The Kostant cone and the combinatorial structure of the root lattice (Q1123250)

Let R be a reduced irreducible root system with base \(B=\{\alpha_ 1,...,\alpha_ n\}\). If \(\Lambda_ R^+\) is the positive part of the root lattice, i.e. the additive semigroup generated by B, then the Kostant cone of \(\Lambda_ R^+\) is the set of elements in \(\Lambda_ R^+\) which can be written as the sum of positive non-simple roots. In an earlier paper [J. Algebra 96, 9-17 (1985; Zbl 0603.17005)] the author has introduced an operation s: \(\Lambda_ R^+\to \Lambda_ R^+\) (or, equivalently s: \({\mathbb{Z}}^ n_+\to {\mathbb{Z}}^ n_+)\) which can be viewed as a projection onto the Kostant cone. In this paper he provides more information on this operation. For example he answers questions such as: - How many elements of \(\Lambda^+_ R\) get mapped to a given element under s? - How are the Kostant partition functions of these elements related? - How do the root weights \(\Pi_ r(\lambda)\) of a highest weight module V(\(\lambda)\) behave under s? (Here \(\Pi_ r(\lambda)\) is the set of differences between the highest weight \(\lambda\) and arbitrary weights of the representation.) In the last part of the paper the combinatorial structure of the set \(E_+(x)\) of all expressions which give \(x\in \Lambda_ R^+\) as a sum of positive roots is studied in some detail.