It is known that the convex hull of permutation matrices yields exactly the stochastic matrices. I am interested in the convex hull of cyclic permutation matrices. Trivially this is a subset of the stochachastic matrices, but what more do we know than that about the relationship to the convex hull of all permutations?
My intuition would be that the chull of cyclic permutations is strictly smaller, but is this true? I have not yet found an example demonstrasting this!
If it is strictly smaller can we say something about those stochastic matrices within the "cyclic cone"? It seems they should hold some common property but i am unable to pinpoint it or find literature about it!