Optimal strategies of the iterated prisoner’s dilemma problem for multiple conflicting objectives

Abstract — In this paper, we present a new paradigm of searching optimal strategies in the game of Iterated Prisoner’s Dilemma using multiple objective evolutionary algorithms. This method is better than the existing approaches, because it not only gives strategies which perform better in the iterated game, but also gives a family of non-dominated strategies, which can be analyzed to see what properties a strategy should have to win in the game. We present the results obtained with this new method, and also the common pattern emerging from the set of non-dominated strategies so obtained