| Self-Adapting Payoff Matrices in Repeated Interactions (2008) | |||||||||||||
Abstract | |||||||||||||
| Abstract-Traditional iterated prisoner's dilemma (IPD) as- in light of the opponent's, is fixed and symmetric. That is, sumed a fixed payoff matrix for all players, which may not be the utility expectation of a strategy on rewards for certain realistic because not all players are the same in the real-world. behaviors does not change (i.e., fixed payoff matrix) and This paper introduces a novel co-evolutionary framework where behaviors stcange (i.e., fixed payoff matrix)a each strategy has its own self-adaptive payoff matrix. This is s lo all straties (i.e., s etrictpayof trIx) framework is generic to any simultaneous two-player repeated These two basic assumptions might not be realistic if the IPD encounter game. Here, each strategy has a set of behavioral game is used as a model to explain outcomes of real-world responses based on previous moves, and an adaptable payoff interactions due to variations between individuals on the matrix based on reinforcement feedback from game interactions payoff matrix [I I]. More importantly, these two assumptions that is specified by update rules. We study how different update payoffcat impontl thesehtworasmins rules affect the adaptation of initially random payoff matrices, havsignificanthimplicati sttogmodellin bavainterand how this adaptation in turn affects the learning of strategy actions because they restrict strategies from adapting (e.g., behaviors. learning) their individual payoff matrices based on feedback Keywords: Evolutionary games, Co-evolution, Iterated of game interactions that reinforces certain behaviors. As Prisoner's Dilemma, Mutualism, Repeated Encounter Games such, cooperative outcomes may not be a result of IPD-like | |||||||||||||
Details der Publikation | |||||||||||||
| |||||||||||||