On stable equivalences with endopermutation source. (Q2343504)

The main theorem of the paper under review is a sufficient criterion for the existence of a stable equivalence of Morita type between block algebras \(\mathcal OGb\) and \(\mathcal OHc\) of two finite groups \(G\) and \(H\); here \(\mathcal O\) is a suitable complete discrete valuation ring with residue field \(k\) of characteristic \(p>0\), and \(b\) and \(c\) are primitive idempotents of the centers \(Z(\mathcal OG)\) and \(Z(\mathcal OH)\), respectively, with a common defect group \(P\). The precise hypotheses for the main theorem are somewhat technical. They require the existence of (almost) source idempotents \(i\) and \(j\) of \(b\) and \(c\), respectively, inducing the same fusion system \(\mathcal F\) on \(P\), and the existence of a suitable \(\mathcal O[G\times H]\)-module \(M\) with vertex \(\Delta P=\{(x,x):x\in P\}\) whose source \(V\), considered as an \(\mathcal OP\)-module in the canonical way, is an \(\mathcal F\)-stable indecomposable endopermutation module. One also requires that \(M\) induces Morita equivalences between the block algebras \(kC_G(Q)e_Q\) and \(kC_H(Q)f_Q\), for any nontrivial subgroup \(Q\) of \(P\); here \(e_Q\) and \(f_Q\) are the primitive idempotents of \(Z(kC_G(Q))\) and \(Z(kC_H(Q))\) determined by \(i\) and \(j\), respectively. The main theorem of the paper can be viewed as a converse to a result by Puig. A special case arises in the context of a hypothetical minimal counterexample to the \(Z^*_p\)-theorem, a situation considered by \textit{E. Biland} [Adv. Math. 259, 67-88 (2014; Zbl 1292.20010)]. Another special case arises in the work of \textit{L. Puig} on blocks with abelian defect groups and Frobenius inertial quotient [Geom. Dedicata 37, No. 1, 9-43 (1991; Zbl 0724.20012)].