Ghost numbers of Group Algebras

TL;DR

This paper introduces the concept of ghost numbers in group algebras to measure the failure of the generating hypothesis in stable module categories, providing computations, bounds, and classifications for various p-groups.

Contribution

It defines ghost numbers for group algebras, relates them to Auslander-Reiten triangles, and computes bounds and classifications for different p-groups.

Findings

Ghost numbers quantify the failure of the generating hypothesis.

Explicit ghost number bounds are provided for various p-groups.

A classification of p-group algebras with small ghost numbers is achieved.

Abstract

Motivated by Freyd's famous unsolved problem in stable homotopy theory, the generating hypothesis for the stable module category of a finite group is the statement that if a map in the thick subcategory generated by the trivial representation induces the zero map in Tate cohomology, then it is stably trivial. It is known that the generating hypothesis fails for most groups. Generalizing work done for $p$-groups, we define the ghost number of a group algebra, which is a natural number that measures the degree to which the generating hypothesis fails. We describe a close relationship between ghost numbers and Auslander-Reiten triangles, with many results stated for a general projective class in a general triangulated category. We then compute ghost numbers and bounds on ghost numbers for many families of $p$-groups, including abelian $p$-groups, the quaternion group and dihedral $2$-groups, and also give a general lower bound in terms of the radical length, the first general lower bound that we are aware of. We conclude with a classification of group algebras of $p$-groups with small ghost number and examples of gaps in the possible ghost numbers of such group algebras.