“…(1) 0 ∈ p(A) if and only if A is normal (N) (meaning a i,i = 0, a i,j ≤ 0, ∀i, j), (see [10,38]) (2) if 0 ∈ p(A), then A describes p(A) optimally (or tightly) if and only if A is normal idempotent (NI) (meaning that, in addition to normality, we have A ⊙ A = A, which requires that a i,j + a jk ≤ a ik , ∀i, j, k) 1 (see [24,30,36]). (3) for each alcoved polytope P containing 0 there exist a unique NI matrix A such that P = p(A) (see Lemma 2.6 in [24] and [30,36]).…”