A lattice version of the KP equation (Q1200351)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | A lattice version of the KP equation |
scientific article |
Statements
A lattice version of the KP equation (English)
0 references
16 January 1993
0 references
Let \(C\) be a compact Riemann surface defined over \(\mathbb{R}\) and \(P \in C\) a point defined over \(\mathbb{R}\). Let us consider \(H^ 1 (C,\mathbb{Z})\) as a subgroup of \(H^ 0 (C, \Omega)^*\) (the dual space of the holomorphic 1-forms on \(C)\). Denote by \(\Lambda^ +\) the elements of \(H^ 1 (C, \mathbb{Z})\) fixed by the natural antiholomorphic involution \(\iota\). The author considers special sets \(\{v_ 1, v_ 2, v_ 3\} \subset \Lambda^ +\) (which are called adapted and acceptable) and special kinds of pairs \((C,P)\) (called good). Let \(\vartheta\) be defined on \(H^ 0 (C, \Omega)^*\), the associated theta function, that can be chosen to be real on \(H^ 0 (C, \Omega)^* (\mathbb{R})\). A part of the article is devoted to the explicit construction (starting with \((C,P))\) of two quasi-homogeneous real polynomials \(f_{N,M}\) and \(g_{N,M}\), in five variables and weights four and five, respectively (here \(N,M\) are relatively prime integers) and to investigate solutions \(A(t)\) and \(B(t)\) of the following differential-difference equations: \[ {dA(t) \over dt} = f_{N,M} \bigl( A(t), B(t), \alpha, \beta, \gamma \bigr), \quad {dB(t) \over dt} = g_{N,M} \bigl( A(t), B(t), \alpha, \beta, \gamma \bigr) \] where \(A\) and \(B\) are functions from \(\mathbb{R}\) to the vector space \(V_{N,M}\) of all real valued functions \(\psi\) on \(\mathbb{Z} \times \mathbb{Z}\) satisfying \(\psi (n+N,m) = \psi (n,m+M) = \psi (n,m)\). A function \(f\) on \(\mathbb{R}^ 3\) is said to be of class \({\mathcal C}\) if the following conditions are verified: 1. \(f(x+1,y,t) = f(x,y+1,t) = f(x,y,t)\) for all \((x,y,t) \in \mathbb{R}^ 3\). 2. For all \(\varepsilon>0\), there exist \((\alpha, \beta, \gamma) \in \mathbb{R}^ 3\), an integer there exist \(N\), a constant there exist \(K\) and functions \(A(t)\), \(B(t)\) from \(\mathbb{R}\) to \(V_{N,N^ 2+1}\) such that: \(\bullet\) \(| f({n \over N}\), \({m \over N^ 2+1}, t) - NA(t) (n,m) - K | < \varepsilon\) \(\bullet\) \(A(t)\), \(B(t)\) satisfy the differential-difference equations above. The main result of the paper is that the class \({\mathcal C}\) contains many of the solutions of the KP equation arising from algebraic geometry. More precisely, the author proves the following: Theorem. Suppose that \(K_ 1 \in H^ 1 (C, \mathbb{Z})\), \(K_ 1^ \iota \in K_ 1 + H^ 1(C, \mathbb{Z})\) and the pair \((C,P)\) is good. Let \(\{v_ 1, v_ 2, v_ 3\}\) be an adapted and acceptable set in \(\Lambda^ +\) and let \(\overline v_ i\) be the image of \(v_ i\) in \(H^ 0 (C,\Omega)^*\). Suppose that the function \(\vartheta (x \overline v_ 1 + y \overline v_ 2 + z \overline v_ 3 + K_ 1)\) does not vanish on \(\mathbb{R}^ 3\). Let \(f(x,y,t) = {\partial^ 2 \over \partial x^ 2} \log \vartheta (x \overline v_ 1 + y \overline v_ 2 + z \overline v_ 3 + K_ 1)\). Then \(f\) is in the class \({\mathcal C}\). Under the assumptions of the theorem there exists a constant \(K\) such that \(f+K\) satisfies the KP equation. Some other constructions are similar to these by \textit{P. Van Moerbeke} and \textit{D. Mumford} [Acta Math. 143, 93-154 (1979; Zbl 0502.58032)].
0 references
KP equation
0 references
Riemann surface
0 references
theta function
0 references