JFSAJournal of Function Spaces and Applications0972-68022090-8997Hindawi Publishing Corporation78038210.1155/2012/780382780382Research ArticleOn the Existence of Variational Principles for a Class of the Evolutionary Differential-Difference EquationsKolesnikovaI. A.SavchinV. M.StepanovV.1University of Russia, Miklukho-Maklaya Street 6, Moscow 117198Russiamsu.ru2012412012201210032011181020112012Copyright © 2012 I. A. Kolesnikova and V. M. Savchin.This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Necessary and sufficient conditions for the existence of variational principles for a given wide class of evolutionary differential-difference operator equation are obtained. The theoretic results are illustrated by two examples.

1. Introduction

We consider the equation:N(u)=0,uD(N),

where D(N) is a domain of the definition of the operator N:D(N)UV, and U,V are normed linear spaces over the field of real numbers .

Later, we will assume that at every point uD(N), there exists the Gâteaux derivative Nu of N defined by the formula (d/dϵ)N(u+ϵh)|ϵ=0=δN(u,h)=Nuh.

The operator N:D(N)V is said to be potential  on the set D(N) relative to a given bilinear form Φ(·,·):V×UR, if there exists a functional FN:D(FN)=D(N)R such thatδFN[u,h]=Φ(N(u),h),uD(N),  hD(Nu).

In that case, we also say that the given equation admits the direct variational formulation.

A problem of the construction of the functional FN upon the given operator N is known as the classical inverse problem of the calculus of variations . Note that practically no one has been solving inverse problem of the calculus of variations for partial differential-difference operators until recently . Let us also note that for a wide classes of partial differential equations, there has been developped the problem of recongnition of variationality upon the structure of corresponding operators [5, 6]. There is a theoretical and practical interst in the extention of these results on partial differential-difference equations .

For what follows, we suppose that D(N) is a convex set, and we also need the following potentiality criterion : Φ(Nuh,g)=Φ(Nug,h),uD(N),  h,gD(Nu). Under this condition, the potential FN is given byFN[u]=01Φ(N(u0+λ(u-u0)),u-u0)dλ+const,

where u0 is a fixed element of D(N).

The functional FN is called the potential of the operator N, and in turn the operator N is called the gradient of the functional FN.

2. Statement of the Problem

Let us consider the following differential-difference operator equation: N(u)λ=-11Pλ(t)ut(t+λτ)+Q(t,u(t+λτ))=0,uD(N),  t[t0,t1]R.

Here, Pλ:U1V1  (λ=-1,0,1) are linear operators depending on t;Q:[t0-τ,t1+τ]×U13V1 is in general a nonlinear operator;  N:D(N)UV;  U  =C1([t0-τ,t1+τ];U1);V=C([t0-τ,t1+τ];V1), where U1,V1 are real normed linear spaces, U1V1.

The domain of definition D(N) is given by the equality: D(N)={uU:u(t)=φ1(t),t[t0-τ,t0],  u(t)=φ2(t),  t[t1,t1+τ]}, where φi,  (i=1,2) are given functions.

Under the solution of (2.1), we mean a function uD(N) satisfying the identity: N(u)λ=-11Pλ(t)ut(t+λτ)+Q(t,u(t+λτ))=0,t[t0,t1]R.

Let us give the following bilinear form: Φ(,)t0t1,dt:V×UR, where the bilinear mapping Φ1  ·,· satisfies the following conditions: v1,v2=v2,v1,v1,v2V1,Dtv,g=Dtv,g+v,Dtg,v,gV.

Our aim is to define the structure of operators Pλ  (λ=-1,0,1) and Q under which (2.1) allows the solution of the inverse problem of the calculus of variations relative to the bilinear form (2.4) such that Dt=d/dt is the antisymmetric operator on D(Nu), that is, Φ(Dth1,h2)=-Φ(Dth2,h1),h1,h2D(Nu).

3. Conditions of Potentiality and the Structure of (<xref ref-type="disp-formula" rid="EEq0.1">2.1</xref>) in the Case of Its Variationality

We denote by K* the operator adjoint to K.

Theorem 3.1.

If Dt*=-Dt on the set D(Nu), then for the existence of the direct variational formulation for the operator (2.1) on the set D(N) relative to (2.4), it is necessary and sufficient that the following conditions hold on the set D(Nu): P-λ+Pλ*|tt-λτ=0,Pλ*t|tt-λτ+Qu(t-λτ)-Qu(t+λτ)*|tt-λτ=0,λ=-1,0,1,uD(N),t(t0,t1).

Proof.

Taking into account formula (2.1), we get Nuh=λ=-11Pλht(t+λτ)+λ=-11Qu(t+λτ)h(t+λτ). The criterion of potentiality takes the following form: t0t1λ=-11{Pλht(t+λτ)+Qu(t+λτ)h(t+λτ)},g(t)dt=t0t1λ=-11{Pλgt(t+λτ)+Qu(t+λτ)g(t+λτ)},h(t)dt, or t0t1[λ=-11{Pλht(t+λτ)+Qu(t+λτ)h(t+λτ)},g(t)-λ=-11(Pλgt(t+λτ)+Qu(t+λτ)g(t+λτ)),h(t)]dt=0,uD(N),  g,hD(Nu). Bearing into account the condition Dt*=-Dt on the set D(Nu), from (3.5), we get t0t1[λ=-11(PλDt+Qu(t+λτ))h(t+λτ),g(t)-λ=-11[-Dt(Pλ*h(t))+Qu(t+λτ)*h(t)],g(t+λτ)]dt=0, or t0t1[λ=-11(P-λDt+Qu(t-λτ))h(t-λτ),g(t)-λ=-11-Pλ*t-Pλ*Dt+Qu*|tt-λτh(t-λτ),g(t)]dt=0. Thus, condition (3.5) can be reduced to the following form: t0t1λ=-11(P-λDt+Qu(t-λτ))h(t-λτ)-(-Pλ*t-Pλ*Dt+Qu*)|tt-λτh(t-λτ),g(t)dt=0,uD(N),g,hD(Nu). This equality is fulfilled identically if and only if λ=-11[(P-λ+Pλ*|tt-λτ)Dt+Pλ*t|tt-λτ+Qu(t-λτ)-Qu(t+λτ)*|tt-λτ]h(t-λτ)=0, for all uD(N). Thus, it is necessary and sufficient that conditions (3.1) hold.

Theorem 3.2.

Conditions (3.1) are held if and only if (2.1) has the following form: N1(u)λ=-11(Rλ*|tt-λτ-R-λ)ut(t-λτ)+(gradΦB[u]-λ=-11Rλtu(t+λτ))=0,uD(N1),t[t0,t1]R.

The operators Rλ and B depend on Pλ(t) and Q(t,u(t+λτ)).

Proof.

If Dt*=-Dt on the set D(N1u) and conditions (3.1) are held, then according to Theorem 3.1, operator (2.1) is potential on the set D(N) relative to a given bilinear form (2.4).

Let us consider the following functional:FN[u]=t0t1[λ=-11Rλ(t)u(t+λτ),ut(t)+B[u]]dt+FN[u0]. It is easy to check that δFN[u,h]=t0t1λ=-11Rλh(t+λτ),ut(t)+λ=-11Rλu(t+λτ),ht(t)+gradΦB[u],h(t)dt=t0t1λ=-11Rλ*|tt-λτut(t-λτ),h(t)-λ=-11Dt(Rλu(t+λτ)),h(t)+gradΦB[u],h(t)dt=t0t1λ=-11Rλ*|tt-λτut(t-λτ),h(t)-λ=-11Rλtu(t+λτ),h(t)-λ=-11R-λut(t-λτ),h(t)+gradΦB[u],hdt=t0t1λ=-11(Rλ*|tt-λτ-R-λ)ut(t-λτ)+(gradΦB[u]-λ=-11Rλtu(t+λτ)),h(t)dt=t0t1N(u),hdt,uD(N),  hD(Nu). Then functional (3.11) is a potential of evolutionary operator (2.1).

This theorem shows the structure of the given kind of differential-difference operator, which admits the solution of the inverse problem of the calculus of variations.

4. ExamplesExample 4.1.

Let us consider the evolutionary differential-difference equation with partial derivatives in the following form: N1(u)λ=-11(aλ(x,t)ut(x,t+λτ)-bλij(x,t)2uxixj(x,t+λτ))=0, where (x,t)Q=Ω×(t0,t1),  t1-t0>2τ,  i,j=1,n¯,aλ(x,t)Cx,t0,1(Q¯),  bλij(x,t)Cx,t2,0(Q¯).

Ω is a bounded domain in n with piecewise smooth boundary Ω.

The domain of definition D(N1) is given by the equality: D(N1)={μunxμ|Γτ=ψμ(x,t),  uU=Cx,t2,1(Ω¯×[t0-τ,t1+τ]):u(x,t)=φ1(x,t),  (x,t)E1=Ω¯×[t0-τ,t0],u(x,t)=φ2(x,t),  (x,t)E2=Ω¯×[t1,t1+τ],μunxμ|Γτ=ψμ(x,t),  μ=0,1}, where Γτ=Ω×[t0-τ,t1+τ], φiC(Ei),  (i=0,1),  ψμC(Γτ),  (μ=0,1) are given functions.

We investigate the existence of variational principle for (4.1) relative to a given bilinear form (2.4).For (4.1), we getN1uh=λ=-11(aλ(x,t)ht(x,t+λτ)-bλij(x,t)hxixj(x,t+λτ)).

We investigate the existence of variational principle for (4.1) relative to a given bilinear form (2.4).

For (4.1), we get N1uh=λ=-11(aλ(x,t)ht(x,t+λτ)-bλij(x,t)hxixj(x,t+λτ)).

Necessary and sufficient conditions of potentiality take the form:   Φ(N1uh,g)-Φ(N1ug,h)-t0t1Ω{(λ=-11Dt(aλ(x,t)g(x,t))λ=-11|tt-λτ+(λ=-11Dxixjbλij(x,t)g(x,t))|tt-λτh(x,t)}dxdt-t0t1Ωλ=-11{aλ(x,t)gt(x,t+λτ)-bλij(x,t)gxixj(x,t+λτ)}h(x,t)dxdt=0,uD(N1),  g,hD(N1u).

From that, we come to the following: -λ=-11(Dt(aλg(x,t))-Dxixj(bλijg(x,t)))|tt-λτ=λ=-11{aλgt(x,t+λτ)-bλijgxixj(x,t+λτ)},gD(N1u).

That is true if and only if -a-λ(x)=aλ(x),λ=1,-1,a0(x)=0,bλij(t)|tt-λτ=b-λij(t),λ=-1,0,1,i,j=1,n¯,(x,t)Q.

Under the fulfilment of that conditions, the corresponding functional is given by FN1[u]=12t0t1Ω{a1(x)ut(x,t-τ)u(x,t)-a1(x)ut(x,t+τ)u(x,t)+b1ij(t-τ)uxi(x,t-τ)uxj(x,t)+b0ij(t)uxi(x,t)uxj(x,t)+b-1ij(t+τ)uxi(x,t+τ)uxj(x,t)}dxdt+const.

Let us consider an example when this criterion of potantiality fails.

Example 4.2.

Consider the equation: N2(u)=ut(t,x)-2u(t,x+2τ)ux(t,x+2τ)-2ux(t,x)u(t,x-2τ)-2u(t,x)ux(t,x-2τ)+12uxxx(t,x+2τ)+12uxxx(t,x-2τ)=0,(t,x)QT=(0,T)×(-,+). Let us note that this equation is a Korteweg-de Vries’ equation τ=0.

We denote D(N2)={uU:Ct,x1,3((0,T)×(-,+)):u|t=0=u0(x),u|t=T=u1(x),lim|x|+Dxnu=0  (n=0,3¯)}. It is easy to be convinced that operator (4.7) is not potential on set (4.8) relative to the bilinear form: Φ(v,g)=0T-+v(t,x)g(t,x)dxdt,vV,  gU. Here, V={vCt,x0,3((0,T)×(-,+)):lim|x|+Dxnv=0,  (n=0,3¯)}.

We define the integrating operator M as Mv(t,x)=-xv(t,y)dy.

Then, the operator, Ñ(u)MN2(u)=-xut(t,y)dy-u2(t,x+2τ)-2u(t,x)u(t,x-2τ)+12uxx(t,x+2τ)+12uxx(t,x-2τ)dy

is potential on set (4.8) relative to bilinear form (4.9). The corresponding functional FN2[u] has the form: FN2[u]=120T-+{u(t,x)-xut(t,y)dy-ux(t,x-τ)ux(t,x+τ)-2u2(t,x+τ)u(t,x-τ)dy}dxdt.

Indeed, using (4.11), we find that δFN2[u,h]=0T-+{-xut(t,y)dy+12uxx(t,x+2τ)+12uxx(t,x-2τ)-2u(t,x)u(t,x-2τ)-u2(t,x+2τ)dy-+{-xut(t,y)dy+12uxx(t,x+2τ)+12uxx(t,x-2τ)-2u(t,x)u(t,x-2τ)}h(t,x)dxdt,uD(N),  hD(Nu).

From the condition δFN[u,h]=0,  uD(N), for all hD(Nu), we obtain Ñ(u)MN2(u)-xut(t,y)dy-u2(t,x+2τ)-2u(t,x)u(t,x-2τ)+12uxx(t,x+2τ)+12uxx(t,x-2τ)dy=0,  uD(N),

This equation is equivalent to (4.7).

Let us note that the formula I[u]=-+(1/2)ux(t,x-τ)ux(t,x+τ)+u2(t,x+τ)u(t,x-τ)dx defines the first integral I[u]=const of (4.7).

Acknowledgment

The work of V. M. Sovchin was partially supported by the Russian Foundation for Basic Research (Projects 09-01-00093 and 09-01-00586).

FilippovV. M.SavchinV. M.ShorokhovS. G.Variational principles for nonpotential operatorsJournal of Mathematical Sciences1994683275398127002210.1007/BF01252319KolesnikovaI. A.PopovA. M.SavchinV. M.On variational formulations for functional differential equationsJournal of Function Spaces and Applications200751891012296014ZBL1139.34046SavchinV. M.Helmholtz’s conditions of potentiality for PDE with deviating argumentsProceedings of the 35th Scientific Conference of the Departement of Physico-Mathematical and Natural ScienceMay 1994M.: RUDN25Part 2. Mathematical SectionsPopovA. M.Potentiality conditions for differential-difference equationsDifferentsial'nye Uravneniya19983434224241668242SavchinV. M.BudochkinaS. A.On the structure of a variational equation of evolution type with a second derivative with respect to t-derivativeDifferentsial'nye Uravneniya200339111812410.1023/A:10251843117012130302SavchinV. M.An operator approach to Birkhoff’s equationVestnik RUDN199522111123El’sgol’tzL. E.Qualitative Methods in Mathematical Analysis19641American Mathematical SocietyTranslations of Mathematical Monographs0170048SkubachevskiiA. L.Elliptic Functional-Differential Equations and Applications1997Basel, SwitzerlandBirkhäuser1437607KamenskiĭG. A.Variational and boundary value problems with deviating argumentDifferentsial'nye Uravneniya1970698134913580273162