Title :
Ostracism for improving cooperation in the iterated prisoner´s dilemma game
Author_Institution :
Dept. of Comput. Sci., Pittsburgh Univ., PA, USA
Abstract :
This paper considers methods for improving cooperation in the Iterated Prisoner´s Dilemma (IPD) game. We consider two previous efforts for improving cooperation that use a bias in selecting opponents in the IPD game: the space-based approach and the tag-based approach. In the biased opponent selection mechanism, opponents are selected using a bias based on, for example, a normal distribution function around the player who tries to choose opponents; i.e., the closer a prospective opponent is to a player, the higher the probability that it will be chosen as an opponent by that player. The cooperation can be achieved by promoting the emergence of cooperators´ clusters; thus, cooperators will play the game with the cooperators within the “kin dusters”. Our research shows that use of ostracism schemes that expel dissimilar players from the kin clusters can greatly support the biased opponent selection mechanism and therefore promote cooperation in the overall environment
Keywords :
cooperative systems; game theory; IPD game; biased opponent selection; cooperation; improving cooperation; iterated prisoner´s dilemma game; Computer science; Gaussian distribution; History; Thin film transistors;
Conference_Titel :
Evolutionary Computation, 1999. CEC 99. Proceedings of the 1999 Congress on
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-5536-9
DOI :
10.1109/CEC.1999.782517