An improved approach for adversarial decision making under uncertainty based on simultaneous game

Abstract

Game theory is used widely when to make decisions in an adversarial environment. There are usually two adversarial players in a game and both of players try to maximize their payoffs. However the classical game assumes the players are rational and do not consider other uncertainty factors such as personal preference. This paper proposes a method to take uncertainty into account when using game theory to make decisions. And Dempster-Shafer evidence theory is used to deal with the uncertainty, then the payoff matrix of game can be modified properly and the new payoff matrix reflects the impact of uncertain factors. Finally the equilibrium solution is computed which is the optimal decision under uncertainty.