Cyclic Shuffle Compatibility

Abstract: Consider a permutation p to be any finite list of distinct positive integers. A statistic is a function St whose domain is all permutations. Let S(p,q) be the set of shuffles of two disjoint permutations p and q. We say that St is shuffle compatible if the distribution of St over S(p,q) depends only on St(p), St(q), and the lengths of p and q. This notion is implicit in Stanley's work on P-partitions and was first explicitly studied by Gessel and Zhuang. One of the places where shuffles are useful is in describing the product in the algebra of quasisymmetric functions. Recently Adin, Gessel, Reiner, and Roichman defined an algebra of cyclic quasisymmetric functions where a cyclic version of shuffling comes into play. The purpose of this paper is to define and study cyclic shuffle compatibility. In particular, we show how one can lift shuffle compatibility results for (linear) permutations to cyclic ones. We then apply this result to cyclic descents and cyclic peaks. We also discuss the problem of finding a cyclic analogue of the major index.