• Nenhum resultado encontrado

TEMA (São Carlos) vol.14 número1

N/A
N/A
Protected

Academic year: 2018

Share "TEMA (São Carlos) vol.14 número1"

Copied!
10
0
0

Texto

(1)

Uma Publiação daSoiedade BrasileiradeMatemátiaApliadaeComputaional.

A review on some results on loal onservation

laws for ertain evolution equations 1

I.L. Freire 2

, CMCC - Centro de Matemátia, Computação e Cognição, UFABC

-Universidade Federal do ABC,Rua Santa Adélia, 166, BairroBangu, 09.210-170

SantoAndré,SP,Brasil

J.C.S. Sampaio 3

, IMECC - Departamento de Matemátia Apliada, UNICAMP

-UniversidadeEstadualdeCampinas,RuaSérgioBuarquedeHolanda,651,Cidade

UniversitáriaZeferinoVaz,BarãoGeraldo,13083-859Campinas,SP,Brasil.

Abstrat. Inthis workwe revisitsome reentresults ononservationlaws for a

lassoffth-orderevolutionequationsuptofth-order.

Key-words. Ibragimov'stheorem,onservationlaws,fth-orderKdVequations.

1. Introdution

Thisworkorrespondstothetalk[6℄,givenbyJ.C.S.Sampaio,duringtheXXXIV

CNMAC -XXXIV CongressoNaional deMatemátiaApliadaeComputaional

-September,17-212012,ÁguasdeLindóia-SP,Brasil,where itwaspresentedand

disussed somereent results[5℄ with respet to how to obtainonservation laws

foranequationwithoutalassialLagrangian.

Itis well known that forequationsarisingfrom theEuler-Lagrangeequations,

theelebratedNoethertheoremprovidesanelegantwayforndingonserved

quan-tities for suh equation, see [1, 17, 20℄. However,when it is onsidered equations

withoutvariationalstruture, it is impossibleto apply Noether's approah. Then

thefollowingquestionarises: howanonendaonservedvetorforanequation

withoutLagrangians?

Arossthelastentury,manyapproaheshavebeenproposedin orderto

over-ame this problem. The interested readeris direted to thereferene [19℄, where

someof these developmentsweredisussed. Inthis paperweuse themost reent

approah proposed in [12℄, where it was shown ageneral result onneting

sym-metries and onservation laws, in order to nd onserved vetorsfor the general

equation

u

t

+

αuu

x

+

βu

2

u

x

+

γu

xxx

+

µu

xxxxx

= 0.

(1.1)

1

ThisworkwassupportedbyFAPESP,grantsno2011/23538-0and2011/19089-6.

2

igor.freireufab.edu.brandigor.leite.freiregmail.om.

3

(2)

Following[2℄, wewould liketo allthis new formulationfor nding onserved

vetorsasIbragimov'stheoremononservationlaws.

Inthis paperitisonsidered theequation(1.1)withthefollowingrestritions:

(α, β)

6

= (0,

0)

and

(γ, µ)

6

= (0,

0)

. Infat,in theasewhenever

(γ, µ) = (0,

0)

itis

obtainedapartiular ase ofthe invisidBurgers equation, whih wasonsidered

in [2, 4℄. Whenever

α

=

β

= 0

it is obtained a linear equation and here we are interestedinnonlinearphenomena.

Apartfromthepreviousmentionedequations,equation(1.1)inludes

TheelebratedKdVequation

u

t

+

αuu

x

+

γu

xxx

= 0;

(1.2)

GeneralKawaharaequation

u

t

+

αuu

x

+

γu

xxx

+

µu

xxxxx

= 0;

(1.3)

SimpliedmodiedKawaharaequation

u

t

+

βu

2

u

x

+

µu

xxxxx

= 0;

(1.4)

SimpliedKawaharaequation

u

t

+

αuu

x

+

µu

xxxxx

= 0

(1.5)

and

Gardner equation

u

t

+ 6(u

+

u

2

)u

x

+

u

xxx

= 0.

(1.6)

Allof theseequationsis omingfromthe elebratedKdV equation,whihwas

deduedbyKortewegandhisstudentde-Vriesformodelingshallowwaveequations.

These equations are employed in Mathematial Physisin orderto desribea lot

of dispersivephenomena, suh asplasmaphenomena, while equationsof thetype

(1.2)arealsousedtomodeltheGreatRedSpot ofJupiter,see[21℄.

Althoughour main purpose is to revisit someof our previousresults given in

[5, 21℄, in thepresentwork wealso presentsomenew onservation lawsfor

equa-tions of the lass (1.1). Namely, the new onserved vetors derived by using the

developments[12,15,16℄are

C

0

=

u

2

,

C

1

=

2

3

αu

3

+ 2µuu

xxxx

2µu

x

u

xxx

+

µu

2

xx

,

(1.7)

forthesimpliedKawaharaequation(1.5),and

C

0

=

u

u

2

,

C

1

= 3u

4

+ 6u

3

+ 3u

2

+ 2uu

xx

+

u

xx

u

2

x

,

(3)

fortheGardnerequation(1.6).

The remaining of this paper is as the follows. In the next setion we revisit

theIbragimov'stheorem on onservationand the neessarytoolsto employ itfor

nding loal onservedvetors. Next,in the setion3, weshow that the lass of

equations(1.1)isnonlinearlyself-adjoint. Thisallowsustondloalonservation

lawsfor theseequations, whihwill bedonein thefollowingsetions. Some parts

ofthisreviewloselyfollowourreferenes[5,21℄.

2. Ibragimov's theory

Inthissetionwerevisitthebasitoolsaboutthereentdevelopmentsstartedwith

thefruitfulwork[12℄,whereIbragimovprovedanewonservationtheorem. Nextwe

introduethe reentonept of nonlinearself-adjointness, proposed by Ibragimov

in[15,16℄.

Thiseld, nowadays,is arih branh in the eld of groupanalysis and ithas

been attrating the attention of a big number of researhers. Many of them are

interested in nding nonlinearly self-adjoint properties of equations, as it an be

seenin[2,3,4,5,7,8,9,10,11, 13,14, 22℄andreferenestherein.

2.1. Adjoint equations and nonlinear self-adjointness

Let

x

= (x

1

,

· · ·

, x

n

)

be

n

independentvariables,

u

=

u(x)

beadependentvariable. Thesetof

k

thorderderivativesof

u

isdenotedby

u

(k)

,where

k

inapositiveinteger number. Considerthesetoffuntions depending on

x, u

and

u

derivativesupto a niteorder.

Aloallyanalytifuntion ofanitenumberofthevariables

x, u

and

u

deriva-tivesisalled adierentialfuntion. Thehighestorder ofderivativesappearingin

thedierentialfuntionisalledtheorderofthisfuntion. Thevetorspaeofall

dierentialfuntionsofniteorderis denotedby

A

.

Theformalsum

δ

δu

=

∂u

+

X

j=1

(

1)

j

D

i

1

· · ·

D

i

j

∂u

i

1

···

i

j

(2.9)

isthewellknownEuler-Lagrangeoperator.

Let

F

∈ A

be a dierential funtion. From this funtion we an obtain a dierentialequation

F

(x, u,

· · ·

, u

(s)

) = 0

(2.10)

andthefollowingnewdierentialfuntion,alled formalLagrangian,givenby

L

=

vF

, where

v

=

v(x)

isanotherdependentvariable.

Equation

(2.10)

is said to be nonlinearly self-adjoint if the equation obtained fromtheadjointequation

F

(x, u, v,

· · ·

, u

(s)

, v

(s)

) :=

δ

L

(4)

bythesubstitution

v

=

φ(x, u)

withaertainfuntion

φ(x, u)

6

= 0

isidentialwith theoriginalequation

(2.10)

,that is,

F

(x, u, v, u

(1)

, v

(1)

,

· · ·

, u

(s)

, v

(s)

)

v=φ(x,u)

= 0.

(2.12)

Whenever

(2.12)

holdsforaertaindierentialfuntion

φ

suhthat

φ

u

6

= 0

and

φ

x

6

= 0

,equation

(2.11)

isalled weakself-adjoint.

In other words: equation (2.11) is said to be nonlinearly self-adjoint if there

existsafuntion

φ

=

φ(x, u)

suhthat

F

|

v=φ

=

λ(x, u,

· · ·

)F,

(2.13)

forsomedierentialfuntion

λ

=

λ(x, u,

· · ·

)

.

2.2. Ibragimov's theorem on onservation laws

Thefollowingresultwasprovedin[12℄.

Theorem2.1 (Ibragimov'stheoremononservationlaws). Let

X

=

ξ

i

∂x

i

+

η

∂u

be any symmetry (Lie point, Lie-Bäklund, nonloal symmetry) of equation

(2.10)

and

(2.11)

be the adjoint equation toequation

(2.10)

. The ombined system

(2.10)

and

(2.11)

hasthe onservation law

D

i

C

i

= 0

,where

C

i

=

ξ

i

L

+

W

L

∂u

i

D

j

L

∂u

ij

+

D

j

D

k

L

∂u

ijk

− · · ·

+D

j

(W

)

L

∂u

ij

D

k

L

∂u

ijk

+

· · ·

+D

j

D

k

(W

)

L

∂u

ijk

− · · ·

+

· · ·

(2.14)

and

W

=

η

ξ

i

u

i

.

2.3. Algorithm

Ibragimov'stheorem on onservation laws an be resumed by the following

algo-rithm(see[12,2℄forfurther details): givenaPDE

F

=

F

(x, u, u

(1)

· · ·

, u

(n)

) = 0,

weonstrutaLagrangian

L

=

vF

.

FromtheEuler-Lagrangeequations,thefollowingsystemisobtained:

F

(x, u, u

(1)

· · ·

, u

(n)

) = 0,

(2.15)

F

(x, u, v,

· · ·

, u

(s)

, v

(s)

) = 0.

(2.16)

(5)

3. Nonlinearlyself-adjoint lassiation ofthe

equa-tion (1.1)

Thefollowingtheoremwasprovedin [5℄ (seealso[21℄).

Theorem 3.2. Equation

(1.1)

, with

(γ, µ)

= 0

6

and

(α, β)

6

= (0,

0)

, is nonlinearly

self-adjoint.

Remarks: Equation(1.1)isalsononlinearlyself-adjointwithouttherequested

hypothesis in the Theorem 3.2. The mentionedhypothesis only reets the

non-linearityoftheequationandthefat thatweareonsideringdispersiveequations.

However,astraightforwardalulationshowsthatweanremovethem. Weleave

thedetailstotheinterestedreader. Itanalsobeuseful,onegoinginthisdiretion,

tosee [2,3,4℄.

Proof. Letusdenotetheleftsideof(1.1)by

F

=

u

t

+

αuu

x

+

γu

xxx

+

µu

xxxxx

.

(3.17)

Then,theadjointequationto

F

= 0

is

F

:=

v

t

(αu

+

βu

2

)v

x

γv

xxx

µv

xxxxx

= 0.

(3.18)

Nowweonlymustanalyzethefollowingtwoases. Below,

a

1

, a

2

and

a

3

are arbi-traryonstants.

Assume

β

= 0

in (1.1). Thenthesubstitution

φ(x, t, u) =

a

1

(x

αtu) +

a

2

u

+

a

3

in (3.18)makestheadjointequationequivalentto theoriginalequation.

Nowsuppose

β

6

= 0

in(1.1). Substituting

φ(x, t, u) =

a

1

u

+

a

2

insteadof

v

in (3.18), weobtainamultipleof(1.1).

Thisprovesourstatement.

4. Non-loal onservation laws

Here we nd nonloal onservation laws for the equation (1.1). We follow the

algorithmpreviouslydisussedinSetion2.3.

Atually,thersttwostepsofsuhamentionedalgorithmhavejustbeendone

beausetheformalLagrangianhadalreadybeenobtainedwhiletheadjointequation

(6)

Withrespetto theseond step,theomponentsofthevetoraregivenby

C

0

=

τ

L

+

W v,

C

1

=

ξ

L

+

W

v(αu

+

βu

2

) +

γv

xx

+

µv

xxxx

D

x

(W

)(γv

x

+

µv

xxx

) +

D

x

2

(W

)(γv

+

µv

xx

)

D

3

x

(W

)(µv

x

) +

D

x

4

(W

)(µv),

(4.19)

where

L

=

vF

,

W

=

η

τ u

t

ξu

x

and

X

=

τ(x, t, u)

∂t

+

ξ(x, t, u)

∂x

+

η(x, t, u)

∂u

isanyLiepointsymmetryof(1.1).

Theobtainedonservedvetordependson

v

beauseitisaonservedvetorto thesystem

u

t

+

αuu

x

+

βu

2

u

x

+

γu

xxx

+

µu

xxxxx

= 0,

v

t

(αu

+

βu

2

)v

x

γv

xxx

µv

xxxxx

= 0.

Suhaonservedvetor

C

= (C

0

, C

1

)

is, then,anonloalonservedvetor.

5. Loal onservation laws

AordingtoIbragimov'stheoremononservationlaws(seealsotheorresponding

algorithm), a onserved vetor to (1.1) and (3.18) is

C

= (C

0

, C

1

)

, whose the

omponents are given by (4.19) and, as it was already pointed out, a nonloal

onservedvetortotheoriginalequation.

Inordertondloal onservedvetors,weusetheLiepointsymmetriesfound

in [18, 21℄ and the fat that under the substitutions given by Theorem 3.2, the

nonloalonservedvetorsbeomealoalonservedvetor. Weillustrate thisfat

at the same time that weproeed ouralulation for establishingthe elds (1.7)

and(1.8).

5.1. Conservation law for the simplied Kawahara equation

ConerningthesimpliedKawaharaequation(1.5),byusingthedilational

symme-try

X

=

x

∂x

+ 5t

∂t

4u

(7)

foundin[18℄, from(4.19)weobtain

C

0

=

4vu

xvu

x

+ 5αtvuu

x

+ 5µtvu

xxxxx

,

C

1

=

xvu

t

4αvu

2

4µuv

xxxx

µxu

x

v

xxxx

5αtvuu

t

5µtu

t

v

xxxx

+ 5µu

x

v

xxx

+

µxu

xx

v

xxx

+ 5µtu

xt

v

xxx

6µv

xx

u

xx

µxv

xx

u

xxx

5µtv

xx

u

xxt

+ 7µv

x

u

xxx

+µxv

x

u

xxxx

+ 5µtv

x

u

xxxt

8µvu

xxxx

5µtvu

xxxxt

.

(5.21)

Substituting

v

=

x

αtu

intotheomponentsabove,wearrivedat

C

0

=

2xu

+

αtu

2

+

D

x

x

2

u

+ 3αxtu

2

5

3

α

2

t

2

u

3

+D

x

5µxtu

xxxx

5µtu

xxx

5αµt

2

uu

xxxx

+ 5αµt

2

u

x

u

xxx

5

2

αµt

2

u

2

xx

,

C

1

=

αxu

2

+

2

3

α

2

tu

3

2µxu

xxxx

+ 2µu

xxx

+ 2αµtuu

xxxx

2αµtu

x

u

xxx

+

αµtu

2

xx

+

D

t

x

2

u

3αxtu

2

+

5

3

α

2

t

2

u

3

D

t

5µxtu

xxxx

+ 5µtu

xxx

+ 5αµt

2

uu

xxxx

5αµt

2

u

x

u

xxx

+

5

2

αµt

2

u

2

xx

Onetransferredtheterm

D

x

(

· · ·

)

from

C

0

to

C

1

,wendtheonservedvetor

C

= (C

0

, C

1

)

,where

C

0

=

2xu

+

αtu

2

,

C

1

=

αxu

2

+

2

3

α

2

tu

3

2µxu

xxxx

+ 2µu

xxx

+ 2αµtuu

xxxx

2αµtu

x

u

xxx

+

αµtu

2

xx

forthesimpliedKahawaraequation.

Now,substituting

v

=

u

into(5.21), itisobtained

C

0

=

4u

2

xuu

x

+ 5αtu

2

u

x

+ 5µtuu

xxxxx

,

C

1

=

xuu

t

4αu

3

12µuu

xxxx

5αtu

2

u

t

5µtu

t

u

xxxx

+ 12µu

x

u

xxx

+

(8)

andthen,

C

0

=

7

2

u

2

+

D

x

xu

2

2

+

5

3

αtu

3

+ 5µtuu

xxxx

5µtu

x

u

xxx

+

5

2

µtu

2

xx

,

C

1

=

7

3

αu

3

7µuu

xxxx

+ 7µu

x

u

xxx

7

2

µu

2

xx

D

t

xu

2

2

+

5

3

αtu

3

+ 5µtuu

xxxx

5µtu

x

u

xxx

+

5

2

µtu

2

xx

.

Afterastraightforwardalulationand multiplyingthenalresultby

7/2

,it

isobtainedthevetor(1.7).

5.2. Conservation law for the Gardner equation

Inorder to establishtheonservedvetor(1.8)for theGardner equation, werst

substitutetheomponentsofthefollowingLiepointsymmetrygenerator

X

3

= (2x

+ 6t)

∂x

+ 6t

∂t

(2u

+ 1)

∂u

intotheEq. (4.19)with

µ

= 0, α

=

β

= 6

and

γ

= 1

. Thus

C

0

=

6vt(6u

+ 6u

2

)u

x

6vtu

xxx

(2u

+ 1)v

(2x

+ 6t)vu

x

,

C

1

=

(2x

+ 6t)vu

t

(2x

+ 6t)vu

xxx

+(2u

+ 1)(6u

+ 6u

2

)v

+ (2u

+ 1)v

xx

+ (2x

+ 6t)v

xx

u

x

+6t(6u

+ 6u

2

)vu

t

+ 6tv

xx

u

t

4v

x

u

x

(2x

+ 6t)v

x

u

xx

6tv

x

u

xt

+ 6vu

xx

+ (2x

+ 6t)vu

xxx

+ 6tvu

xxt

.

(5.22)

Now,setting

v

=

u

into

(5.22)

,itisobtained

C

0

=

36tu

2

u

x

36tu

3

u

x

6tuu

xxx

2u

2

u

2xuu

x

6tuu

x

,

C

1

=

2xuu

t

+ 6tuu

t

+ 18u

3

+ 12u

4

+ 6u

2

+ 2uu

xx

+

u

xx

+ 36tu

2

u

t

+36tu

3

u

t

+ 6tu

t

u

xx

4(u

x

)

2

6tu

x

u

xt

+ 6uu

xx

+ 6tuu

xxt

,

whihareequivalentto

C

0

=

u

u

2

D

x

(xu

2

+ 3tu

2

+ 6tuu

xx

3t(u

x

)

2

+ 9tu

4

+ 12tu

3

),

C

1

=

3u

4

+ 6u

3

+ 3u

2

+ 2uu

xx

+

u

xx

(u

x

)

2

+D

t

(xu

2

+ 3tu

2

+ 6tuu

xx

3t(u

x

)

2

+ 9tu

4

+ 12tu

3

).

(9)

6. Conlusion

Inthis review paperwehaverevisited somereent resultsonnetingLie

symme-triesandonservationlawsforequationsnotneessarilybeingfromEuler-Lagrange

equations. Moreover,wehaveillustratedourdisussionwithsomenewonservation

lawsobtainedviaIbragimov'sapproah. Suhonservationlawsareestablished,at

leastusingsuhanewdevelopment,forthersttimein thepresentreview.

Theinterested readeran foundmore loal onservation laws forequations of

the type (1.1), where we established onservation lawsfor the generalKawahara

equation(1.3)andmodiedKawaharaequation(1.5). Moreover,someonservation

lawsfortheKdVequationanalsobeestablishedusingthoseobtainedresults,see

also [16℄. Morereently,somenewlassesof evolutionequations upto fthorder

havebeendisovered,see[7℄.

Resumo. Nestetrabalhorevisitamosalgunsresultadosreentessobreleisde

on-servaçãodeumalassedeequaçõesevolutivasatéquintaordem.

Palavras-have. Teorema de Ibragimov, leis deonservação, equaçõesKdV de

quintaordem.

Referenes

[1℄ G.W. Bluman, S. Kumei, Symmetries and Dierential Equations, Applied

MathematialSienes 81,Springer,NewYork,1989.

[2℄ I.L. Freire,Conservationlawsforself-adjointrstorder evolutionequations,J.

Nonlin. Math. Phys.,18(2011),279290.

[3℄ I.L.Freire,Self-adjointsub-lassesofthirdandfourth-orderevolutionequations,

Appl.Math. Comp.,217(2011),94679473.

[4℄ I.L. Freire, Newonservationlawsfor invisid Burgersequation, Comp. Appl.

Math., 31(2012),559567.

[5℄ I.L. Freire, J. C. S. Sampaio, Nonlinear self-adjointnessof a generalized

fth-orderKdVequation, J.Phys. A:Math.Theor., 45(2012),032001.

[6℄ I.L. Freire, J.C.S. Sampaio, Conservation laws forKawaharaequations, Anais

doXXXIVCNMAC,2012.

[7℄ I.L. Freire, Newlassesofnonlinearlyself-adjointevolutionequationsof

third-andfth-order,Commun. NonlinearSi. Numer.Simulat.,18(2013),493499.

[8℄ M.L.Gandarias,Weakself-adjointdierentialequations,J.Phys. A,44(2011),

262001.

[9℄ M.L. Gandarias, Weak self-adjointness and onservation laws for a porous

(10)

[10℄ M.L.Gandarias,M.Redondo,M.S.Bruzón,Someweakself-adjoint

Hamilton-Jaobi-Bellmanequationsarisinginnanialmathematis,Nonlin.Anal.RWA,

13(2012),340347.

[11℄ M.L.Gandarias,M.S.Bruzón,SomeonservationlawsforaforedKdV

equa-tion,Nonlin. Anal.RWA,13(2012),26922700.

[12℄ N.H.Ibragimov,Anewonservationtheorem,J.Math.Anal.Appl.,333(2007),

311328.

[13℄ N.H.Ibragimov,M.Torrisi,R.Trainà,Quasiself-adjointnonlinearwave

equa-tions,J. Phys. A:Math. Theor., 43(2010),442001-442009.

[14℄ N.H.Ibragimov,M.Torrisi,R.Trainà,Self-adjointnessandonservationlaws

ofageneralizedBurgersequation,J.Phys. A:Math.Theor., 44(2011),

145201-145206.

[15℄ N.H.Ibragimov,Nonlinearself-adjointnessandonservationlaws,J.Phys. A:

Math. Theor., 44(2011),432002432006.

[16℄ N.H.Ibragimov,Nonlinearself-adjointnessin onstrutingonservation laws,

ArhivesofALGA, 7/8(2011),190.

[17℄ N.H. Ibragimov, Transformation groups applied to mathematial physis,

TranslatedfromtheRussianMathematis anditsAppliations(Soviet Series),

D. ReidelPublishingCo.,Dordreht,1985.

[18℄ H. Liu, J. Li, L. Liu, Lie symmetry analysis, optimal systems and exat

so-lutions to the fth-order KdV types of equations, J. Math. Anal. Appl., 368

(2010),551558.

[19℄ R. Naz,F.M. Mahomed, D.P. Mason, Comparison of dierent approahes to

onservationlawsforsomepartialdierentialequationsinuidmehanis,Appl.

Math. Comput.,205(2008),212230.

[20℄ P.J.Olver,AppliationsofLiegroupstodierentialequations,Springer,New

York,(1986).

[21℄ J.C.S. Sampaio, Simetrias de Lie e leis de onservação de equações

evoluti-vas do tipo Korteweg-de Vries', Dissertação de Mestrado, Programa de

Pós-GraduaçãoemMatemátiaApliada,UFABC,SantoAndré-SP,2012.

[22℄ M.Torrisi,R.Trainà,Quasiself-adjointnessofalassofthirdordernonlinear

Referências

Documentos relacionados

An expression for interval-valued intuitionistic version of the Atanassov’s intu- itionistic fuzzy index is introduced in Section 3, in order to study properties of IvIFIs, which

Then, the main objective of the paper is to present the FGM long-term bivariate survival copula model assuming promotion time models as marginal distributions, and to de-

However, from our previous results regarding the bidimensional Lane-Emden systems (and it still holds to the multidimensional case), such a particular system naturally arise as

[r]

In the context of the present work, and based in a previous result (op. [13]), we extend for octonions the concept of a 2L-periodical function, that we will call Octonionic

Para finalizar, apresentamos o artigo da ´area de computac¸˜ao cient´ıfica: “A faster pixel-decimation. method for block motion estimation

In despite of several results in the literature to compute inverse solutions in the hyper- bolic heat transfer with a number of different methods, and recovery initial conditions on

What makes the SK algorithm more convenient in such cases is that, in addition to dividing the treatments in groups without overlapping, its result also uses a probabilistic