A classification of the irreducible algebraic \(\mathcal A\)-hypergeometric functions associated to planar point configurations (Q2451119)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | A classification of the irreducible algebraic \(\mathcal A\)-hypergeometric functions associated to planar point configurations |
scientific article |
Statements
A classification of the irreducible algebraic \(\mathcal A\)-hypergeometric functions associated to planar point configurations (English)
0 references
26 May 2014
0 references
\(\mathcal{A}\)-hypergeometric functions were introduced in the 1980s by Gelfand, Graev, Kapranov and Zelevinsky. They are generalizations of classical hypergeometric functions of Gauss, Appell and Horn. The interesting problem of algebraicity of these functions for specified values of parameters was solved by Schwarz (1873), Appel-Laurecella, Horn (see [\textit{T. Sasaki}, J. Fac. Sci., Univ. Tokyo, Sect. I A 24, 565--573 (1977; Zbl 0388.33003)]) and others. In this paper the investigation of algebraicity is extended to the case of all \(\mathcal{A}\)-hypergeometric functions where \(\mathcal{A}\in\mathbb{Z}^2\) and \(\mathcal{A}\in\mathbb{Z}^3\). We give the definitons of such \(\mathcal{A}\)-hypergeometric functions. Let \(\mathcal{A}=\{\mathbf{a_1}, \dots, \mathbf{a_N}\}\) be a finite subset of \(\mathbb{Z}^r\) such that \(\mathbb{Z}A=\mathbb{Z}^r\) and there exists a linear form \(h\) on \(\mathbb{R}^r\) such that \(h(\mathbf{a}_i)=1\) for all \(i\). The \textit{lattice of relations} of \(\mathcal{A}\) is \(\mathbb{L}=\{(l_1,\dots,l_N)\in\mathbb{Z}^N ~ | ~ l_1 \mathbf{a}_1+\dots+l_N \mathbf{a}_N=0\}\). Let \(\mathbf{\beta}\in\mathbb{C}^r\) and denote by \(\partial_i\) the differential operator \(\frac{\partial}{\partial z_i}\). The \(\mathcal{A}\)-hypergeometric \textit{system associated to \(\mathcal{A}\) and \(\mathcal{\beta}\)}, denoted by \(H_{\mathcal{A}}(\mathbf{\beta})\), consists of two sets of differential equations: {\parindent=6mm \begin{itemize}\item[-] the \textit{structure equations}: for all \(l=(l_1,\dots,l_N)\in\mathbb{L}\) \[ \square_{l} \Phi=\left(\prod_{l_i>0}\partial_{i}^{l_i}\right)\Phi-\left(\prod_{l_i<0}\partial_{i}^{-l_i}\right)\Phi=0 \] \item[-] the \textit{homogeneity} or \textit{Euler equations}: for \(1 \leq i \leq r\) \[ a_{1i}z_1\partial_1\Phi+\dots+a_{Ni}z_N\partial_N\Phi=\beta_i\Phi. \] \end{itemize}} In order to formulate the principal result following definitions are necessary. Definition 1. \(\mathcal{A}\) is called a \textit{pyramid} if \(\mathcal{A}=\mathcal{A}'\cup\{\mathbf{a}_i\}\), where all elements of \(\mathcal{A}'\) lie in an \(r-2\) dimensional hyperplane. Definition 2. \(\mathcal{A}\) is called \textit{normal} if \(C(\mathcal{A})\cap\mathbb{Z}^r=\mathbb{Z}_{\geq 0}\mathcal{A}\), where \(C(\mathcal{A})\) is a real non-negative cone spanned by the elements of \(\mathcal{A}\). Definition 3. \(H_{\mathcal{A}}(\mathbf{\beta})\) is called \textit{resonant} if \(\beta+\mathbb{Z}^r\) contains a point in a face of \(C(\mathcal{A})\). Otherwise \(H_{\mathcal{A}}(\mathbf{\beta})\) is called \textit{non-resonant}. The main result of this paper is the following theorem. Theorem. If \(\mathcal{A}\) is a finite normal subset of \(\mathbb{Z}^2\), lying on an affine line, then there exists \(\mathbf{\beta}\in\mathbb{Q}^2\) such that \(H_{\mathcal{A}}(\mathbf{\beta})\) is non-resonant and has algebraic solutions. If \(\mathcal{A}\) is a finite normal non-pyramidal subset of \(\mathbb{Z}^3\), lying in an affine plane, then there exists \(\mathbf{\beta}\in\mathbb{Q}^3\) such that \(H_{\mathcal{A}}(\mathbf{\beta})\) is non-resonant and has algebraic solutions if and only if the polygon \(P(\mathcal{A})=\{x\in\mathbb{R}^2|(x,1)\in Q(\mathcal{A})\},\) where \(Q(\mathcal{A})\) is the convex hull of \(\mathcal{A},\) is isomorphic to one of the polygons drawn in Fig. 1 of the paper and called: ``The polygons \(P(\mathcal{A})\) such that there exists \(\beta\) such that \(H_{\mathcal{A}}(\beta)\) has irreducible algebraic solutions.'' Now we give a short characterization of the paper under review. At first the author gives the description of algebraic functions using a combinatorial criterion and triangulations of polygons. In Sec. 3 an analysis of one-dimensional sets is given. Instead of the two-dimensional sets \(\mathcal{A}\) the author mostly uses the polygons \(P(\mathcal{A})\). In Sec. 4 the computation of all polygons with at most 2 interior lattice points and the determination of irreducible algebraic functions are given. In Sec. 5 it is shown that all polygons with sufficiently many interior and boundary lattice points contain smaller subpolygons. To this subpolygons the problem of finding irreducible algebraic functions can be reduced. Sec. 6 is devoted to an investigation of polygons with many interior points and few boundary points.
0 references
hypergeometric functions
0 references
algebraic functions
0 references
polygons
0 references
0 references
0 references