A theoretical analysis at cooperative behavior in multi-agent Q-learning

L.R. Waltman, U. Kaymak

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

7 Citations (Scopus)
11 Downloads (Pure)

Abstract

A number of experimental studies have investigated whether cooperative behavior may emerge in multi-agent Qlearning. In some studies cooperative behavior did emerge, in others it did not. This paper provides a theoretical analysis of this issue. The analysis focuses on multi-agent Q-learning in iterated prisoner’s dilemmas. It is shown that under certain assumptions cooperative behavior may emerge when multi-agent Q-learning is applied in an iterated prisoner’s dilemma. An important consequence of the analysis is that multi-agent Q-learning may result in non-Nash behavior. It is found experimentally that the theoretical results presented in this paper are quite robust to violations of the underlying assumptions.
Original languageEnglish
Title of host publicationIEEE Symposium on Approximate Dynamic Programming and Reinforcement Learning (ADPRL 2007), 1-5 April 2007, Honolulu
Place of PublicationPiscataway
PublisherInstitute of Electrical and Electronics Engineers
Pages84-91
ISBN (Print)1-4244-0706-0
DOIs
Publication statusPublished - 2007

Fingerprint

Dive into the research topics of 'A theoretical analysis at cooperative behavior in multi-agent Q-learning'. Together they form a unique fingerprint.

Cite this