Achieving cooperation with many prisoners in the NIPD. (2016)

• Record Type: Journal Article
• Authors: Agudo, J. Enrique; Fyfe, Colin
• Is Part Of: International journal of computational science and engineering. Volume 12:Number 1(2016)
• Page Start: 3
• Page End: 16
• Publication Date: 2016
• Subjects: iterated prisoners&amp -- #39 -- dilemma -- IPD -- cross entropy -- reinforcement learning -- cooperation -- many prisoners -- game theory; Computer science -- Mathematics -- Periodicals; Computer simulation -- Mathematical aspects -- Periodicals; Computational intelligence -- Periodicals; 004.015105
• Languages: English
• ISSNs: 1742-7185

Abstract

This paper discusses an empirical investigation into the N-person's iterated prisoners' dilemma (NIPD), a standard problem from game theory. We use both the cross entropy method and reinforcement learning and achieve cooperation with much greater sizes of population than we have previously been able to do with genetic algorithms and artificial immune systems. Our experimental results give some insight into the circumstances where cooperation might develop.