• No results found

Commun Nonlinear Sci Numer Simulat

N/A
N/A
Protected

Academic year: 2022

Share "Commun Nonlinear Sci Numer Simulat"

Copied!
13
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

ContentslistsavailableatScienceDirect

Commun Nonlinear Sci Numer Simulat

journalhomepage:www.elsevier.com/locate/cnsns

Research paper

Nonlocal hydrodynamic type of equations

Metin Gürses

a

, Aslı Pekcan

b,

, Kostyantyn Zheltukhin

c

a Department of Mathematics, Faculty of Sciences, Bilkent University, Ankara 06800, Turkey

b Department of Mathematics, Hacettepe University, Ankara 06800, Turkey

c Department of Mathematics, Middle East Technical University, Ankara 06800, Turkey

a r t i c l e i n f o

Article history:

Received 27 June 2019 Revised 2 January 2020 Accepted 1 March 2020 Available online 2 March 2020 Keywords:

Hydrodynamic equations Lax representations Conserved quantities Nonlocal reductions

a b s t r a c t

Weshowthattheintegrableequations ofhydrodynamictypeadmitnonlocalreductions.

WefirstconstructsuchreductionsforageneralLaxequationandthengiveseveralexam- ples.Thereducednonlocalequationsareofhydrodynamictypeandintegrable.Theyadmit Laxrepresentationsandhencepossessinfinitelymanyconservedquantities.

© 2020 Elsevier B.V. All rights reserved.

1. Introduction

Letusconsiderthefollowingsystemofevolutionequationsin(1+1)-dimension,

qit=Fi

(

qk,rk,qkx,rkx,qkxx,rkxx,...

)

, (1)

rti=Gi

(

qk,rk,qkx,rkx,qkxx,rkxx,...

)

, (2)

forall i,k=1,2,...,N.Here Fi andGi, (i=1,2,...,N)are functionsof thedynamical variables qi(x,t), ri(x,t), andtheir partialderivativeswithrespecttox.IftheabovesystemhasaLaxpairthenitisanintegrablesystem.Therearelocaland nonlocalreductionsofthecoupledsystem(1)and(2).Thelocalreductionsaregivenby

ri

(

x,t

)

=

ρ

qi

(

x,t

)

, (3)

and

ri

(

x,t

)

=

ρ

¯qi

(

x,t

)

, (4)

where

ρ

is a real constant and a bar over a letter denotes complex conjugation. By these reductions the systems of Eqs.(1)and(2)reducetoonesystemforqi’s

qit=F˜i



qj,qxj,qxxj,...



, i,j=1,2,...,N, (5)

Corresponding author.

E-mail addresses: gurses@fen.bilkent.edu.tr (M. Gürses), aslipekcan@hacettepe.edu.tr (A. Pekcan), zheltukh@metu.edu.tr (K. Zheltukhin).

https://doi.org/10.1016/j.cnsns.2020.105242 1007-5704/© 2020 Elsevier B.V. All rights reserved.

(2)

so that thesystem (2) consistently reduces to the above system (5). Here corresponding to (3) we have F˜=F

|

r=ρq and F˜=F

|

r=ρ¯qcorresponding to (4).There are also nonlocal reductions introduced by Ablowitz and Musslimani[1–3] given as

ri

(

x,t

)

=

ρ

qi

( ε

1x,

ε

2t

)

, (6)

and

ri

(

x,t

)

=

ρ

¯qi

( ε

1x,

ε

2t

)

, (7)

for i=1,2,...,N.Here

ρ

is a real constant and

ε

21=

ε

22=1.If

ε

1=

ε

2=1the above reductions turnto be local reduc- tions.When(

ε

1,

ε

2)=(1,−1),(−1,1),(−1,−1),wereducethesystem(1)and(2)bythesenonlocalreductionstononlocal timereflectionsymmetric(T-symmetric),spacereflectionsymmetric(S-symmetric),orspace-timereflectionsymmetric(ST- symmetric)differentialequations.

Thenonlocal reductionsofsystemsofintegrableequationswere firstconsistentlyappliedto thenonlinearSchrödinger system,whichledtothenonlocalnonlinearSchrödingerequation(nNLS)[1–3].Ithasbeenlatershownthattherearemany othersystemswheretheconsistentnonlocalreductionsarepossible[4–29].Forarecentreviewofthissubjectsee[27].See also[30]forthediscussionofsuperpositionofnonlocalintegrableequationsand[31]fortheoriginofnonlocalreductions.

One wayof obtaininga systemofintegrableequations,inGelfand-Dikiiformalism [32], istoconstructa Laxoperator onsomeLiealgebra.Theexamplesofsuchalgebrasarethematrixalgebra,algebraofdifferentialoperators,andalgebraof Laurentseries(see[33–35]and references there in).In most casesthe Lax operator is polynomial.It is a polynomial of the spectralparameterinthematrixalgebra,theoperatorDxinthealgebraofdifferentialoperators,andanauxiliaryvariablep (momentum)inthecaseofthealgebraofLaurentseries.

WhenwewritetheLaxequationonthealgebraofLaurentserieswederiveequationsofthehydrodynamictype,thatis asystemoffirstorderquasilinearpartialdifferentialequations

unt=

k

m=1

hmn

(

u

)

umx, n=1,2,...,k, (8)

where u=(u1,. . .,uk). The derived systems are integrable. This means they admit recursion operators and multi- Hamiltonianrepresentation(see[33,36–38]).

Fora hydrodynamictype systemwealso havelocalreductions(3),(4),andnonlocal reductions(6),(7)consistent for

ρ

2=1.Localreductionsforthehydrodynamictypesystemswereconsideredin[36].Inthisworkweaddresstotheproblem ofnonlocalreductionsfortheLaxequationsonthealgebraofLaurentseries.

Thelayoutofthepaperisasfollows.InSection2we giveashortreviewonthehydrodynamictypeof(dispersionless) equations andtheir Lax representations.In Section 3 we give nonlocal reductionsforreal andcomplex valuedfields for N=2.Weobtain someexplicitexamples ofnonlocalhydrodynamictypeofequationswithsome conservedquantities.In Section4wediscussthenonlocalreductionsingeneralandgiveexamplesforhighervaluesofN.

2. Laxequations

In thissection we describethe algebra ofLaurentseries,introduce necessary definitionsto write aLax equation, and givesomeexamples.Inlatersectionswegivenonlocalreductionsoftheseequations.

WeconsideranalgebraA,givenby A=

 α

=

−∞

α

n

(

x,t

)

pn :

α

n

(

x,t

)

Y



, (9)

whereYiseitherC(S1)-thespaceofperiodicfunctionsorS(R)-thespaceofsmooth asymptotically decreasingfunctions onR.APoissonbracketonA,isgivenby

{

f,g

}

k=pk



f

p

g

x

f

x

g

p



, f,gA, (10)

wherek∈Z.

OnthealgebraAwecandefineatracefunctional trk

α

=

I

Resk

α

dx,

α

A, (11)

where Resk

α

=

α

k−1(x,t). The set Iis either S1 forthe spaceof periodic functionsor Rfor the spaceof asymptotically decayingfunctions.Usingthetracefunctionalwedefineanon-degenerate,ad-invariant,symmetricpairing

( α

,

β )

= trk

α

·

β

,

α

,

β

A. (12)

ThealgebraAadmitsadecompositionintosub-algebras,closedwithrespecttothePoissonbracket,

A=A≥−k+1A<−k+1, (13)

(3)

where

A≥−k+1=

α

≥−k+1=



−k+1

α

n

(

x,t

)

pn:

α

n

(

x,t

)

Y

(14)

and

A<−k+1=

α

<−k+1=−k

−∞

α

n

(

x,t

)

pn:

α

n

(

x,t

)

Y

. (15)

So,wecandefiner-matrixmappingonA(see[35]andreferencestherein) Rk=1

2

 

≥−k+1



<−k+1



,

where≥−k+1and<−k+1areprojectionsonthesub-algebrasA≥−k+1 andA<−k+1respectively.

NowforaLaxoperator L=pN−1+

N−2

i=−1

piSi

(

x,t

)

, N∈N, (16)

wecanconsiderahierarchyoftheLaxequations(see[36])

L

tn =

Rk

(

LN−1m +n

)

,L

k=



L

m N−1+n

≥−k+1,L



k, n=1,2,..., (17)

wherem=0,1,2,...,(N− 2) isfixed. This is an integrable hierarchythat is the Lax equationsadmit multi-Hamiltonian representation(see[33]).Inparticulartheconservedquantitiesforthehierarchyaregivenby

Qn= trkLN−1m +n, n=1,2,.... (18)

Wenote that to considernonlocal reductionsitismore convenientto write theEq.(17)interms ofzerosof theLax operator.Thatisweset

L= 1 p

N

j=1

(

p− uj

)

. (19)

LetusgivesomeexamplesoftheLaxequationsandtheirconservedquantities.First,weconsiderexamplesofLaxequa- tionscorrespondingtotheLaxoperator

L= 1

p

(

p− u

)(

p

v )

. (20)

Example1. WetakethealgebraAwithk=0andconsidertheLaxEq.(17)withaLaxoperator(20).Forn=2,weobtain theshallowwaterwavessystem

1

2ut2=

(

u+

v )

ux+u

v

x,1

2

v

t2=

(

u+

v ) v

x+

v

ux. (21)

Taking n=3,4,..., we can have generalized symmetriesof the shallowwater waves system. Forinstance, the first two symmetriesare

1

3ut3=

(

u2+4u

v

+

v

2

)

ux+

(

2u2+2u

v ) v

x,1

3

v

t3=

(

2

v

2+2u

v )

ux+

( v

2+4u

v

+u2

) v

x, (22) and

1

4ut4 =



u3+9u2

v

+9u

v

2+

v

3



ux+



3u3+9u2

v

+3u

v

2

 v

x, 1

4

v

t4 =



3

v

3+9

v

2u+3

v

u2



ux+



v

3+9

v

2u+9

v

u2+u3

 v

x.

(23)

Theabovesystemsadmitinfinitelymanyconservedquantities.Thefirstthreeoftheconservedquantitiesare Q2=

I

(

u2

v

+u

v

2

)

dx, (24)

Q3=

I

(

u3

v

+3u2

v

2+u

v

3

)

dx, (25)

Q4=

I

(

u4

v

+6u3

v

2+6u2

v

3+6u

v

4

)

dx. (26)

(4)

Example2. WetakethealgebraAwithk=1andconsidertheLaxEq.(17)withaLaxoperator(20).Forn=1we obtain theTodasystem

ut1=u

v

x,

v

t1=

v

ux. (27)

Takingn=2,3,...,wecanhavegeneralizedsymmetriesoftheTodasystem.TheTodasystemadmitsinfinitelymanycon- servedquantities.Thefirstthreeoftheconservedquantitiesare

Q1=

I

(

u+

v )

dx, (28)

Q2=

I

(

u2+4u

v

+

v

2

)

dx, (29)

Q3=

I

(

u3+9u2

v

+9u

v

2+

v

3

)

dx. (30)

LetusgiveexamplesoftheLaxequationwithmoregeneralLaxoperator.

Example3. WetakethealgebraAwithk=0andconsidertheLaxEq.(17)withaLaxoperator L=1

p

(

p− u

)(

p

v )(

p− w

)

, (31)

andm=1.Forn=1weobtainthefollowingsystem 8

3ut1 =



−u2+

v

2+w2+4uv+4uw+6vw



ux+



2u2+2uv+6uw

 v

x+



2u2+6uv+2uw



wx, 8

3

v

t1 =



2

v

2+2uv+6vw



ux+



v

2+u2+w2+4uv+4vw+6vw

 v

x+



2

v

2+6uv+2vw



wx, 8

3wt1 =



2w2+6wv+2uw



ux+



2w2+2wv+6uw

 v

x+



−w2+

v

2+u2+4wv+4uw+6uv



wx.

(32)

Taking n=2,3,..., we can have generalizedsymmetries ofthe above system. The above system admitsinfinitely many conservedquantities.Thefirsttwooftheconservedquantitiesare

Q1=

I

(

u2+

v

2+w2− 2u

v

− 2uw− 2

v

w

)

dx, (33)

Q2 =

I

(

u4+

v

4+w4− 4

(

u3

v

− u3w− u

v

3

v

3w− uw3

v

w3

)

+6

(

u2

v

2+u2w2+6

v

2w2

)

+36

(

u2

v

w+u

v

2w+u

v

w2

))

dx, (34)

Example4. Forsimplicity,wetakethealgebraAwithk=0andconsidertheLaxEq.(17)withaLaxoperator L=1

p

(

p− u

) (

p− u1

) (

p

v ) (

p

v

1

)

, (35)

andm=0.Forn=1weobtainthefollowingsystem

ut1 =

(

u1

v

+u1

v

1+

vv

1

)

ux+

(

u

v

+u

v

1

)

u1x+

(

uu1+u

v

1

) v

x+

(

uu1+u

v ) v

1x, u1t1=

(

u

v

+u

v

1+

vv

1

)

u1x+

(

u1

v

+u1

v

1

)

ux+

(

uu1+u1

v

1

) v

x+

(

u1u+u1

v ) v

1x,

v

t1 =

(

uu1+u

v

1+u1

v

1

) v

x+

(

u1

v

+

vv

1

)

ux+

(

u

v

+

vv

1

)

u1x+

(

u

v

+u1

v ) v

1x,

v

1t1 =

(

uu1+u

v

+u1

v ) v

1x+

(

u1

v

1+

vv

1

)

ux+

(

u

v

1+

vv

1

)

u1x+

(

u

v

1+u1

v

1

) v

x. (36) Taking n=2,3,..., we can have generalizedsymmetries ofthe above system. The above system admitsinfinitely many conservedquantities.Thefirsttwooftheconservedquantitiesare

Q1=

I

uu1

vv

1dx, (37)

Q2=

I

(

u2u21

v

2

v

1+u2u21

vv

21+u2u1

v

2

v

21+uu21

v

2

v

21

)

dx. (38)

3. NonlocalreductionsfortwocomponentLaxoperator

TostudynonlocalreductionsoftheLaxEq.(17),westartwiththecaseoftwocomponentsystems.Thatiswestudythe hierarchy

L

tn =

Ln≥−k+1; L

k, n=1,2,... (39)

withtheLaxoperator(20).

(5)

3.1. Nonlocalreductionsforrealvalueddependentvariables

Inthissectionweassumethatourdependentvariablesarerealvaluedandconsiderreductionoftheform

v (

x,t

)

=

ρ

u

( ε

1x,

ε

2t

)

=

ρ

uε. (40)

ThereducedequationistheLaxEq.(39)withtheLaxoperator L= 1

p

(

p− u

)(

p

ρ

uε

)

. (41)

Tofindadmissiblereductionsweintroducethefollowingchangeofvariables

˜

x=

ε

1x, t˜n=

ε

2tn, p˜=

ρ

p. (42)

InnewvariablestheLaxoperatorbecomes L=

ρ

p˜

( ρ

p˜− uε

)( ρ

p˜

ρ

u

)

=

ρ

L˜, (43)

whereL˜= 1

˜

p(p˜− u)(p˜

ρ

uε).TheLaxEq.(39)takestheform

ε

2

t˜n

( ρ

L˜

)

=

( ρ

p˜

)

k



ρ ∂

p˜

( ρ

nL˜n≥−k+1

) ε

1

x˜

( ρ

L˜

)

ε

1

x˜

( ρ

nL˜n≥−k+1

) ρ ∂

p˜

( ρ

L˜

)



(44)

or

ρ

k+n+1

ε

2

ε

1

t˜n

(

˜L

)

=p˜k



p˜L˜n≥−k+1

x˜L˜

x˜L˜n≥−k+1

p˜L˜



. (45)

It is easy to see that if

ρ

k+n+1

ε

2

ε

1=1 then the hierarchy(39) withLax operator (41) is invariant under thechange of variablesandwehaveaconsistentnonlocalreduction.

Since

ρ

2=1,wehaveonecondition

ρ

k

ε

2

ε

1=1forall oddnandanothercondition

ρ

k+1

ε

2

ε

1=1forallevenn.Thus, toconsidernonlocalreductionofthewholehierarchy(39)wehavetosplititintotwohierarchies

L

t2n =

L2≥−kn +1; L

k, n=1,2,... (46)

and

L

t2n−1 =

L2≥−kn−1+1; L

k, n=1,2,.... (47)

Taking

ρ

,

ε

1,and

ε

2satisfying

ρ

k+1

ε

2

ε

1=1, (48)

weobtainnonlocalreductionsforthehierarchy(46).Notethattheseequationsadmitthefollowingconservedquantities

Q2n= trkL2n, n=1,2,..., (49)

whereLisgivenby(41). Taking

ρ

,

ε

1,and

ε

2satisfying

ρ

k

ε

2

ε

1=1, (50)

weobtainnonlocalreductionsforthehierarchy(47).Notethattheseequationsadmitthefollowingconservedquantities

Q2n−1= trkL2n−1, n=1,2,..., (51)

whereLisgivenby(41).

Letusgivesomeexamples.Asafirstexampleweconsidertheshallowwaterwavessystem. Notethat fromnowonin allexamplesinthetextwelettn=t foralln.

Example5. Tohavenonlocalreductionsoftheshallowwaterwavessystem(21),weneedtoworkwiththehierarchy(46). Taking

ρ

,

ε

1,and

ε

2satisfying(48)weobtainpossiblereductions.

a. S-symmetricreduction,

ρ

=−1,

ε

1=−1,

ε

2=1,is 1

2ut

(

x,t

)

=

(

u

(

x,t

)

− u

(

−x,t

) )

ux

(

x,t

)

− u

(

x,t

)

ux

(

−x,t

)

. (52)

Ithasinfinitelymanysymmetries,withthefirstsymmetry,thereductionof(23),being

(6)

1

4ut

(

x,t

)

=



u3

(

x,t

)

− 9u2

(

x,t

)

u

(

−x,t

)

+9u

(

x,t

)

u2

(

−x,t

)

− u3

(

−x,t

) 

ux

(

x,t

)



3u3

(

x,t

)

− 9u2

(

x,t

)

u

(

−x,t

)

+3u

(

x,t

)

u2

(

−x,t

) 

ux

(

−x,t

)

. (53)

Theaboveequationshaveinfinitelymanyconservedquantities.Thefirsttwoconservedquantitiesare Q2=

I

(

−u2

(

x,t

)

u

(

−x,t

)

+u

(

x,t

)

u2

(

−x,t

))

dx, (54)

Q4=

I

(

−u4

(

x,t

)

u

(

−x,t

)

+6u3

(

x,t

)

u2

(

−x,t

)

− 6u2

(

x,t

)

u3

(

−x,t

)

+u

(

x,t

)

u4

(

−x,t

))

dx. (55)

b. T-symmetricreduction,

ρ

=−1,

ε

1=1,

ε

2=−1,is 1

2ut

(

x,t

)

=

(

u

(

x,t

)

− u

(

x,−t

) )

ux

(

x,t

)

− u

(

x,t

)

ux

(

x,−t

)

. (56)

Ithasinfinitelymanysymmetries,withthefirstsymmetry,thereductionof(23),being 1

4ut

(

x,t

)

=



u3

(

x,t

)

− 9u2

(

x,t

)

u

(

x,−t

)

+9u

(

x,t

)

u2

(

x,−t

)

− u3

(

x,−t

) 

ux

(

x,t

)



3u3

(

x,t

)

− 9u2

(

x,t

)

u

(

x,−t

)

+3u

(

x,t

)

u2

(

x,−t

) 

ux

(

x,−t

)

. (57)

Theaboveequationshaveinfinitelymanyconservedquantities.Thefirsttwoconservedquantitiesare Q2=

I

(

−u2

(

x,t

)

u

(

x,−t

)

+u

(

x,t

)

u2

(

x,−t

))

dx, (58)

Q4=

I

(

−u4

(

x,t

)

u

(

x,−t

)

+6u3

(

x,t

)

u

(

x,−t

)

2− 6u2

(

x,t

)

u

(

x,−t

)

3+6u

(

x,t

)

u4

(

x,−t

))

dx. (59) c. ST-symmetricreduction,

ρ

=1,

ε

1=−1,

ε

2=−1,is

1

2ut

(

x,t

)

=

(

u

(

x,t

)

+u

(

−x,−t

) )

ux

(

x,t

)

+u

(

x,t

)

ux

(

−x,−t

)

. (60)

Ithasinfinitelymanysymmetrieswiththefirstsymmetry,thereductionof(23),being 1

4ut

(

x,t

)

=



u3

(

x,t

)

+9u2

(

x,t

)

u

(

−x,−t

)

+9u

(

x,t

)

u2

(

−x,−t

)

+u3

(

−x,−t

) 

ux

(

x,t

)

+



3u3

(

x,t

)

+9u2

(

x,t

)

u

(

−x,−t

)

+3u

(

x,t

)

u2

(

−x,−t

) 

ux

(

−x,−t

)

. (61)

Theaboveequationshaveinfinitelymanyconservedquantities.Thefirsttwoconservedquantitiesare Q2=

I

(

u2

(

x,t

)

u

(

−x,−t

)

+u2

(

x,t

)

u

(

−x,−t

))

dx, (62)

Q4=

I

(

u4

(

x,t

)

u

(

−x,−t

)

+6u3

(

x,t

)

u

(

−x,−t

)

2+6u2

(

x,t

)

u3

(

−x,−t

)

+u

(

x,t

)

u4

(

−x,−t

))

dx. (63) Wecanalsoconsiderthereductionsfortheoddsymmetriesoftheshallowwaterwavesequations.

Example6. Tohavenonlocalreductionsofthesystem(22),weneedtoworkwiththehierarchy(47).Taking

ρ

,

ε

1,and

ε

2, satisfying(50)weobtainonepossiblereduction.

ST-symmetricreduction,

ρ

=1,

ε

1=−1,

ε

2=−1,is 1

3ut

(

x,t

)

=



u

(

x,t

)

2+4u

(

x,t

)

u

(

−x,−t

)

+u2

(

−x,−t

) 

ux

(

x,t

)

+



2u2

(

x,t

)

+2u

(

x,t

)

u

(

−x,−t

) 

ux

(

−x,−t

)

. (64)

Theabovesystemhasinfinitelymanysymmetriesandconservedquantities.Thefirsttwoconservedquantitiesare Q3=

I

(

u3

(

x,t

)

u

(

−x,−t

)

+3u2

(

x,t

)

u2

(

−x,−t

)

+u

(

x,t

)

u

(

−x,−t

)

3

)

dx, (65) Q5 =

I

(

u5

(

x,t

)

u

(

−x,−t

)

+10u4

(

x,t

)

u2

(

−x,−t

)

+20u3

(

x,t

)

u3

(

−x,−t

)

+10u2

(

x,t

)

u4

(

−x,−t

)

+u

(

x,t

)

u5

(

−x,−t

))

dx. (66) LetusalsoconsiderthereductionsoftheTodasystem.

Example7. TohavenonlocalreductionsoftheTodasystem(27),weneedtoworkwiththehierarchy(47).Taking

ρ

,

ε

1,and

ε

2 satisfying(50)wecanobtainthefollowingreductions:S-symmetricreduction,

ρ

=−1,

ε

1=−1,

ε

2=1;T-symmetricre- duction,

ρ

=−1,

ε

1=1,

ε

2=−1;ST-symmetricreduction,

ρ

=1,

ε

1=−1,

ε

2=−1.Forinstance,theST-symmetricreduction oftheTodasystemis

ut

(

x,t

)

=u

(

x,t

)

ux

(

−x,−t

)

. (67)

(7)

Ithasinfinitelymanysymmetries,withthefirstsymmetry 1

3ut

(

x,t

)

=2u

(

x,t

) 

u

(

x,t

)

u

(

−x,−t

)

+u2

(

−x,−t

) 

ux

(

x,t

)

+u

(

x,t

) 

u2

(

x,t

)

+4u

(

x,t

)

u

(

−x,−t

)

+u2

(

x,t

) 

ux

(

−x,−t

)

, (68)

whichis the ST-symmetricreduction of thesecond equation inthe hierarchy (39). It alsohas infinitelymany conserved quantities.Thefirsttwoconservedquantitiesare

Q1=

I

(

u

(

x,t

)

+u

(

−x,−t

))

dx, (69)

Q3=

I

(

u3

(

x,t

)

+9u2

(

x,t

)

u

(

−x,−t

)

+9u

(

x,t

)

u2

(

−x,−t

)

− u3

(

−x,−t

))

dx. (70) Remark3.1. Weconsideredthereductionsforthewholehierarchy(39).Wecanalsoconsidernonlocalreductionsjustfor oneequationfromthehierarchy.Forinstance,takingk=0andn=2in(39)wegetthefollowingequations:

1

2ut=

(

u+

v )

ux+u

v

x, (71)

1

2

v

t=

(

u+

v ) v

x+

v

ux. (72)

Letusconsider nonlocalreductionsoftheseequationsonly. Assumethat

v

(x,t)=

ρ

u(

ε

1x,

ε

2t)=

ρ

uε,where

ε

21=

ε

22=1, and

ρ

isarealconstant,thentheabovesystemtakesform

1

2

ρ

ut=

(

uε+

1

ρ

u

)

ux+uuεx, (73)

1

2

ε

1

ε

2ut=

(

uε+

ρ

u

)

ux+uuεx. (74)

Forconsistency,we musthave

ρ

=

ε

1

ε

2 which isthesameconditionasequality(48).Therefore workingwiththe whole hierarchyweobtainallpossiblecases.

3.2.Nonlocalreductionsforcomplexvalueddependentvariables

Inthissectionweassumethatalldependentvariablesarecomplexvaluedfunctions.Tostudyreductionsofsuchequa- tionsitisconvenienttointroduceaconstantintheLaxequation (39)bysettingnewtimet=at,a∈C.Soweconsidera hierarchy

a

L

tn =

Ln≥−k+1; L

k, n=1,2,.... (75)

Notethatafterthechangeofthetimevariableweshalluseagainttodenotenewtime.

Thereductionisoftheform

v (

x,t

)

=

ρ

u¯

( ε

1x,

ε

2t

)

=

ρ

u¯ε. (76)

Tofindadmissiblereductionsweusethefollowingchangeofvariables

˜

x=

ε

1x, t˜n=

ε

2tn, p˜=

ρ

p. (77)

InnewvariablestheLaxoperatorbecomes L=

ρ

˜

p

( ρ

p˜− uε

)( ρ

p˜

ρ

u¯

)

=

ρ

¯L, (78)

whereL=1

˜

p(p˜− u)(p˜

ρ

u¯ε)andtheLaxEq.(39)takestheform a

ε

2

t˜n

( ρ

¯L

)

=

( ρ

p˜

)

k



ρ ∂

p˜

( ρ

n

(

¯L

)

n≥−k+1

) ε

1

x˜

( ρ

¯L

)

ε

1

x˜

( ρ

n

(

¯L

)

n≥−k+1

) ρ ∂

p˜

( ρ

¯L

)



. (79)

Takingcomplexconjugatesofbothsidesintheaboveequalitywecanget a

ρ

k+n+1

ε

2

ε

1

t˜n

(

L

)

=p˜k



p˜

(

L

)

n≥−k+1

x˜L

x˜

(

L

)

n≥−k+1

p˜L



. (80)

Ifa

ρ

k+n+1

ε

2

ε

1=a then the hierarchy(75)is invariant under thechange ofvariables andwe havea consistent nonlocal reduction.

(8)

Since

ρ

2=1wehaveoneconditiona

ρ

k

ε

2

ε

1=aforalloddnandanotherconditiona

ρ

k+1

ε

2

ε

1=aforallevenn.Thus, toconsiderlocalreductionofthehierarchy(75)wehavetosplititintotwohierarchies(46)and(47)asintherealcase.

Letusgiveoneexampleofthereductionforcomplexvalueddependentvariables.

Example8. Wetakek=0andconsiderthehierarchy(75)withneven.Thefirstsystemcorrespondington=2is aut = 2

(

u+

v )

ux+2u

v

x,

avt = 2

(

u+

v ) v

x+2

v

ux. (81)

Taking

ρ

,

ε

1,and

ε

2 satisfyinga

ρε

2

ε

1=aweobtainpossiblereductions.

a. S-symmetricreduction,

ε

1=−1,

ε

2=1,is a

2ut

(

x,t

)

=

(

u

(

x,t

)

+

ρ

u¯

(

−x,t

))

ux

(

x,t

)

+

ρ

u

(

x,t

)

u¯x

(

−x,t

)

, (82)

where we can take

ρ

=1 and the constant a equals to a pure imaginary number or

ρ

=−1 and the constant a equalstoarealnumber.Theaboveequation hasinfinitelymanysymmetriesandconservedquantities.Thefirsttwo conservedquantitiesare

Q2=

I

(

−u2

(

x,t

) ρ

u¯

(

−x,t

)

+u

(

x,t

) ρ

u¯2

(

−x,t

))

dx, (83)

Q4=

I

(

−u4

(

x,t

) ρ

u¯

(

−x,t

)

+6u3

(

x,t

) ρ

u¯2

(

−x,t

)

− 6u2

(

x,t

) ρ

u¯3

(

−x,t

)

+u

(

x,t

) ρ

u¯4

(

−x,t

))

dx. (84)

b. T-symmetricreduction,

ε

1=1,

ε

2=−1,is a

2ut

(

x,t

)

=

(

u

(

x,t

)

+

ρ

u¯

(

x,−t

))

ux

(

x,t

)

+

ρ

u

(

x,t

)

u¯x

(

x,−t

)

, (85) where we can take

ρ

=1 and the constant a equals to a pure imaginary number or

ρ

=−1 and the constant a equalstoarealnumber.Theaboveequation hasinfinitelymanysymmetriesandconservedquantities.Thefirsttwo conservedquantitiesare

Q2=

I

(

−u2

(

x,t

) ρ

u¯

(

x,−t

)

+u

(

x,t

) ρ

u¯2

(

x,−t

))

dx, (86)

Q4=

I

(

−u4

(

x,t

) ρ

u¯

(

x,−t

)

+6u3

(

x,t

) ρ

u¯2

(

x,−t

)

− 6u2

(

x,t

) ρ

u¯3

(

x,−t

)

+u

(

x,t

) ρ

u¯4

(

x,−t

))

dx. (87) c. ST-symmetricreduction,

ε

1=−1,

ε

2=−1,is

a

2ut

(

x,t

)

=

(

u

(

x,t

)

+

ρ

u¯

(

−x,−t

))

ux

(

x,t

)

+

ρ

u

(

x,t

)

u¯x

(

−x,−t

)

, (88) where we can take

ρ

=1 and the constant a equals to a real number or

ρ

=−1 andthe constant a equals to a pureimaginarynumber.Theaboveequation hasinfinitelymanysymmetriesandconservedquantities.Thefirsttwo conservedquantitiesare

Q2=

I

(

−u2

(

x,t

) ρ

u¯

(

−x,−t

)

+u

(

x,t

) ρ

u¯2

(

−x,−t

))

dx, (89)

Q4=

I

(

−u4

(

x,t

) ρ

u¯

(

−x,−t

)

+6u3

(

x,t

) ρ

u¯2

(

−x,−t

)

− 6u2

(

x,t

) ρ

u¯3

(

−x,−t

)

+u

(

x,t

) ρ

u¯4

(

−x,−t

))

dx. (90) 4. ReductionswithageneralLaxoperator

TostudynonlocalreductionsoftheLaxEq.(17)withthegeneralLaxoperator(19),wecandividethevariablesuiinto pairsandsetvariablesthatdonothaveapairtozero.Foreachpairofvariablesweconsiderarelationsimilarto(40)and proceedasbefore.Asanexampleletusconsidertwospecialcases.

4.1. ThreecomponentLaxoperator

InthissectionwetaketheLaxoperator L=1

p

(

p− u

)(

p

v )(

p− w

)

, (91)

andconsiderthehierarchy(17)withm=1.Thuswehave

L

tn =



L≥−k12+n+1; L



k, n=1,2,.... (92)

Referenties

GERELATEERDE DOCUMENTEN

Of course, in the final analysis we should be happy with the doctoral dissertation by De Vries and the Korteweg-De Vries paper, espe- cially since it is certainly conceivable that

The invention relates to a process for injection moulding of objects consisting of a number of layers of different materials, in which process the different materials ar

Based on experimental evidence obtained in triclinic DMM-TCNQ2, it will be argued that in zero field there is an instability for a lattice modulation, that approaches

(a) and (b) Estimation of time varying parameter hðtÞ in Problem 5.4 from observational data for different values of signal to noise ratio... It is assumed that the initial function H

So far the complete classification has been done for some evolution types of autonomous equations.lY5 There are some partial attempts of the classification of

In this article an algorithm of classification of integrable discrete chains of form 共1兲 is sug- gested based on the notion of the characteristic Lie algebra 共see also Refs...

A method of integrable discretization of the Liouville type nonlinear partial differential equations is suggested based on integrals.. New examples of discrete Liouville type models

The only restriction is that if there are any numbered equations inside the subequations environment that break out of the subequation numbering sequence, they would have to be