Mackey formula for large \(q\) (Q1268095): Difference between revisions
From MaRDI portal
Latest revision as of 15:48, 28 May 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Mackey formula for large \(q\) |
scientific article |
Statements
Mackey formula for large \(q\) (English)
0 references
19 June 2001
0 references
The Mackey formula for a finite group \(G\) gives an expression for the restriction to a subgroup \(K\) of an induced representation from a subgroup \(H\). In the case of finite reductive groups one has the operations of Harish-Chandra induction (or parabolic induction) and Lusztig induction, and one can ask whether the Mackey formula holds for these operations. For Harish-Chandra induction it is known that the formula holds, and in this paper the author considers Lusztig induction. Let \(G\) be a connected, reductive algebraic group defined over \(\mathbb{F}_q\), where \(q\) is a power of the prime \(p\) and let \(F\colon G\to G\) a Frobenius morphism and \(G^F\) be the finite reductive group of \(F\)-fixed points of \(G\). Let \({\mathcal R}(G^F)\) denote the Grothendieck group of finite-dimensional \(\overline\mathbb{Q}_\ell G^F\)-modules, where \(\ell\) is a prime number distinct from \(p\). If \(L\) is an \(F\)-stable Levi subgroup of a (not necessarily \(F\)-stable) parabolic subgroup \(P\) of \(G\), we have the Lusztig induction map \(R_{L\subset P}^G\colon{\mathcal R}(L^F)\to{\mathcal R}(G^F)\) and its adjoint \(^* R_{L\subset P}^G\colon{\mathcal R}(G^F)\to{\mathcal R}(L^F)\). Let \(M\) and \(Q\) denote another \(F\)-stable Levi subgroup and parabolic subgroup, respectively, where \(M\subseteq Q\). Then the Mackey formula states that \[ {^* R}_{L\subset P}^G\circ R_{M\subset Q}^G=\sum_{g\in L^F\backslash{\mathcal S}_G(L,M)^F/M^F}R_{L\cap{^g M}\subset L\cap{^g Q}}^L\circ{^* R}_{L\cap{^g M}\subset P\cap{^g M}^{^g M}}\circ(\text{ad }g), \] where \({\mathcal S}_G(L,M)\) is the set of \(g\in G\) such that \(L\cap{^g M}\) contains a maximal torus of \(G\), and \(\text{ad }g\) denotes the map \({\mathcal R}(M^F)\to{\mathcal R}({^g M}^F)\) induced by the conjugation by \(g\). The author proves that for sufficiently large \(q\), the Mackey formula holds for \(G^F\). The author gives two proofs of that result. In his first proof, he obtains a bound for \(q\) which depends only on the root data associated with \(G\). The bound found is given by \(q>1+i(G)^2l(G)\), where \(i(G)\) denotes the greatest of the numbers \(|Z'/Z^{\prime\circ}|\), where \(Z'\) runs over the set of the centers of all the connected reductive subgroups of \(G\) of the same rank as \(G\) (here \(Z^{\prime\circ}\) is the connected component of \(Z'\)), and where \(l(G)\) denotes the maximum of the orders (modulo the centre) of the isolated semisimple elements of the groups \(G^{\prime*}\), with \(G'\) running over the set of dual groups of connected reductive subgroups of \(G\) of the same rank as \(G\). The author states that his proof is inspired by the proof of orthogonality of Deligne-Lusztig characters given by \textit{P.~Deligne} and \textit{G.~Lusztig} in their orignal paper [Ann. Math., II. Ser. 103, 103-161 (1976; Zbl 0336.20029)]. The second proof of the Mackey formula is based on the theory of character sheaves developed by Lusztig, assuming that \(p\) is ``almost good'' for \(G\) and that \(q\) is sufficiently large. In this case the bound for \(q\) is not explicit.
0 references
reductive groups
0 references
finite fields
0 references
character sheaves
0 references
Mackey formula
0 references
induced representations
0 references
Harish-Chandra induction
0 references
Lusztig induction
0 references
Frobenius morphisms
0 references
Grothendieck groups
0 references
0 references