Minkowski summands of cubes (Q2680939)

If \(P,Q\) are convex polytopes, then the authors call \(Q\) a weak Minkowski summand of \(P\) and write \(Q \preceq P\) if there is a scalar \(\lambda > 0\) and another polytope \(R\) such that \(Q + R = \lambda P\), where \(Q + R = \{q + r : q \in Q, r \in R\}\). Observe that a translate \(Q + v\) of \(Q\) will also satisfy \(Q + v \preceq P\), and so it is convenient to factor out translations. There are various ways of seeing that the weak Minkowski summands of \(P\) form a polyhedral cone, called the type cone, for instance using the theory of representations of \textit{P. McMullen} [Geom. Dedicata 2, 83--99 (1973; Zbl 0273.52006)]. If scalars are also factored out, then a Gale diagram of a polar polytope to \(P\) provides an alternative picture. In this paper, the authors consider the type cones of polygons, cubes and generalizations of the latter. In particular, they sketch a proof of the fact that the type cone of a polytope which is combinatorially isomorphic to a direct sum of simplices is a cone over a simplex; in the Gale diagram picture, the family of weak Minkowski summands corresponds to a simplex.