In a coalitional game, the coalitions are weakly ordered according to their worths in the game. When moreover a power index is given, the players are ranked according to the real numbers they are assigned by the power index. If any game inducing the same ordering of the coalitions generates the same ranking of the players then, by definition, the game is (ordinally) stable for the power index, which in turn is ordinally invariant for the game. If one is interested in ranking players of a game which is stable, re-computing the power indices when the coalitional worths slightly fluctuate or are uncertain becomes useless. Bivalued games are easy examples of games stable for any power index which is linear. Among general games, we characterize those that are stable for a given linear index. Note that the Shapley and Banzhaf scores, frequently used in AI, are particular semivalues, and all semivalues are linear indices. To check whether a game is stable for a specific semivalue, it suffices to inspect the ordering of the coalitions and to perform some direct computation based on the semivalue parameters.

The order polytopes we consider here are the linear order polytope, the interval order polytope, the semiorder polytope and the partial order polytope. Among their known facet defining inequalities (FDIs), many have their coefficients in {-1, 0, 1}. We consider the problem of finding all of these particular FDIs. The problem is easy for the partial order polytope. For the interval order polytope, we prove that the solution consists of the so-called io-clique inequalities of Müller and Schulz [5]. We present a characterisation of the {-1, 0, 1}-FDIs for the semiorder polytope. The similar problem for the linear order polytope remains open and seems harder to solve because of the large variety of known examples of FDIs which fall in this class.

Abstract Optimality of a linear inequality in finitely many graph invariants is defined through a geometric approach. For a fixed number of graph vertices, consider all the tuples of values taken by the invariants on a selected class of graphs. Then form the polytope which is the convex hull of all these tuples. By definition, the optimal linear inequalities correspond to the facets of this polytope. They are finite in number, are logically independent, and generate precisely all the linear inequalities valid on the class of graphs. The computer system GraPHedron, developed by some of the authors, is able to produce experimental data about such inequalities for a “small” number of vertices. It greatly helps in conjecturing optimal linear inequalities, which are then hopefully proved for any number of vertices. Two examples are investigated here for the class of connected graphs. First, all the optimal linear inequalities for the stability number and the number of edges are obtained. To this aim, a problem of Ore (1962) related to the Turán Theorem (1941) is solved. Second, several optimal inequalities are established for three invariants: the maximum degree, the irregularity, and the diameter. © 2008 Wiley Periodicals, Inc. NETWORKS, 2008

The weak order polytopes are studied in Gurgel and Wakabayashi [ Discrete Math., 175 (1997), pp. 163--172], Gurgel and Wakabayashi [The Complete Pre-Order Polytope: Facets and Separation Problem, manuscript, 1996], and Fiorini and Fishburn [Weak order polytopes, submitted]. We make use of their natural, affine projection onto the partition polytopes to determine several new families of facets for them. It turns out that not all facets of partition polytopes are lifted into facets of weak order polytopes. We settle the cases of all facet-defining inequalities established for partition polytopes by Grötschel and Wakabayashi [Math. Programming, 47 (1990), pp. 367--387]. Our method, although rather simple, allows us to establish general families of facets which contain two particular cases previously requiring long proofs.