Abstract
We study symmetries, invariant solutions, and conservation laws for the dispersionless Veselov–Novikov equation. The emphasis is placed on cases when the odes involved in description of the invariant solutions are integrable by quadratures. Then we find some non-invariant solutions, in particular, solutions that are polynomials of an arbitrary degree \(N \ge 3\) with respect to the spatial variables. Finally we compute all conservation laws that are associated to cosymmetries of second order.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
We consider the dispersionless Veselov-Novikov equation (dVN) [11] written in the form
This equation describes the propagation of the high frequency electromagnetic waves in certain nonlinear media, see [12] and references therein. Nontrivial t-independent solutions of dVN are related to the problem of existence of the first integrals for the geodesic or magnetic geodesic flows on a two-torus, [23]. Equation (1) is the dispersionless reduction of the Nizhnik–Veselov–Novikov equation [17, 24]. The Lax representation [21]
of dVN is the dispersionless reduction of the Lax representation of the Nizhnik–Veselov–Novikov equation. In [2] the Lax representation (2) was used to construct two-dimensional reductions of dVN.
In the present paper we study exact solutions and conservation laws of dVN. We find the contact symmetry algebra and the explicit form for the transformations from the contact symmetry pseudogroup of dVN. Then we employ the pseudogroup to find the optimal system of one-dimensional subalgebras of the symmetry algebra. We factorize dVN with respect to the symmetries from the optimal system and obtain two-dimensional partial differential equations (pdes) (8) and (50) for the invariant solutions as well as their Lax representations. Then we find the symmetry algebras and their optimal systems of one-dimensional subalgebras for equations (8) and (50). The factorization with respect to the subalgebras provide the collection of ordinary differential equations (odes) that describe invariant solutions to (8) and (50). We find some cases when the obtained odes are integrable by quadratures, thus providing exact solutions for dVN. Further, we study solutions that are not invariant with respect to contact symmetries. In particular, we find a class of solutions to dVN that are polynomials in x and y of arbitrary degree. Finally we find the whole set of conservation laws that are associated to cosymmetries defined on the second order jets.
2 Preliminaries
The presentation in this section closely follows [13, 15, 25]. Let \(\pi :{\mathbb {R}}^n \times {\mathbb {R}} \rightarrow {\mathbb {R}}^n\), \(\pi :(x^1, \dots , x^n, u) \mapsto (x^1, \dots , x^n)\), be a trivial bundle, and \(J^\infty (\pi )\) be the bundle of its jets of infinite order. The local coordinates on \(J^\infty (\pi )\) are \((x^i,u,u_I)\), where \(I=(i_1, \dots , i_n)\) are multi-indices, and for every local section \(f :{\mathbb {R}}^n \rightarrow {\mathbb {R}}^n \times {\mathbb {R}}\) of \(\pi \) the corresponding infinite jet \(j_\infty (f)\) is a section \(j_\infty (f) :{\mathbb {R}}^n \rightarrow J^\infty (\pi )\) such that \(u_I(j_\infty (f)) =\displaystyle {\frac{\partial ^{\#I} f}{\partial x^I}} =\displaystyle {\frac{\partial ^{i_1+\dots +i_n} f}{(\partial x^1)^{i_1}\dots (\partial x^n)^{i_n}}}\). We put \(u = u_{(0,\dots ,0)}\). Also, we will simplify notation in the following way, e.g., in the case of \(n=3\): we denote \(x^1 = t\), \(x^2= x\), \(x^3= y\), and \(u_{(i,j,k)}=u_{{t \dots t}{x \dots x}{y \dots y}}\) with i times t, j times x, and k times y.
The vector fields
\((i_1,\dots , i_k,\dots , i_n)+1_k = (i_1,\dots , i_k+1,\dots , i_n)\), are called total derivatives. They commute everywhere on \(J^\infty (\pi )\) and are annihilated by the ideal of contact forms \(\langle du_I - \sum _{k=1}^{n} u_{I+1_{k}}\,dx^k \,\,\vert \,\, \#I \ge 0 \rangle \).
The evolutionary vector field associated to an arbitrary smooth function \(\varphi :J^\infty (\pi )\) \(\rightarrow \) \({\mathbb {R}}\) is defined as
with \(D_I=D_{(i_1,\dots \,i_n)} =D^{i_1}_{x^1} \circ \dots \circ D^{i_n}_{x^n}\).
A pde \(F(x^i,u_I) = 0\) of order \(s \ge 1\) with \(\# I \le s\) defines the submanifold \(\mathcal {E}=\{(x^i,u_I)\in J^\infty (\pi )\,\,\vert \,\,D_K(F(x^i,u_I))=0,\,\,\# K\ge 0\}\) in \(J^\infty (\pi )\).
A function \(\varphi :J^\infty (\pi ) \rightarrow {\mathbb {R}}\) is called a (generator of an infinitesimal) symmetry of equation \(\mathcal {E}\) when \({\mathbf {E}}_{\varphi }(F) = 0\) on \(\mathcal {E}\). The symmetry \(\varphi \) is a solution to the defining system
where \(\ell _{\mathcal {E}} = \ell _F \vert _{\mathcal {E}}\) with the differential operator
The symmetry algebra \({\mathrm {Sym}} (\mathcal {E})\) of equation \(\mathcal {E}\) is the linear space of solutions to (3) endowed with the structure of a Lie algebra over \({\mathbb {R}}\) by the Jacobi bracket \(\{\varphi ,\psi \} = {\mathbf {E}}_{\varphi }(\psi ) - {\mathbf {E}}_{\psi }(\varphi )\). The algebra of contact symmetries \({\mathrm {Sym}}_0 (\mathcal {E})\) is the Lie subalgebra of \({\mathrm {Sym}} (\mathcal {E})\) defined as \({\mathrm {Sym}} (\mathcal {E}) \cap C^\infty (J^1(\pi ))\). The point symmetries are the contact symmetries whose generators are polynomials in \(u_{x^i}\) of degree 1.
The contact symmetry pseudogroup of a pde \(\mathcal {E} \subset J^\infty (\pi )\) is the collection of all the local diffeomorphisms \(\varGamma _\infty :J^\infty (\pi ) \rightarrow J^\infty (\pi )\) that preserve the submanifold \(\mathcal {E}\) as well as the ideal of contact forms.
Let \(\phi \) be a symmetry of \(\mathcal {E}\), then the \(\phi \)–invariant solution of \(\mathcal {E}\) is the solution of the compatible system \(F= 0\), \(\phi =0\).
For a pde \(\mathcal {E}\) in three independent variables t, x, y a conservation law, [18, § 4.3], [25, Ch. 5], is a horizontal two-form
closed with respect to the horizontal differential \(d_h\), which means that
on \(\mathcal {E}\). Functions \(P_i\) are smooth functions on \(\mathcal {E}\). The conservation law is referred to as a trivial conservation law when there exists a horizontal one-form \(\omega \) such that \(\varOmega = d_h\omega \). Nontrivial conservation laws are associated with cosymmetries of equation \(\mathcal {E}\), see discussion in [14, Ch. 1]. Let \({\tilde{P}}_i\) be arbitrary extensions of \(P_i\) on \(J^{\infty }(\pi )\), then for \({\tilde{\varOmega }} = {\tilde{P}}_1 \,dx \wedge dy + {\tilde{P}}_2 \,dy \wedge dt + {\tilde{P}}_3 \, dt \wedge dx\) there holds \(d_h{\tilde{\varOmega }} = \psi \cdot F\,dt \wedge dx \wedge dy\) for some function \(\psi \) on \(J^\infty (\pi )\). The restriction \(\psi \vert _{\mathcal {E}}\) depends on \(\varOmega \) only and is called the generating function or characteristic of the conservation law \(\varOmega \). The conservation law is trivial if and only if its generating function vanishes. Generating functions are solutions to equation
with \(\ell _{\mathcal {E}}^{*} = \ell _{F}^{*}\vert _{\mathcal {E}}\), where the adjoint operator to \(\ell _F\) is
A solution of (4) is referred to as a cosymmerty of equation \(\mathcal {E}\).
3 Symmetries of dVN
3.1 The symmetry algebra
Direct computationsFootnote 1 show that the contact symmetry algebra \({\mathrm {Sym}}_0({\mathrm {dVN}})\) is generated by functions
where \(A= A(t)\) and \(B=B(t)\) below are arbitrary smooth functions of t. Actually, all the contact symmetries of (1) turn out to be point symmetries. The structure of the Lie algebra \({\mathrm {Sym}}_0({\mathrm {dVN}})\) is defined by the commutator table
Direct computations show that the contact symmetry pseudogroup of equation (1) is generated by the infinite prolongations of the (local) diffeomorphisms \(\varGamma _0 :(t,x,y,u) \mapsto ({\tilde{t}},{\tilde{x}},{\tilde{y}},{\tilde{u}})\) of the form
where \(\varepsilon \ne 0\), \(B_i=B_i(t)\) are arbitrary functions, \(B_i^{\prime } =\displaystyle {\frac{dB_i}{dt}}\), and \(B_0^{\prime }(t) \ne 0\). In other words, substitution for (5) into dVN written in the tilded variables yields (1).
3.2 The optimal system of one-dimensional subalgebras
Since the symmetry algebra of dVN is infinite-dimensional and depends on 6 arbitrary functions of one variable, the problem of examining all invariant solutions is very complicated. To overcome the difficulty, we use the following observation: transformations from the symmetry pseudogroup (5) preserve equation (1), while changing the symmetry generators. Therefore we can classify the orbits of the action of (5) on the \({\mathrm {Sym}}_0({\mathrm {dVN}})\). In order to use symmetries for computing invariant solutions we consider symmetries whose generators depend explicitly on at least one of the variables u, \(u_t\), \(u_x\), or \(u_y\).
Proposition 1
Each symmetry
from \({\mathrm {Sym}}_0({\mathrm {dVN}})\) with \(A_0^2 +\mu ^2 +A_{1,1}^2+A_{2,1}^2 \not \equiv 0\) is equivalent with respect to the action of the pseudogroup (5) to one of symmetries
Proof
Let \(A_0(t) \ne 0\). Put \(\varepsilon =1\) and consider solutions \(B_0(t)\), ... , \(B_5(t)\) to the system of odes
Direct computations show that for \({\tilde{\chi }}_1 =-{\tilde{u}}_{{\tilde{t}}}+\mu \,(3\,{\tilde{u}}-{\tilde{x}}\,{\tilde{u}}_{{\tilde{x}}} -{\tilde{y}}\,{\tilde{u}}_{{\tilde{y}}})\) there holds \(\varGamma _1^{*}({\tilde{\chi }}_1) =\varPhi \), where \(\varGamma _1\) is the first prolongation of (5). Therefore the inverse diffeomorphism maps \(\varPhi \) to \({\tilde{\chi }}_1\).
When \(A_0(t) \equiv 0\) and \(\mu \ne 0\), put \(\varepsilon = \mu ^{1/3}\) and define (5) by \(B_0=t\), \(B_1 = \mu ^{-1}\,A_{1,1}\), \(B_2 = \mu ^{-1}\,A_{2,1}\), \(B_3 = \frac{1}{2}\,\mu ^{-2}\,(\mu \,A_{1,2}-A_{1,1}\,A_{1,1}^{\prime })\), \(B_4 = \frac{1}{2}\,\mu ^{-2}\,(\mu \,A_{2,2}-A_{2,1}\,A_{2,1}^{\prime })\), \(B_5 = \frac{1}{6}\,\mu ^{-3}\,(2\,\mu ^2\,A_3+\mu \,(A_{1,1}\,A_{1,2}+A_{2,1}\,A_{2,2}) -A_{1,1}^2\,A_{1,1}^{\prime } -A_{2,1}^2\,A_{2,1}^{\prime })\). Then we have \(\varGamma _1^{*}({\tilde{\chi }}_2) =\varPhi \vert _{A_0=0}\).
Suppose now that \(A_0(t) \equiv 0\), \(\mu =0\), \(A_{1,1}(t) \ne 0\), and \(A_{2,1}(t) \ne 0\). Put \(\varepsilon =1\) and define functions \(B_i\), \(A({\tilde{t}})\) by equations \(B_0^{\prime } = A_{1,1}^{-3}\), \(B_1^{\prime } = A_{1,2}A_{1,1}^{-2}\), \(B_2^{\prime } = (A_{2,2}+(A_{1,1}A_{2,1}^{\prime }-A_{2,1}A_{1,1}^{\prime })\,B_2)\,A_{1,1}^{-1}A_{2,1}^{-1}\), \(B_3 = \frac{1}{2}\,(A_{1,1}A_{2,1}^{\prime }-A_{2,1}A_{1,1}^{\prime })\,B_2^2-A_3\,A_{1,1}^{-1}\), \(B_4 = 0\), and \(A(B_0(t))= A_{2,1}(t)/A_{1,1}(t)\). Then \(\varGamma _1^{*}({\tilde{\chi }}_3) =\varPhi \vert _{A_0=0,\mu =0}\).
Finally, when \(A_0(t) \equiv 0\), \(\mu =0\), \(A_{1,1}(t) \equiv 0\), and \(A_{2,1}(t) \ne 0\), put \(\varepsilon =1\) and define functions \(B_i\), \(A({\tilde{t}})\) by equations \(B_0^{\prime } = A_{2,1}^{-3}\), \(B_1 \equiv 0\), \(B_2 = A_{2,2}A_{2,1}^{-2}\), \(B_3 \equiv 0\), \(B_4 = -A_{3}\,A_{2,1}^{-1}\), and \(A(B_0(t))= A_{1,2}(t)\,A_{2,1}(t)\). Then \(\varGamma _1^{*}({\tilde{\chi }}_4) =\varPhi \vert _{A_0=0,\mu =0, A_{1,1} =0}\). \(\square \)
4 Invariant solutions
In this section we analyze reductions of dVN with respect to the symmetries \(\chi _1\), ... , \(\chi _4\).
4.1 Reduction w.r.t. \(\chi _1\)
The \(\chi _1\)–invariant solutions of dVN satisfy equation (1) and
so they are of the form
We introduce function
for convenience of the further computations. Substitution for (6), (7) into (1) shows that U(z, w) is a solution to equation \(\mathcal {E}_\mu \) defined by
This equation admits a Lax representation. Indeed, symmetry \(\chi _1\) has the lift \((\chi _1,{\hat{\chi }}_1)\) with \({\hat{\chi }}_1= -q_t+\mu \,\left( \frac{3}{2}\,q-x\,q_x-y\,q_y\right) \) to the Lax representation (2). Solutions to equation \({\hat{\chi }}_1=0\) are of the form \(q=\mathrm {e}^{\frac{3}{2}\,\mu \,t}Q(s,w)\). Substituting this into (2) yields the Lax representation
for equation (8). We put \(Q_s=S\) and obtain another Lax representation for (8):
Now we analyze invariant solutions to equation (8). We consider cases \(\mu \ne 0\) and \(\mu =0\) separately.
4.1.1 Case \(\mu \ne 0\).
When \(\mu \ne 0\), the symmetry algebra \({\mathrm {Sym}}_0(\mathcal {E}_\mu )\) of equation (8) is generated by functions \(\eta _1 = 3\,U-s\,U_s-w\,U_w\), \(\eta _2 = -U_s\), \(\eta _3 = -U_w\), \(\eta _4 = s\), \(\eta _5 = w\), \(\eta _6 =1\). The structure of this Lie algebra is defined by the commutators in Table 1.
The adjoint representation of the Lie group \(G_\mu \) associated with the finite-dimensional Lie algebra \({\mathrm {Sym}}_0(\mathcal {E}_\mu )\) is defined by the Lie series
where \({\mathrm {ad}}\,\eta _i (\eta _j) = \{\eta _i, \eta _j\}\). This representation is given by Table 2. In this table the (i, j)-th entry is the expression for \({\mathrm {Ad}}_{\tau \eta _i} (\eta _j)\). Using this table one can classify the orbits of action of the adjoint representation of the Lie group \(G_\mu \) on its Lie algebra \({\mathrm {Sym}}_0(\mathcal {E}_\mu )\):
Proposition 2
Each symmetry of equation \(\mathcal {E}_\mu \) with \(\mu \ne 0\) is equivalent under the action of the adjoint representation of \(G_\mu \) to one of the symmetries
where \(\alpha , \beta , \gamma \in {\mathbb {R}}\).
Proof
is obtained by the standard computation, see, e.g., [18, § 3.3]. \(\square \)
Thus each invariant solution to equation (8) can be obtained by an action of appropriate superposition of transformations \({\mathrm {Ad}}_{\tau \,\eta _i} (\eta _j)\) from \(\zeta _k\)–invariant solutions. Below we examine such solutions.
4.1.1.1. Solutions invariant w.r.t. \(\zeta _1\).
The \(\zeta _1\)–invariant solutions satisfy (8) and \(\zeta _1 =0 \), therefore they are of the form
where W is a solution to the ode
We could not obtain the general solution to this ode. Instead, we find a family of particular solutions given by
where \(a_1 a_2 \ne 0\). Then (6), (7), (9) give the t-independent solution
to dVN.
4.1.1.2. Solutions invariant w.r.t. \(\zeta _2\).
For \(\zeta _2\)–invariant solutions there holds \(\zeta _2=0\), or
When \(\alpha \ne 0\), this equation gives
Substituting for (12) into (8) yields
This equation is integrable by quadratures. When \(\beta = \alpha ^2\,\gamma \), \(\alpha \ne -1\), the general solution to (13) is
where \(c_0\), \(c_1\), \(c_2\) are arbitrary constants. This function produces solution
of dVN.
Remark 1
If u is a solution to (1), then for arbitrary functions \(b_0(t)\), \(b_1(t)\), \(b_2(t)\) the linear combination \(u+b_0(t)+b_1(t)\,x+b_2(t)\,y\) is a solution to dVN as well. So we can put \(c_0=c_1=0\) in (14) without loss of generality. In what follows we will ignore the linear in x and y terms in solutions of dVN. \(\square \)
When \(\beta \ne \alpha ^2\,\gamma \), \(\alpha \ne -1\), equation (13) has the first integral
hence the general solution in this case is of form
where function \(H(\tau )\) is defined by formula
This function can be expressed in terms of the Lambert W function, [6], while the expression is too complicated to be written here.
When \(\alpha =-1\), \(\beta \ne \gamma \), the general solution of (13) reads
to dVN of the form
cf. Remark 1.
Finally, when \(\alpha = 0\) we have solution
of equation (8). This produces the following solution to dVN:
4.1.1.3. Solutions invariant w.r.t. \(\zeta _3\)
The \(\zeta _3\)–invariant solution of equation (8) has the form
the corresponding solution of dVN is obtained from (18) by renaming \(x \leftrightarrow y\), \(\beta \leftrightarrow \gamma \).
Remark 2
For each solution \(u=f(t,x,y)\) of dVN expression \(u=f(t,y,x)\) defines a solution as well. \(\square \)
4.1.2 Case \(\mu =0\)
Now consider equation \(\mathcal {E}_0\) obtained by putting \(\mu =0\) in (8). Notice that solutions of this equation produce t-independent solutions to dVN, and t-independent solutions are defined up to a nonzero constant multiple. The symmetry algebra \({\mathrm {Sym}}_0(\mathcal {E}_0)\) of this equation has generators \(\xi _1 =-s\,U_s-w\,U_w\), \(\xi _2 =-U_s\), \(\xi _3 =-U_w\), \(\xi _4 =s\), \(\xi _5 =w\), \(\xi _6 =1\), and \(\xi _7 =U\), with the commutators given in Table 3.
The adjoint action of the symmetry group \(G_0\) of equation \(\mathcal {E}_0\) on \({\mathrm {Sym}}_0(\mathcal {E}_0)\) is defined by the Table 4. Then direct computations give the optimal system of one-dimensional subalgebras of \({\mathrm {Sym}}_0(\mathcal {E}_0)\):
Proposition 3
Each symmetry of equation \(\mathcal {E}_0\) is equivalent under the action of the adjoint representation of \(G_0\) to one of the following symmetries:
4.1.2.1. Solutions invariant w.r.t. \(\sigma _1\).
The \(\sigma _1\)–invariant solutions to \(\mathcal {E}_0\) have the form
Substituting this into (8) gives the ode
where
and
We did not find the general solution to this equation for any \(\alpha \). For each \(\alpha \) equation (19) admits solutions of the form
where \(z_0\) is a root of equation \(z_0\,(z_0^3+1) =0\). Thus we get solution
of equation (1). We found some other solutions of (19) for \(\alpha \in \{1, 2, 3, 4\}\).
Equation (19) with \(\alpha = 1\)
is integrable by quadratures, its general solution
produces solution to dVN of the form
Notice that integral in (22) can not be expressed in elementary functions, [3, § VIII], [22, Ch. 3, § 14].
When \(\alpha =2\), equation (19) reduces to \(W_{zzz}=0\), its general solution \(W=c_0+c_1\,z+c_2\,z^2\), \(c_i \in {\mathbb {R}}\), gives solution of dVN
Remark 3
Expression \(u = b_1(t) \, x^2 + c\, x \,y + b_2(t) \, y^2\) with arbitrary functions \(b_1(t)\), \(b_2(t)\) and constant c defines a trivial solution to dVN. We will ignore such solutions below. \(\square \)
For \(\alpha =3 \) we obtain two algebraic solutions
and
of (19). They give solutions
and
of dVN.
When \(\alpha =4\), equation (19) admits the polynomial solution \(W = 17\,z^4-36\,z^3 -90\, z^2 -36 \,z+17\). This corresponds to solution of dVN of the form
4.1.2.2. Solutions invariant w.r.t. \(\sigma _2\).
For \(\sigma _2\)–invariant solutions of \(\mathcal {E}_0\) there holds \(\sigma _2 =-U_w+\alpha \,U=0\), therefore these solutions have the form
Substituting this into \(\mathcal {E}_0\) gives the reduced equation \(V_s\,V_{sss} + V_{ss}^2+2\,\alpha ^3\,V\,V_s =0\). Integrating this twice, we obtain the first order ode
with arbitrary constants \(\gamma \) and \(\delta \). This equation is integrable by quadratures, and the general solution can be expressed in elliptic functions, [10, Ch. V, § 21]. To prove this claim, consider transformation
that maps the curve \(R^3=P^3+\gamma \,P+\delta \) onto the curve
with \(\beta = 6\,\delta \, \gamma ^{-3/2}\). We have \(2\,r\,dr = -2\,(\beta + 6\,p +6\,p^3)\,dp\), hence
Then (29), (6), (7), and \(\mu =0\) give solution
where function V is defined implicitly by equations
The integral in (34) includes Legendre’s integrals, therefore x can be expressed as an elliptic function of p.
When polynomial \(V^3 +\gamma \,V +\delta \) has multiple roots, the general solution to equation (30) can be expressed in elementary functions. Indeed, the presence of the triple root implies \(\gamma = \delta =0\), and then solution to (30) is \(V = c\,\mathrm {e}^{-\alpha \,s}\), \(c\in {\mathbb {R}}\). This corresponds to solution
of dVN. For a double root we have \(V^3+\gamma \,V+\delta =(V+2\,\varepsilon )\,(V-\varepsilon )^2 \) for some \(\varepsilon \ne 0\). Then we obtain solution to dVN of the form (32), where function V is defined in the implicit form by equations
with \(c \in {\mathbb {R}}\).
There are some other values of parameters \(\gamma \) and \(\delta \) when integral \(\int (V^3+\gamma \,V+\delta )^{-\frac{1}{3}}\,dV\) can be expressed by elementary functions. From the results of [4] (see also [9]) we obtain three following cases when function V(x) in (32) admits elementary expressions in implicit form:
(i) for \(\gamma = 0\), \(\delta \ne 0\) there holds
(ii) when \(\gamma \ne 0\), \(\delta = 0\), we get
(iii) conditions \(\gamma ^3 =-6\,\delta ^2 \ne 0\) imply
where
with \(T=(\tau ^2+\gamma \,\tau +\delta )^{\frac{1}{3}}\).
4.1.2.3. Solutions invariant w.r.t. \(\sigma _3\), \(\sigma _4\), \(\sigma _5\), and \(\sigma _6\).
For symmetries \(\sigma _3\), \(\sigma _4\), \(\sigma _5\), and \(\sigma _6\) the reduced odes are of the form \(W_{zzz}=0\). This gives only trivial solutions to dVN, cf. Remark 3.
4.1.2.4. Solutions invariant w.r.t. \(\sigma _7\), \(\sigma _8\), \(\sigma _9\), and \(\sigma _{10}\).
For symmetries \(\sigma _7\), ..., \(\sigma _{10}\) we found neither general nor particular solutions of the reduced equations. Below we write the forms of the invariant solutions of dVN that include functions W and reduced odes that define these functions.
For \(\sigma _7\) we have
where W(z) satisfies
The \(\sigma _8\)–invariant solutions are
where for W(z) there holds
Symmetry \(\sigma _9\) produces the invariant solution
with the reduced ode
Finally, for \(\sigma _{10}\) we get
with the defining equation for W(z) of the form
4.2 Reduction of dVN w.r.t. \(\chi _2\)
The \(\chi _2\)–invariant solutions of dVN satisfy \(\chi _2 = 3\,u-x\,u_x-y\,u_y =0\), therefore they have the form \(u=x^3\,U(t,z)\), \(z =y\,x^{-1}\), where U is a solution to
This equation admits a Lax representation. To show this, consider the lift \((\chi _2, {\hat{\chi }}_2)\) of \(\chi _2\) to the Lax representation (2) with \({\hat{\chi }}_2 = \frac{3}{2}\,q-x\,q_x-y\,q_y\). Solutions to \({\hat{\chi }}_2= 0\) have the form \(q= x^{\frac{3}{2}}\,Q(t,z)\). Substituting this into (2) yields
For \(S=Q_z\) we obtain another Lax representation
where
The symmetry algebra of equation (50) is generated by the family \(A(t)\,U_t+A^{\prime }(t)\,U+\frac{1}{18}\,A^{\prime \prime }(t)\,(z^3+1)\) with arbitrary function A(t). This symmetry provides invariant solutions to (50) of the form \(U=\frac{1}{18}\,(W(z)-A^{\prime }(t)\,(z^3+1))\,A(t)^{-1}\), where W is a solution to ode (19) with \(\alpha =3\). Hence solutions (20), (24), (25) of (19) generate the following solutions of dVN:
and
with \(c_1\,c_2 \ne 0\).
4.3 Reduction of dVN w.r.t. \(\chi _3\)
For \(\chi _3\)–invariant solutions we have \( u_x+A\,u_y+\frac{1}{2}\,A^{\prime }\,y^2 =0\), thus such solutions have the form
Substituting this into (1) and denoting \(V= W_{zz}\), we get the nonlinear pde of first order
We did not find the general solution to (55), instead we obtain three families of particular solutions for this pde.
First, we note that substitution \(V= - \frac{1}{2}\,A^{\prime }A^{-1}\,(A^3-1)^{-1}\,z\) reduces (55) to the ode \(A^{\prime \prime }= (2\,A^3+1)\,(A^{\prime })^2\,A^{-1}\,(A^3-1)^{-1}\) for function A. Integrating this when \(A \ne {\mathrm {const}}\), we get solution of dVN of the form
where function A(t) is defined implicitly by equation
and \(c_0\), \(c_1\ne \) are constants.
When \(A (t) \equiv \varepsilon \in {\mathbb {R}} \backslash \{1\}\), equation (55) gets the form \(V_t +2\,(\varepsilon ^3-1)\,V\,V_z =0\). The (multi-valued) solutions of this equation have the form \(2\,(\varepsilon ^3-1)\,t\,V + G(V) = z\), where G is an arbitrary function of one variable. They produce the family of solutions
where function W is defined implicitly by ode
When \(A (t) \equiv 1\), we have \(W_{tzz} = 0\), and hence
with arbitrary functions H, \(p_0\), \(p_1\) of one variable. We can put \(p_0 \equiv p_1 \equiv 0\) without loss generality in accordance with Remark 3.
4.4 Reduction of dVN w.r.t. \(\chi _4\)
Solutions of dVN that are invariant with respect to \(\chi _4\) satisfy equation \(u_y = A(t) \,x\) and therefore they have the form \(u=A(t)\,x\,y+W(t,x)\). When \(A \ne 0\), substituting this into (1) gives \(W_{xxx} = A^{\prime }A^{-1}\), so we have solutions of dVN in the form
with arbitrary function \(p_2\).
Finally, for \(A \equiv 0\), equation dVN is satisfied identically, so
is a solution to dVN for arbitrary function W of two variables.
5 Non-invariant solutions
Some of the solutions obtained in Sect. 4 admit natural generalizations that are not invariant with respect to symmetries from \({\mathrm {Sym}}_0({\mathrm {dVN}})\).
For example, solution (62) is a particular case of the family of solutions \(u = f(t,x)+g(t,y)\) with arbitrary functions f and g. The \(\sigma _2\)– invariant solutions from subsection 4.1.2.2 are included is the set of separable solutions of the form \(u=X_{\alpha }(x)\,Y_{-\alpha }(y)\), where \(X_{\alpha }\) and \(Y_{\alpha }\) are independently defined by either one of systems (33)–(34), (36)–(37), (40)–(41), or equation (38), or (39).
Nontrivial polynomial solutions (11), (14), (18), (27), (28), (51), (53), (56)–(57), (61) lead to idea to consider solutions of the form
Below we present such solutions for \(N \ge 3\).
5.1 \(N=3\)
Substituting (63) into (1) yields system of odes
for unknown functions \(T_{21}\), \(T_{12}\), \(T_{11}\) and arbitrary functions \(T_{30}\), \(T_{03}\), \(T_{20}\), \(T_{02}\) of t.
Solutions (11), (14), (18), (27), (51), (53), (56)–(57), (61) are particular cases of solutions given by system (63)–(64) for appropriate choices of functions \(T_{30}\), \(T_{03}\), \(T_{20}\), \(T_{02}\).
Analysis of system (64) gives two families of solutions to dVN. The first one reads
where \(T_{21}\), \(T_{12}\), \(T_{11}\), and \(T_{02}\) are arbitrary functions of t such that \(T_{21} \ne 0\) and \(T_{12} \ne 0\). When \(T_{21} \equiv T_{12} \equiv 0\), we obtain solution
with arbitrary functions \(T_{11}\), \(T_{30}\), \(T_{20}\), \(T_{02}\) of t such that \(T_{11} \ne 0\).
5.2 \(N=4\)
Substituting (63) with \(N=4\) into (1) and analyzing the resulting system we get four families of solutions of dVN:
Solution (68) with \(T \equiv 1\) coincides with solution (28).
5.3 \(N=5\)
For \(N =5\) we obtain the family of solutions
where \(T_{50}\) is an arbitrary nonzero function of t, while \(T_{30}\), \(T_{03}\), and \(T_{11}\) satisfy the following system of odes:
5.4 \(N \ge 6\)
For each \(N \ge 6\) we find the family of solutions
where \(T_N \), \(T_{N-1}\), and \(T_{20}\) are arbitrary functions of t such that \(T_N \ne 0\), \(T_{N-1} \ne 0\), while functions \(T_k\) with \(k \in \{4, \dots N-2\}\), \(T_{30}\), \(T_{03}\), and \(T_{11}\) obey the following system of odes:
6 Cosymmetries and conservation laws
Cosymmetries of equation (1) are solutions \(\psi \) to equation (4), which is of the form
Direct computations show that each solution \(\psi \in C^\infty (J^2(\pi ))\) of equation (75) is a linear combination of the following cosymmetries:
Here \(A=A(t)\) are arbitrary (smooth) functions.
The conservation laws \(\varOmega _1\), ... , \(\varOmega _7\) associated to cosymmetries \(\psi _1\), ... , \(\psi _7\) are given by the formulas
where
7 Concluding remarks
The results of the paper can be summarized as follows. Employing the methods of the Lie symmetry group analysis we have found a number of exact solutions for the dispersionless Veselov–Novikov equation, including solutions in elementary or elliptic functions (17), (33)–(34), (36)–(37), (38), (39), (40)–(41), (52), or functions represented by quadratures (15)–(16), (22)–(23), (58)–(59). We have indicated ordinary differential equations (10), (19), (43), (45), (47), (49) that describe all other invariant solutions. We have studied some non-invariant solutions and found a broad set of polynomial solutions (65)—(74). Furthermore, we have presented all the local conservation laws of order up to two.
While odes (10), (19), (43), (45), (47), (49) are not integrable by quadratures in general, their origin as reductions of Lax-integrable pdes (8) and (50) allows one to hope that the method of prolongation structures in the version implemented in [16] could be applicable to examine these odes. Likewise, the methods of weak symmetries [19, 20], nonclassical symmetry reductions, see [5] and references therein, conditional symmetries [7, 8], or stable-range approach [26], can be useful to generate new non-invariant solutions of the dispersionless Veselov–Novikov equation. We intend to address these issues in our future work.
Data Availability
The authors confirm that the data supporting the findings of this study are available within the article.
Notes
Computations of symmetries, their commutators, cosymmetries, and conservation laws were supported by the Jets software [1].
References
Baran, H., Marvan, M.: Jets: A software for differential calculus on jet spaces and diffieties. Available on-line at http://jets.math.slu.cz
Bogdanov, L.V., Konopelchenko, B.G., Moro, A.: Symmetry constraints for real dispersionless Veselov-Novikov equation. J. Math. Sci. 136(6), 4411–4418 (2006)
Chebyshev, P.L.: (Tchebychef). Sur l’intégration des différentielles irrationelles. Journal de mathématiques pures at appliquées. I. serie, XVIII (1853), 87-111, Œuvres T. I, pp. 142–168, Reprint. Chelsea, N.Y., (1961)
Chebyshev, P.L.: (Tchebychef). Sur l’intégration des différentielles les plus simples parmi celles qui contiennent une racine cubique. Matem. Sbornik II (1867), 71–78, Œuvres T. II, pp. 41–47, Reprint. Chelsea, N.Y., (1961)
Clarkson, P.A.: Nonclassical symmetry reductions of nonlinear partial differential equations. Math. Comput. Model. 18, 45–68 (1993)
Corless, R.M., Gonnet, G.H., Hare, D.E.G., Jeffrey, D.J., Knuth, D.E.: On the Lambert W function. Adv. Comput. Math. 5, 329–359 (1996)
Fushchich, W.I., Shtelen, W.M., Serov, N.I.: Symmetry Analysis and Exact Solutions of the Equations of Mathematical Physics. Kluwer, Dordrecht (1993)
Fushchich, W., Tsyfra, I.: On reduction and solutions of nonlinear wave equation with broked symmetry. J. Phys. A, Math. Gen. 20, 45–60 (1987)
Golubev, V.V.: P.L. Chebyshev’s works on integration of algebraic functions. Scientific Heritage of P.L. Chebyshev. First Issue, Mathematics. Academy of Science Publishing, Moscow, 1945, pp. 88–121 (in Russian)
Hardy, G.H.: The integration of Functions of a Single Variable, 2nd edn. Cambridge University Press, Cambridge (1916)
Konopelchenko, B., Martínez Alonso, L.: Nonlinear dynamics on the plane and integrable hierarchies of infinitesimal deformations. Stud. Appl. Math. 109, 313–336 (2002)
Konopelchenko, B., Moro, A.: Integrable equations in nonlinear geometrical optics. Stud. Appl. Math. 113, 325–352 (2004)
Krasil’shchik, J., Verbovetsky, A.: Geometry of jet spaces and integrable systems. J. Geom. Phys. 61, 1633–1674 (2011)
Krasil’shchik, J., Verbovetsky, A., Vitolo, R.: The Symbolic Computation of Integrability Structures for Partial Differential Equations. Springer, Berlin (2017)
Krasil’shchik, I.S., Vinogradov, A.M.: Nonlocal trends in the geometry of differential equations: symmetries, conservation laws, and Bäcklund transformations. Acta Appl. Math. 15, 161–209 (1989)
Morris, H.C., Dodd, R.K.: Infinite dimensional Lie algebras and the direct determination of deformation problems for equations of Painlevé type. Proc. Roy. Irish Acad. 83A, 127–143 (1983)
Nizhnik, L.P.: Integration of multidimensional nonlinear equations by the method of the inverse problem. Sov. Phys. Dokl. 25, 706–708 (1981)
Olver, P.J.: Applications of Lie Groups to Differential Equations, 2nd edn. Springer, Berlin, New York, Heidelberg (2000)
Olver, P.J., Rosenau, Ph: The construction of special solutions to partial differential equations. Phys. Lett. A 114, 107–112 (1986)
Olver, P.J., Vorob’ev, E.M.: Nonclassical conditions symmetries of partial differential equations. CRC Handbook of Lie Group Analysis of Differential Equations, V.3, Modern Trends, CRC Press, Boca Raton, pp. 291–328 (1996)
Pavlov, M.V.: Modified dispersionless Veselov–Novikov equation and corresponding hydrodynamic chains. arXiv:nlin/0611022
Ritt, J.F.: Integration in Finite Terms. Liouville’s Theory of Elementary Methods. Columbia University Press, New York (1948)
Taimanov, I.A.: On first integrals of geodesic flows on a two-torus. Proc. Steklov Inst. Math. 275, 225–242 (2016)
Veselov, A.P., Novikov, S.P.: Finite-zone two-dimensional potential Schrödinger operators. Explicit formulas and evolution equations. Sov. Math. Dokl. 30, 588–591 (1984)
Vinogradov, A.M., Krasil’shchik, I.S. (eds.): Symmetries and Conservation Laws for Differential Equations of Mathematical Physics [in Russian], Moscow: Factorial, 2005; English transl. prev. ed.: I.S. Krasil’shchik, A.M. Vinogradov (eds.) Symmetries and Conservation Laws for Differential Equations of Mathematical Physics. Transl. Math. Monogr., 182, Amer. Math. Soc., Providence, RI, (1999)
Xiaoping, Xu: Algebraic Approaches to Partial Differential Equations. Springer, Berlin (2013)
Acknowledgements
The work of O.M. was partially supported by the Faculty of Applied Mathematics of AGH UST statutory tasks within subsidy of Ministry of Science and Higher Education (Poland).
The work of J.-H.C. was partially supported by the Ministry of Science and Technology (Taiwan), research project number MOST 109-2115-M-606-001.
The authors are very grateful to M.V. Pavlov for stimulating discussions and important comments.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Ethical approval
The authors declare that they have adhered to the ethical standards of research execution.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Morozov, O.I., Chang, JH. The dispersionless Veselov–Novikov equation: symmetries, exact solutions, and conservation laws. Anal.Math.Phys. 11, 126 (2021). https://doi.org/10.1007/s13324-021-00563-8
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s13324-021-00563-8