Classification of multiplicity free symplectic representations (Q855706): Difference between revisions

Let \(V\) be a finite dimensional complex vector space with a symplectic form \(\omega.\) Then we can form the Weyl algebra \(\mathcal{W}(V)\) which, by definition, is generated by \(V\) with relations \[ v\cdot w-w\cdot v=\omega(v,w). \] The symplectic group Sp\((V)\) acts on \(\mathcal{W}(V)\) by algebra automorphisms. Assume a connected reductive group \(G\) is acting linearly on \(V\) preserving the symplectic structure \(\omega\). Then \(V\) is called a \textit{symplectic representation} of \(G\), and can be thought of as a homomorphism \(\rho:G\rightarrow \text{Sp}(V).\) This symplectic representation is multiplicity free (``MFSR'') if \(\mathcal{W}(V)^G\) is commutative. Said in another way, there is a Lie algebra homomorphism \(\mathfrak s\mathfrak p(V)\mapsto\mathcal{W}(V)\) whose adjoint integrates to the \(\text{Sp}(V)\)-action, and the symplectic representation is multiplicity free if there exists a Lie algebra homomorphism \[ \tilde{\rho}:\mathfrak g\rightarrow\mathcal{W}(V) \] such that \(\mathcal{W}(V)^G\) is the commutant of \(\tilde{\rho}(\mathfrak g)\) in \(\mathcal{W}(V).\) Two instances of MFSRs have been studied extensively, the first one is given by Howe's reductive dual pairs: Let \[ G=G_1\times G_2=\text{Sp}_{2m}(\mathbb{C})\times \text{SO}_n(\mathbb{C})\text{ acting on }V=\mathbb{C}^{2m}\otimes \mathbb{C}^n\text{ or } \] \[ G=G_1\times G_2=\text{GL}_m(\mathbb{C})\times \text{GL}_n(\mathbb{C})\text{ acting on }V=(\mathbb{C}^m\otimes\mathbb{C}^n)\oplus(\mathbb{C}^m\otimes\mathbb{C}^n)^\ast. \] Then the algebras generated by \(\mathfrak g_1\) and \(\mathfrak g_2\) are mutual commutants inside \(\mathcal{W}(V).\) The other class previously studied, generalizes the second case above: Let \(G\to \text{GL}(U)\) be any finite-dimensional representation. Then \(V=U\oplus U^\ast\) carries the \(G\)-invariant symplectic structure \(\omega(u_1+u_2^\ast,u_2+u_2^\ast)=\langle u_1^\ast,u_2\rangle-\langle u_2^\ast,u_1\rangle.\) In this situation one can identify \(\mathcal{W}(V)\) with the algebra \(\mathcal{P}\mathcal{D}(U)\) of polynomial coefficient linear differential operators on \(U\). One deduces that \((G,V)\) is multiplicity free if and only if the algebra \(\mathbb{C}[U]\) of polynomial functions on \(U\) is multiplicity free, i.e., does not contain any simple \(G\)-module more than once. Such modules, called multiplicity free spaces are classified by Kac, Leahy, Benson and Ratcliff. In this article, the author extends this work by classifying all multiplicity free symplectic representations. Let \(\mathcal{Z}(\mathfrak g)\) be the center of the universal enveloping algebra \(\mathcal{U}(\mathfrak g)\). The morphism \(\tilde{\rho}:\mathfrak g\rightarrow\mathcal{W}(V)\) induces algebra homomorphisms \[ \mathcal{U}(\tilde{\rho}):\mathcal{U}(\mathfrak g)\rightarrow\mathcal{W}(V)\text{ and }\mathcal{U}(\tilde{\rho})^G:\mathcal{Z}(\mathfrak g)\rightarrow\mathcal{W}(V)^G. \] In general \(V\) is a MFSR if and only if \(\mathcal{W}^G\) is a finitely generated \(\mathcal{Z}(\mathfrak g)\)-module. The Weyl algebra is filtered by putting \(V\) in degree 1. The associated graded version of \(\mathcal{U}(\tilde{\rho})\) is the moment map \[ m:V\rightarrow\mathfrak g^\ast:v\mapsto[\xi\mapsto\frac{1}2\omega(\xi v,v)]. \] Then \(V\) is a MFSR if and only if almost all \(G\)-invariants on \(V\) are pull-backs of coadjoint invariants, i.e., \(\mathcal{O}(V)^G\) is a finitely generated \(\mathcal{O}(\mathfrak g^\ast)^G\)-module. The central result about MFSRs is the following: If \(G\rightarrow \text{Sp}(V)\) is a MFSR, the algebra \(\mathcal{O}(V)^G\) is a polynomial ring, and \(\mathcal{O}(V)\) is a free \(\mathcal{O}(V)^G\)-module. Now, this preliminary results makes it possible for the author to classify first all saturated indecomposable MFSRs by listing them in two nice tables, then all MFSRs are classified as various combinations of this. The author gives the complete toolbox for this setup, and all theorems are proved in a structured, readable way.