Evolution of cooperation in Axelrod tournament using cellular automata

Author(s)

Schimit, P. H. T. ^[1] ; Santos, B. O. ^[1] ; Soares, C. A. ^[1]

Affiliation

^[1] Univ Nove Julho, Dept Informat, BR-02112000 Sao Paulo, SP - Brazil

• Document type: Journal article
• Source: PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS; v. 437, p. 204-217, NOV 1 2015.
• Web of Science Citations: 5

Abstract

Results of the Axelrod Tournament were published in 1981, and since then, evolutionary game theory emerged as an idea for understanding relations, like conflict and cooperation, between rational decision-makers. Robert Axelrod organized it as a round-robin tournament where strategies for iterated Prisoner's Dilemma were faced in a sequence of two players game. Here, we attempt to simulate the strategies submitted to the tournament in a multi-agent context, where individuals play a two-player game with their neighbors. Each individual has one of the strategies, and it plays the Prisoner's Dilemma with its neighbors. According to actions chosen (cooperate or defect), points of life are subtracted from their profiles. When an individual dies, some fitness functions are defined to choose the most successful strategy which the new individual will copy. Although tit-for-tat was the best strategy, on average, in the tournament, in our evolutionary multi-agent context, it has not been successful. (C) 2015 Elsevier B.V. All rights reserved. (AU)

• FAPESP's process: 15/01032-9 - Evolution of cooperation and population dynamics using cellular automata and complex networks
• Grantee: Pedro Henrique Triguis Schimit
• Support Opportunities: Regular Research Grants