Cluster combinatorics and Artin group presentations
Hugh Thomas
Associated to a finite reflection group W is a lattice of
noncrossing partitions NC(W). It is of interest because it provides a
Garside structure on the Artin group A corresponding to W other than the
usual one. This "dual" structure has certain technical advantages. In
particular, it has been instrumental in constructing some Eilenberg-Mac
Lane spaces for A.
>
If W is crystallographic, it was observed by Chapoton than the number of
elements in NC(W) is the same as the number of clusters in the
corresponding cluster algebra. A bijection was later given by Reading. The
bijection was reinterpreted in representation-theoretic terms and extended
to include affine type by Ingalls and the speaker. The goal of this talk
is to discuss how this representation-theoretic approach can be
productively applied to the study of Artin groups. I will give a uniform
proof (for finite, crystallographic type) of the correctness of the
presentation at the heart of the dual Garside structure for finite type
Artin groups. (This was originally proved by Birman-Ko-Lee for type A, and
by Bessis, on a type-by-type basis, for general finite type.) I will also
explain the conjectural generalization of this result: a presentation for
arbitrary Artin groups which is expressed in terms of exceptional sequences
of quiver representations.