Geometric invariant theory and generalized eigenvalue problem. II (Q407792)

Let \(G \subseteq \hat{G}\) be complex connected reductive groups. Stated generally, this paper considers the question of what irreducible representations of \(G\) appear in a given irreducible representation of \(\hat{G}\). More precisely, fixing maximal tori \(T\subseteq \hat{T}\) and Borel subgroups of \(G\) and \(\hat{G}\), one needs to understand the cone \(\mathcal{LR}(G,\hat{G})\) generated by the dominant characters \((\nu,\hat{\nu})\) of \(T\times\hat{T}\) such that \(V_\nu^*\) is a submodule of \(V_{\hat{\nu}}\), where \(V_\nu\) (resp. \(V_{\hat{\nu}}\)) denotes the \(G\)-module (resp. \(\hat{G}\)-module) associated with \(\nu\) (resp. \(\hat{\nu}\)). The author presents a list of inequalities describing the facets of \(\mathcal{LR}(G,\hat{G})\) which intersect the interior of the dominant chamber, and proves that this list is minimal in the sense that the inequalities are pairwise distinct and none can be omitted. He also moves on to consider smaller faces of the cone. The main results generalize work of \textit{A. Berenstein} and \textit{R. Sjamaar} [J. Am. Math. Soc. 13, No. 2, 433--466 (2000; Zbl 0979.53092)] and \textit{P. Belkale} and \textit{S. Kumar} [Invent. Math. 166, No. 1, 185--228 (2006; Zbl 1106.14037)], among others. The proofs use tools from Geometric Invariant Theory which, as well as giving the results, allow the author to show geometrically why some of the inequalities describing \(\mathcal{LR}(G,\hat{G})\) in the paper of Berenstein-Sjammar cited above are redundant.