Frobenius categories versus Brauer blocks. The Grothendieck group of the Frobenius category of a Brauer block. (Q2378562)

This book has been long in the making. It is the author's considered account of a body of work inspired by his philosophical position that the categorical point of view is the correct approach to much of modular representation theory. In preprint form, the book was a major influence on the highly active new area ``\(p\)-local finite groups'', which involves interactions with algebraic topology. In particular, Puig's abstract Frobenius categories are called ``saturated fusion systems'' in [\textit{C.~Broto}, \textit{R.~Levi} and \textit{B.~Oliver}, J. Am. Math. Soc. 16, No. 4, 779-856 (2003; Zbl 1033.55010)]. The Alperin-Broué theory of Brauer pairs equips the subgroups of the defect group of a block with a conjugacy relation resembling that between the subgroups of the Sylow \(p\)-subgroup of a finite group. The goal of this book is to give abstract, categorical axioms which make this resemblance precise; to develop this abstract theory of Frobenius categories as far as possible; and then to apply it back to block theory. Several group-theoretic concepts are generalized to the case of a Frobenius category, including normalizers and centralizers. The full subcategory of selfcentralizing subgroups determines the Frobenius category. The analogy with the case of nilpotent blocks suggests the nilcentralized subgroups as an important object of study. The author obtains an Alperin Fusion Theorem for Frobenius categories. A Frobenius category has a hyperfocal subgroup. The notion of solvability can be generalized to Frobenius categories, and all solvable Frobenius categories arise from \(p\)-solvable finite groups. Assembling tools to attack the Alperin Weight Conjecture is one major theme of the book. In view of the work of Knörr and Robinson, the author considers chains of nilcentralized Brauer pairs. Each pair admits a central extension of the Weyl group by the centre of the \(p\)-group, but one wants to demonstrate that these extensions may be performed in a coherent way. Puig shows that a coherent choice of \(k^*\)-extensions is possible, whilst leaving open the question of an extension of the whole subcategory. The Grothendieck group of a Frobenius category is the inverse limit of the composition of this extension functor and the modular Grothendieck group. One main result is Puig's rephrasing of the conjecture as the statement that the \(\mathbb{Z}\)-ranks of the Grothendieck groups of a block and of its Frobenius category coincide. In the longer term, it might also be possible to use the Grothendieck group of a Frobenius category to study Dade's Conjecture. The main theme of the later chapters is a systematic study of localities, which are extensions of the Frobenius category. Ideally one would like to have a perfect locality: this corresponds to the centric linking systems of Broto, Levi and Oliver, [op.~cit.]. Their work on bisets and fusion leads Puig to the concept of a basic locality: though not perfect, this is canonical. The book concludes with some notes on the ongoing hunt for a perfect locality. An appendix assembles background material on the cohomology of small categories.