Rigid analytic uniformization of curves and the study of isogenies (Q2462625)

Whereas the isogenies between abelian varieties \(A_1,A_2\) are hard to write down in algebraic terms, the description becomes very simple if one considers their complex uniformization, i.e.\ their representation \(A_i \cong \mathbb{C}^g / \Lambda_i\) as complex tori: Every isogeny \(\phi : A_1 \rightarrow A_2\) corresponds to a bijective \(\mathbb{C}\)-linear map \(\tilde{\phi} : \mathbb{C}^g \rightarrow \mathbb{C}^g\) with \(\tilde{\phi}(\Lambda_1) \subseteq \Lambda_2\). In particular, the isogeny can be described in terms of an integral \(2g \times 2g\)-matrix. The main idea of the current article is to employ non-archimedian uniformization for the same purpose, under the assumption that the abelian varieties are Jacobians of hyperelliptic curves, and to provide computational examples. Non-archimedian uniformization takes place in the category of rigid-analytic spaces due to Tate. After reviewing the basic notions of this theory, it is explained how an abelian variety over \(\mathbb{C}_p\), considered as a rigid-analytic space, can be represented as a multiplicative torus \((\mathbb{G}_m)^g / \Gamma\), where \(\Gamma\) is a group of automorphisms on \((\mathbb{G}_m)^g\) that can be specified by a matrix of integers. Given two such tori in terms of their matrices, the author provides a computational criterion for the tori to be isogenous. A subsequent section recalls the notion of Mumford curves, their relevance for the uniformization of algebraic curves, and the construction of an non-archimedian analogue of the Abel-Jacobi map based on theta functions. These facts are the essential ingredients for an algorithm which, for a hyperelliptic curve given by its defining equation, outputs the matrix of integers that belongs to the corresponding uniformized Jacobian. The author presents a sample computation which disproves the existence of isogenies between the Jacobians of some randomly chosen hyperelliptic curves. Finally, it is shown that for high genus this method is more efficient than the classical, complex-analytic approach.