Axioms for the outcomes of negotiation in matrix games (Q1566510): Difference between revisions

The author proposes an axiomatic base for some generalization of a matrix game. An element of a payoff array is an \(n\)-vector which is indexed with \(n\)-tuple. Similar to ordinary matrix games each of \(n\) players chooses a concrete value of the index arranging to increase his payoff, but this decision it not terminal. Each player may change his mind many times in response to the strategies of the others. He has complete information about the payoff array and the decisions of all players. The author examines a set \(X(G)\) of all possible payoffs. To provide a ``true'' distribution and to apply mathematical methods the author imposes some axioms on \(X(G)\). Namely, nine ones are introduced, the most significant seems to be the Pareto-optimality of \(X(G)\) and the floor axiom. The last one means that if the players cannot finish the process of choosing the indices, then somebody declares the termination of the game and announces the result. It takes into account rational expectation of any player. Various examples demonstrate necessity of all axioms. The set \(X(G)\) is unique, if it exists. To construct some element of \(X(G)\) the author presents rules for the players. The author investigates the influence of permutation of players during their choice of indices and proves that due to axioms the set \(X(G)\) remains invariant.