. Herod
M. Hi kman
M. Grundland
V.A.
[143
, The use of Maple in the sear
h for symmetries, in: Pro
. 14th IMACS World
Congress on Computational and Applied Mathemati
s, Atlanta, Georgia, 1994, Ed.: W.F.
Ames, vol. 1 (IMACS, New Brunswi
k, New Jersey, 1994) 226-229.
[144
, Solution of Dierential Equations by Means of One-Parameter Groups, Resear
h
Notes in Mathemati
s 63 (Pitman, Boston, Massa
husetts, 1982).
[145
, Dierential Equations and Group Methods for S
ientists and Engineers (CRC
Press, Bo
a Raton, Florida, 1992).
[146
, Non
lassi
al Symmetry Redu
tions and Exa
t Solutions of Nonlinear Partial Differential Equations. Ph.D Thesis, Department of Mathemati
s (University of Exeter, Exeter,
United Kingdom, 1993).
[147
, Transformation Groups Applied to Mathemati
al Physi
s (D. Reidel,
Dordre
ht, The Netherlands, 1985).
[148
, Primer on Group Analysis (Znanie, Mos
ow, U.S.S.R., 1989) in Russian.
[149
, Essay on Group Analysis of Ordinary Dierential Equations (Znanie,
Mos
ow, U.S.S.R., 1991) in Russian.
[150
, Group analysis of ordinary dierential equations and the invarian
e prin
iple in mathemati
al physi
s (for the 150th anniversary of Sophus Lie, Russian Math. Surveys
47 (1992) 89-156.
[151
, Linearization of ordinary dierential equations, in: Dierential Equations
with Appli
ations to Mathemati
al Physi
s. Eds.: W.F. Ames, E.M. Harrell II and J.V. Herod.
Mathemati
s in S
ien
e and Engineering 192 (A
ademi
Press, New York, 1993) 177-188.
[152
, Sophus Lie and harmony in mathemati
al physi
s, on the 150th anniversary of his birth, The Mathemati
al Intelligen
er 16, no. 1 (1994) 20-28.
[153
, Ed., CRC Handbook of Lie Group Analysis of Dierential Equations.
Volume 1: Symmetries, Exa
t Solutions, and Conservation Laws (CRC Press, Bo
a Raton,
Florida, 1994).
[154
, Ed., CRC Handbook of Lie Group Analysis of Dierential Equations.
Volume 2: Appli
ations in Engineering and Physi
al S
ien
es (CRC Press, Bo
a Raton,
Florida, 1994) in press.
[155
and
, Eds., Modern Group Analysis: Theory and Appli
ation in Mathemati
al Modelling. Pro
eedings 5th Int. Conf. on Group Analysis and Dierential Equations, University of the Witwatersrand, Johannesburg, South Afri
a, January 16-22,
1994. Will appear in the new international journal on `Lie Groups and their Appli
ations' 1
(1994) no. 1.
[156
, A REDUCE program for nding symmetries of nonlinear evolution equations with
uniform rank, Comp. Phys. Comm. 42 (1986) 351-357.
M. Hi
kman
J.M. Hill
J.M. Hill
S. Hood
N.H. Ibragimov
N.H. Ibragimov
N.H. Ibragimov
N.H. Ibragimov
N.H. Ibragimov
N.H. Ibragimov
N.H. Ibragimov
N.H. Ibragimov
N.H. Ibragimov
M. Ito
F.M. Mahomed
[157
[158
and
A REDUCE program for nding
onserved densities of partial differential equations with uniform rank, Comp. Phys. Comm. 38 (1985) 415-419.
M. Ito
F. Kako
[159
, Les systemes d'equations aus derivees partielles, J. Math. Pure et Appl. 3 (1920)
65-151.
[160
, Le
ons sur les Systemes d'Equations aux Derivees (Gauthier-Villars, Paris,
1929); Equations aux Derivees Partielles (CNRS, Paris, 1956).
[161
, Einfuhrung in die Theorie der Systeme von Dierentialglei
hungen (B.G. Teubner, Leipzig, 1934).
[162
, Separation of Variables for Riemannian Spa
es of Constant Curvature
(Longman, Essex, 1986).
[163
, On the equations des
ribing modulational instabilities due to nonlinear
intera
tion between high and low frequen
y waves, Phys. Lett. A 136 (1989) 216-220.
[164
, Innitesimal Symmetries: A Computational Approa
h. Ph.D. Thesis.
Department of Mathemati
s. (Twente University of Te
hnology, Ens
hede, The Netherlands,
1985); CWI Tra
t 34 (Center for Mathemati
s and Computer S
ien
e, Amsterdam, 1987).
[165
, Software to Compute Innitesimal Symmetries of Exterior Dierential
Systems, with Appli
ations, A
ta Appl. Math. 16 (1989) 207-229.
[166
and
, Eds. Algebrai
and Geometri
Stru
tures
in Dierential Equations, Pro
. Meeting, Twente University of Te
hnology, Ens
hede, The
Netherlands, 1993, to appear as spe
ial issue of A
ta Appl. Math. (1994).
[167
, Dierential Algebra and Algebrai
Groups (A
ademi
Press, New York,
1973).
[168
, Dierential algebrai
groups (A
ademi
Press, New York, 1984).
[169
and
, Symmetries and Integrals. Preprint 15 (International
Sophus Lie Center & Matematisk Institutt, Universitetet i Oslo, Norway, 1993).
[170
and
, A program for symmetry analysis of dierential
equations, in: Pro
. ISSAC '91, Bonn, Germany, 1991, Ed.: S.M. Watt (ACM Press, New
York, 1991) 315-316.
[171
, Graded manifolds, graded Lie theory and prequantization, in: Pro
. Symposium on Dierential Geometri
al Methods in Mathemati
al Physi
s, Bonn, Germany, 1975,
Eds.: K. Bleuler and A. Reetz. Le
ture Notes in Mathemati
s 570 (Springer Verlag, New
York, 1977) 177-306.
[172
and
, Geometry of Jet Spa
es
and Nonlinear Partial Dierential Equations. Adv. Stud. Contemp. Math. 1 (Gordon &
Brea
h, New York, 1986).
M. Janet
M. Janet
hler
E. Ka
E.G. Kalnins
V.I. Karpman
P.H.M. Kersten
P.H.M. Kersten
P.H.M. Kersten
B. Komrakov
V.V. Kornyak
V. Ly hagin
W.I. Fush hi h
B. Kostant
A.M. Vinogradov
[173
and
A.M. Vinogradov
161-209.
[174
, Le
tures on Involutive Systems of Partial Dierential Equations (Publ. So
.
Math., Sao Paulo, Brasil, 1967).
[175
and
, Eds., Symmetries and Nonlinear Phenomena (World S
ienti
, Singapore, 1988).
[176
and
, Non-
lassi
al symmetry redu
tions: example of the Boussinesq equation, J. Phys. A: Math. Gen. 22 (1989) 2915-2924.
[177
and
, Symmetries and
onditional symmetries of dierentialdieren
e equations, J. Math. Phys. 34 (1993) 3713-3730.
[178
and
Group Methods in Integrating Nonlinear Dynami
al
Systems (Nauka, Mos
ow, U.S.S.R., 1985) in Russian.
[179
, Sophus Lie's 1880 Transformation Group Paper. Translated by M. A
kerman,
omments by R. Hermann, Lie Groups: History, Frontiers on Appli
ations 1 (Mathemati
al
S
ien
es Press, Brookline, Massa
husetts, 1975).
[180
, Sophus Lie's 1884 Dierential Invariants Paper. Translated by M. A
kerman,
omments by R. Hermann, Lie Groups: History, Frontiers on Appli
ations 3 (Mathemati
al
S
ien
es Press, Brookline, Massa
husetts, 1976).
[181
, Theorie der Transformationsgruppen. Volumes I, II & III (B.G. Teubner, Leipzig,
1888, 1890, 1893).
[182
, Vorlesungen uber Dierentialglei
hungen mit Bekannten Innitesimalen Transformationen (B.G. Teubner, Leipzig, 1891, 1912; Chelsea Publishing Company, New York, 1967).
[183
, Gesammelte Abhandlungen. Vol. 1-7 (B.G. Teubner, Leipzig, 1922-1929).
[184
, Vorlesungen uber Continuierli
he Gruppen, Abteilung III (Chelsea Publishing Company, New York, 1971).
[185
and
, Theorie der Transformationsgruppen II (B.G. Teubner, Leipzig,
1890).
[186
, Equivalen
e Transformations for Classes of Dierential Equations. Ph.D. Thesis. Department of Mathemati
s (University of British Columbia, Van
ouver, Canada, 1992).
[187
, An ordinary dierential equation solver for REDUCE, in: Pro
. ISSAC '88, Rome, Italy, 1988, Ed.: P. Gianni. Le
ture Notes in Computer S
ien
e 358 (Springer
Verlag, Berlin, 1988) 196-205.
[188
, Dierential Grobner Bases, Ph.D Thesis, Department of Computer S
ien
e (University of Sydney, Sydney, Australia, 1991).
M. Kuranishi
D. Levi
P. Winternitz
D. Levi
P. Winternitz
D. Levi
P. Winternitz
A.N. Leznov
M.V. Savel
ev,
S. Lie
S. Lie
S. Lie
S. Lie
S. Lie
S. Lie
S. Lie
F. Engel
I.G. Lisle
M.A.H. Ma Callum
E.L. Mansfield
[189
, DIFFGROB2: A Symboli
Algebra Pa
kage for Analyzing Systems of
PDE using Maple, User's Manual for Release 2, Preprint, Department of Mathemati
s (University of Exeter, Exeter, United Kingdom, 1993).
[190
and
, Appli
ations of the dierential algebra pa
kage
Digrob2 to redu
tions of PDE, in: Pro
. 14th IMACS World Congress on Computational
and Applied Mathemati
s, Atlanta, Georgia, 1994, Ed.: W.F. Ames, vol. 1 (IMACS, New
Brunswi
k, New Jersey, 1994) 336-339.
[191
and
, Appli
ations of the Dierential Algebra Pa
kage
Digrob2 to Classi
al Symmetries of Dierential Equations, Preprint M94/18, Department of
Mathemati
s (University of Exeter, Exeter, UK, 1994).
[192
and
, Dierential Grobner Bases, Preprint 92-10, S
hool
of Mathemati
s, Physi
s, Computer S
ien
e, and Ele
troni
s (Ma
quarie University, Sydney,
Australia, 1992).
[193
, Ed., Geometri
Approa
hes to Dierential Equations, Le
ture Notes in Mathemati
s 810 (Springer Verlag, New York, 1979).
[194
and
, GROEBNER: A Grobner basis pa
kage, in:
REDUCE User's Manual, Version 3.4, Ed.: A.C. Hearn (Rand Corporation, Santa Moni
a,
California, 1991).
[195
and
, The symmetry approa
h to
lassi
ation of integrable equations, in: What is Integrability?, Springer Le
ture Notes in Nonlinear
Dynami
s, Ed.: V.I. Zakharov (Springer Verlag, New York, 1990) 115-184.
[196
, Lie Theory and Spe
ial Fun
tions (A
ademi
Press, New York, 1968).
[197
, Symmetry Groups and Their Appli
ations, Pure and Applied Mathemati
s
50 (A
ademi
Press, New York, 1972).
[198
, Symmetry and Separation of Variables (Addison-Wesley, Reading, Massa
husetts, 1977).
[199
, Ed., Ba
klund Transformations, the Inverse S
attering Method, Solitons, and
Their Appli
ations. Le
ture Notes in Mathemati
s 515 (Springer Verlag, New York, 1976).
[200
, Invariante Variationsprobleme, Na
hr. Konig. Gesell. Wissen. Gottingen,
Math.-Phys. Kl. (1918) 235-257; English translation: Transport Theory and Stat. Phys. 1
(1971) 186-207.
[201
, Group analysis of MHD equations, Atti. Sem. Mat. Fis. Univ. Modena 33
(1984) 21-34.
[202
, Intera
tive REDUCE programs for
al
ulating,
lassi
al, non-
lassi
al, and
Lie-Ba
klund symmetries of dierential equations, Manual of the Program, Preprint GT
Math:062090-051, S
hool of Mathemati
s (Georgia Institute of Te
hnology, Atlanta, Georgia, 1990).
E.L. Mansfield
E.L. Mansfield
E.L. Mansfield
E.L. Mansfield
P.A. Clarkson
P.A. Clarkson
E.D. Fa kerell
R. Martini
ller
H. Melenk, H.M. Mo
W. Miller, Jr.
W. Miller, Jr.
W.
Miller
R.M. Miura
E. Noether
M.C. Nu i
M.C. Nu i
W. Neun
V.V. Sokolov
[203
, Intera
tive REDUCE programs for
al
ulating,
lassi
al, non-
lassi
al, and
approximate symmetries of dierential equations, in: Pro
. 13th IMACS World Congress,
Dublin, Ireland, 1991. Eds.: R. Vi
hnevetsky and J.J.H. Miller (Criterion Press, Dublin,
Ireland, 1991) 349-350.
[204
, Intera
tive REDUCE programs for
al
ulating,
lassi
al, non-
lassi
al, and
Lie-Ba
klund symmetries of dierential equations, in: Computational and Applied Mathemati
s, II: Dierential Equations. Sele
ted and revised papers from the IMACS 13th World
Congress, Dublin, Ireland, 1991. Eds: W.F. Ames and P.J. Van der Houwen (Elsevier S
ien
e
Publishers, Amsterdam, The Netherlands, 1992) 345-350.
[205
, Symmetries and symboli
omputation, in: Dierential Equations with Appli
ations to Mathemati
al Physi
s. Mathemati
s in S
ien
e and Engineering 192. Eds.: W.F.
Ames, E.M. Harrell II and J.V. Herod (A
ademi
Press, New York, 1993) 249-260.
[206
, Non
lassi
al symmetries and Ba
klund transformations, J. Math. Anal. Appl.
178 (1993) 294-300.
[207
, Iterating the Non
lassi
al Symmetry Method, Te
h. Rep. 1993-14, Departimento de Matemati
a (Universita di Perugia, Perugia, Italy, 1993).
[208
, Intera
tive REDUCE programs for
al
ulating Lie-point, non-
lassi
al, LieBa
klund, and approximate symmetries of dierential equations: Manual and
oppy disk, in:
CRC Handbook of Lie Group Analysis of Dierential Equations. Volume 3: New Trends (CRC
Press, Bo
a Raton, Florida, 1995), Chapter 14.
[209
and
, Classi
al and non
lassi
al symmetries for the Helmholtz
equation, J. Math. Anal. Appl. 178 (1993) 584-591.
[210
and
, The non
lassi
al method is more general than the dire
t
method for symmetry redu
tions. An example of the Fitzhugh-Nagumo equation, Phys. Lett.
A 164 (1992) 49-56.
[211
, Algorithms for nding the stru
ture of solutions of a system of linear partial
dierential equations, in: Pro
. ISSAC '94, Oxford, UK, 1994, Eds.: J. von zur Gathen and
M. Giesbre
ht (ACM Press, New York, 1994) 216-223.
[212
, Standard bases of dierential ideals, in: Applied Algebra, Algebrai
Algorithms, and Error-Corre
ting Codes. Pro
. 8th Int. Conf. AAECC-8, Eds.: L. Huguet and A.
Poli. Le
ture Notes in Computer S
ien
e 508 (Springer Verlag, Berlin, 1991) 304-321.
[213
, Symmetry and expli
it solutions of partial dierential equations, Appl. Num.
Math. 10 (1992) 307-324.
[214
, Appli
ations of Lie Groups to Dierential Equations. Graduate Texts in Mathemati
s 107 (Springer Verlag, New York, 1986 & 1993{2nd edition).
[215
, Dire
t redu
tion and dierential
onstraints, Pro
. R. So
. London A. 444
(1994) 509-523.
M.C. Nu
i
M.C. Nu i
M.C. Nu i
M.C. Nu i
M.C. Nu i
M.C. Nu i
M.C. Nu i
M.C. Nu i
T. Oaku
F. Ollivier
P.J. Olver
P.J. Olver
P.J. Olver
W.F. Ames
P.A. Clarkson
[216
, Equivalen
e, Invariants and Symmetry. (Cambridge University Press, Cambridge, UK, 1994)
[217
and
, The
onstru
tion of spe
ial solutions to partial dierential
equations, Phys. Lett. A 114 (1986) 107-112.
[218
and
, Group-invariant solutions of dierential equations, SIAM J.
Appl. Math. 47 (1987) 263-278.
[219
, Partially Invariant Solutions of Dierential Equations. Ph.D Thesis, Department of Mathemati
s (University of Minnesota, Minneapolis, Minnesota, 1989).
[220
, Group Properties of Dierential Equations (A
ademy of S
ien
es,
Siberian Bran
h, U.S.S.R., Novosibirsk, 1962); English translation by G.W. Bluman (1967)
unpublished.
[221
, Group Analysis of Dierential Equations (Nauka, Mos
ow, U.S.S.R.,
1972); English translation: W.F. Ames Ed. (A
ademi
Press, New York, 1982).
[222
, A Global Formulation of the Lie Theory of Transformation Groups, Memoirs of
the Ameri
an Mathemati
al So
iety 22 (Ameri
an Mathemati
al So
iety, Providen
e, Rhode
Island, 1957 & 1989).
[223
, Computations in dierential and dieren
e modules, A
ta Appl. Math.
16 (1989) 167-189.
[224
and
, Lo
al Jet Bundle Formulation
of Ba
klund Transformations. Mathemati
al Physi
s Studies 1 (D. Reidel, Dordre
ht, The
Netherlands, 1979).
[225
and
, Automati
determination of dynami
al symmetries of ordinary
dierential equations, Computing 41 (1989) 297-307.
[226
, Sophus Lie (Nauka, Leningrad, 1983).
[227
, Systems of Partial Dierential Equations and Lie Pseudogroups (Gordon
& Brea
h, New York, 1978); translated by A.M. Vinogradov for MIR (1983).
[228
, Dierential Galois Theory (Gordon & Brea
h, New York, 1983).
[229
, Lie Pseudogroups and Me
hani
s (Gordon & Brea
h, New York, 1988).
[230
, Le
ture Notes of `Advan
ed Course on Partial Dierential Equations and
Group Theory'. New Perspe
tives in Computer Algebra and Appli
ations to Engineering S
ien
es. Pro
. ERCIM Advan
ed Course, Gesells
haft fur Mathematik und Datenverarbeitung,
Bonn, Germany, April 6-10, 1992.
[231
, Automized
omputing of dening equations for a Lie group, Izvestiya AN
Belor. SSR, Ser. Phys.-Math. 2 (1985) 33-37, in Russian
[232
, Similarity redu
tions of partial dierential equations, J. Phys. A: Math. Gen. 25
(1992) 2631-2640.
P.J. Olver
P.J. Olver
P.J. Olver
P. Rosenau
P. Rosenau
J.R. Ondi h
L.V. Ovsiannikov
L.V. Ovsiannikov
R.S. Palais
E.V. Pankrat'ev
H.J. Pohle
T. Wolf
J.F. Pommaret
J.F. Pommaret
J.F. Pommaret
M.D. Popov
E. Pu i
W.F. Shadwi k
[233
G.J. Reid
G.J. Reid
G.J. Reid
G.J. Reid
A. Boulton
G.J. Reid
D.K. M Kinnon
G.J. Reid
A.D. Wittkopf
A.D. Wittkopf
C. Riquier
J.F. Ritt
C. Rogers
C. Rogers
P. Rosenau
A.D. Wittkopf
W.F. Ames
W.A. Shadwi k
J.L. S hwarzmeier
[247
and
Systems of Quasilinear Equations and Their
Appli
ations to Gas Dynami
s, 2nd Edition, Nauka, Mos
ow, 1978; Translations of Mathemati
al Monographs 55 (AMS, Providen
e, Rhode Island, 1983).
[248
and
, Eds., Lie Methods in Opti
s, Pro
. CIFMOCIO Workshop, Leon, Guanajuato, Mexi
o, 1985. Le
ture Notes in Physi
s 250 (Springer
Verlag, New York, 1986).
[249
and
, Generalizations of Noether's theorem in
lassi
al me
hani
s,
SIAM Review 23 (1981) 467-494.
[250
and
, REDUCE-pro
edures for the study of adjoint symmetries of se
ond-order dierential equations, J. Sym. Comp. 13 (1992) 683-693.
[251
, Group analysis of some partial dierential equations arising in appli
ations,
Contemp. Math. 54 (1986) 35-44.
[252
, Group Theoreti
Methods in Bifur
ation Theory, Le
ture Notes in Mathemati
s 762 (Springer Verlag, New York, 1979).
[253
and
, Lie Groups and Algebras with Appli
ations to Physi
s,
Geometry, and Me
hani
s, Applied Mathemati
al S
ien
es 61 (Springer Verlag, New York,
1986).
[254
, EXCALC, a system for doing
al
ulations in the
al
ulus of modern dierential geometry{User's Manual, distributed with REDUCE 3.3 (1987).
[255
, EXCALC - A pa
kage for
al
ulations in modern dierential geometry, in:
Computer Algebra in Physi
al Resear
h. Pro
. IV Int. Conf. Dubna, U.S.S.R., 1990, Eds.:
D.V. Shirkov, V.A. Rostovtsev and V.P. Gerdt (World S
ienti
, Singapore, 1990) 71-80.
[256
and
, Algorithmi
methods for Lie pseudogroups, in:
Modern Group Analysis: Advan
ed Analyti
al and Computational Methods in Mathemati
al
Physi
s. Pro
. Int. Workshop A
ireale, Catania, Italy, 1992, Eds.: N.H. Ibragimov, M. Torrisi
and A. Valenti (Kluwer A
ademi
Publishers, Dordre
ht, The Netherlands, 1993) 337-344.
[257
, A REDUCE pa
kage for determining Lie symmetries of ordinary and partial
dierential equations, Comp. Phys. Comm. 27 (1982) 179-186.
[258
, Automati
ally determining symmetries of ordinary dierential equations, in:
Pro
. EUROCAL '83, London, England, 1983, Ed.: J.A. van Hulzen. Le
ture Notes in Computer S
ien
e 162 (Springer Verlag, New York, 1983) 45-54.
[259
, The Riquier-Janet theory and its appli
ation to non-linear evolution equations,
Physi
a D 11 (1984) 243-251.
[260
, Automati
ally determining symmetries of partial dierential equations, Computing 34 (1985) 91-106; Addendum: Computing 36 (1986) 279-280.
B.L. Ro
zdestvensk
i
N.N. Janenko
n
hez Mondrago
n
J. Sa
K.B. Wolf
W. Sarlet
F. Cantrijn
W. Sarlet
J. Vanden Bonne
C.C.A. Sastri
D.H. Sattinger
D.H. Sattinger
O.L. Weaver
fer
E. S
hru
fer
E. S
hru
, W.M. Seiler
J. S
hu
F. S hwarz
F. S hwarz
F. S hwarz
F. S hwarz
J. Calmet
[261
F. S hwarz
F. S hwarz
F. S hwarz
F. S hwarz
F. S hwarz
F. S hwarz
F. S hwarz
F. S hwarz
F. S hwarz
J.F. Pommaret
J.L. S hwarzmeier
L.I. Sedov
P. Rosenau
[273
W.M. Seiler
R. Sesahdri
T.Y. Na
W.A. Shadwi k
J. Sherring
J. Sherring
J. Sherring
M.F. Singer
W.-H.
Steeb
S. Steinberg
S. Steinberg
S. Steinberg
H. Stephani
J. Thomas
V.L. Topunov
A.M. Tresse
A.M. Tresse
G. Prin e
[290
T. Tsujishita
311-363.
[291
, Formal geometry of systems of dierential equations, in: Sugaku Expositions
3 (AMS, Providen
e, Rhode Island, 1990) 25-73; translation of Sugaku 35 (1983) 332-357, in
Japanese.
[292
, Di. Geom. Appl. 1 (1991) 3-34.
[293
, Computer algebrai
determination of symmetries and
onservation laws of
PDEs, in: Pro
. ISMM Int. Symposium on Computer Appli
ations in Design, Simulation and
Analysis, New Orleans, Louisiana, 1990, Ed.: E.K. Park (A
ta Press, Anaheim, California,
1990) 310-312.
[294
, SYMCON: A MACSYMA Pa
kage for the Determination of Symmetries and
Conservation Laws of PDEs. Preprint, Department of Engineering S
ien
es (Trinity University, San Antonio, Texas, 1990)
[295
, PDELIE: A partial dierential equation solver, MACSYMA Newsletter 9,
no. 1 (1992) 1-13.
[296
, PDELIE: A partial dierential equation solver II, MACSYMA Newsletter 9,
no. 2-4 (1992) 5-20.
[297
, PDELIE: A partial dierential equation solver III, MACSYMA Newsletter
11 (1994) to appear.
[298
, PDELIE: Symboli
Software for the Analysis of Partial Dierential Equations
by Lie Group Methods, User's Manual, Preprint, Department of Engineering S
ien
es (Trinity
University, San Antonio, Texas, USA, 1994).
[299
, Lo
al Bifur
ation and Symmetry (Pitman, London, 1982).
[300
, LiE, A software pa
kage for Lie group
omputations, Euromath
Bulletin 1, no. 2 (1994) 83-94.
[301
and
, Eds. Symmetries and Integrability of Dieren
e
Equations, Pro
. Workshop, Esterel, Quebe
, Canada, May 22-29, 1994. CRM Pro
eedings
and Le
ture Notes (Ameri
an Mathemati
al So
iety, Providen
e, Rhode Island, 1994).
[302
, Lo
al symmetries and
onservation laws, A
ta Appl. Math. 2 (1984)
21-78.
[303
, Symmetries and
onservation laws of partial dierential equations:
Basi
notations and results, A
ta Appl. Math. 16 (1989) 3-21.
[304
, Ed., Symmetries of Partial Dierential Equations, Part I, A
ta Appl.
Math. 15, nos. 1 & 2 (1989); Part II, ibid. 16, no. 1 (1989); Part III, ibid. 16, no. 2 (1989).
[305
, Ed., Symmetries of Partial Dierential Equations. Conservation laws
- Appli
ations - Algorithms (Kluwer A
ademi
Publi
ations, Dordre
ht, The Netherlands,
1989).
T. Tsujishita
T. Tsujishita
P. Vafeades
P. Vafeades
P. Vafeades
P. Vafeades
P. Vafeades
P. Vafeades
A. Vanderbauwhede
L. Vinet, D. Levi
A.M. Vinogradov
A.M. Vinogradov
A.M. Vinogradov
A.M. Vinogradov
P. Winternitz
[306
and
, Nonlo
al symmetries and the theory of
overings, A
ta Appl. Math. 2 (1984) 79-96.
[307
, Redu
tion and quotient equations for dierential equations with symmetries, A
ta Appl. Math. 23 (1991) 1-24.
[308
, Symmetries of
ompatibility
onditions for systems of dierential equations, A
ta Appl. Math. 26 (1992) 61-86.
[309
, Dierential form analysis using MACSYMA, in: Pro
. 1977 MACSYMA
User's Conferen
e, Berkeley, California, 1977, Ed.: E.M. Andersen (NASA CP-2012, Washington, D.C., 1977) 71-74.
[310
, Foundations of Dierential Manifolds and Lie Groups (Springer Verlag,
New York, 1983).
[311
, Similarity Methods with Referen
e to a High-order Nonlinear Diusion
Equation. Ph.D Thesis, Mathemati
al Institute (University of Oxford, Oxford, United Kingdom, 1989).
[312
, Lie groups and solutions of nonlinear dierential equations, in: Nonlinear
Phenomena, Ed.: K.B. Wolf. Le
ture Notes in Physi
s 189 (Springer Verlag, New York, 1983)
263-331.
[313
, Group theory and exa
t solutions of partially integrable dierential systems,
in: Partially Integrable Nonlinear Evolution Equations and their Physi
al Appli
ations. NATO
ASI Series C: Math. and Phys. S
s. 310, Eds.: R. Conte and N. Bo
ara (Kluwer A
ademi
Publishers, Dordre
ht, The Netherlands, 1990) 515-567.
[314
, Conditional symmetries and
onditional integrability for nonlinear systems,
in: Group Theoreti
al Methods in Physi
s. Pro
. XVIII International Colloquium, Mos
ow,
U.S.S.R., 1990, Eds.: V.V. Dodonov and V.I. Man'ko. Le
ture Notes in Physi
s 382 (Springer
Verlag, Berlin, 1991) 263-271.
[315
, Lie group and solutions of nonlinear partial dierential equations, in: Integrable Systems, Quantum Groups, and Quantum Field Theories. NATO ASI Series C: Math.
and Phys. S
s. 409, Eds.: L.A. Ibort and M.A. Rodriguez (Kluwer A
ademi
Publishers,
Dordre
ht, The Netherlands, 1993) 429-495.
[316
, Ed., Lie Methods in Opti
s II, Pro
. CIFMA-CIO Workshop, Co
oyo
, Mexi
o,
1988. Le
ture Notes in Physi
s 352 (Springer Verlag, New York, 1990).
[317
, An analyti
algorithm for de
oupling and integrating systems of nonlinear partial
dierential equations, J. Comp. Phys. 60 (1985) 437-446.
[318
, Zur analytis
hen Untersu
hing und exakten Losung von Dierentialglei
hungen
mit Computeralgebrasystemen. Dissertation B (Friedri
h S
hiller Universitat, Jena, Germany,
1989).
A.M. Vinogradov
E.M. Vorob'ev
E.M. Vorob'ev
H.D. Wahlquist
F.W. Warner
S.M. Williams
P. Winternitz
P. Winternitz
P. Winternitz
P. Winternitz
K.B. Wolf
T. Wolf
T. Wolf
[319
, A pa
kage for the analyti
investigation and exa
t solutions of dierential equations, in: Pro
. EUROCAL '87, Leipzig, GDR, 1987, Ed.: J.H. Davenport, Le
ture Notes in
Computer S
ien
e 378 (Springer Verlag, Berlin, 1989) 479-491.
[320
, The symboli
integration of exa
t PDEs, Preprint S
hool of Mathemati
al S
ien
es, Queen Mary and Westeld College (University of London, London, 1992).
[321
, An e
ien
y improved program LIEPDE for determining Lie-symmetries of
PDEs, in: Modern Group Analysis: Advan
ed Analyti
al and Computational Methods in
Mathemati
al Physi
s. Pro
. Int. Workshop A
ireale, Catania, Italy, 1992, Eds.: N.H. Ibragimov, M. Torrisi, and A. Valenti (Kluwer A
ademi
Publishers, Dordre
ht, The Netherlands,
1993) 377-385.
[322
, A Program for Applying Symmetries of PDEs, Presentation at CATHODE Workshop on Computer Algebra Tools for Handling ODEs, London, UK, 21-24 Mar
h 1994.
[323
and
, The
omputer algebra pa
kage CRACK for investigating PDEs,
in: Pro
. ERCIM Advan
ed Course on Partial Dierential Equations and Group Theory, Bonn,
1992, Ed.: J.F. Pommaret (Gesells
haft fur Mathematik und Datenverarbeitung, Sankt Augustin, Germany, 1992); Also: Manual for CRACK added to the REDUCE Network Library,
S
hool of Mathemati
al S
ien
es, Queen Mary and Westeld College (University of London,
London, 1992) 1-19.
[324
and
, Heuristi
s for overdetermined systems of PDEs, in: Pro
. Meeting on Arti
ial Intelligen
e in Mathemati
s, Glasgow, United Kingdom, 1991, Eds.: J.H.
Johnson, S. M
Kee and A. Vella (Clarendon Press, Oxford, 1993) in press.
[325
and
, Investigating symmetries and other analyti
al properties of ODEs
with the
omputer algebra pa
kage CRACK, Preprint S
hool of Mathemati
al S
ien
es, Queen
Mary and Westeld College (University of London, London, 1994).
[326
, Felix Klein and Sophus Lie (Birkhauser, Basel, 1988).
[327
and
, Symmetries,
onserved quantities,
and hierar
hies for some latti
e systems with soliton stru
ture, J. Math. Phys. 32 (1991)
1908-1918.
[328
, Le
ture Notes on Geometri
al Aspe
ts of Partial Dierential Equations.
Series on Soviet and East European Mathemati
s 9 (World S
ienti
, Singapore, 1992).
[329
, Computer
lassi
ation of the integrable
oupled KdV-like systems with
unit main matrix, J. Symb. Comp. 15 (1993) 85-90.
[330
, Algebrai
programming and dierential forms, The Mathemati
a Journal 2, no.
1 (1992) 88-90.
[331
, Dierential Forms Pa
kage, Wolfram Resear
h In
., Champaign, Illinois, MathSour
e 0205-221, 1994.
T. Wolf
T. Wolf
T. Wolf
T. Wolf
T. Wolf
A. Brand
T. Wolf
A. Brand
T. Wolf
A. Brand
I.M. Yaglom
V.V. Zharinov
A. Yu. Zharkov
F. Zizza
F. Zizza
B. Fu hssteiner