Advertisement

Cooperation is not always so simple to learn

  • M. Mailliard
  • F. Amblard
  • C. Sibertin-Blanc
  • P. Roggero
Part of the Springer Series on Agent Based Social Systems book series (ABSS, volume 3)

Abstract

In this paper, we propose to study the influence of different learning mechanisms of social behaviours on a given multi-agent model (Sibertin-Blanc et al. 2005). The studied model has been elaborated from a formalization of the organized action theory (Crozier and Friedberg 1977) and is based on the modelling of control and dependency relationships between resources and actors. The proposed learning mechanisms cover different possible implementations of the classifier systems on this model. In order to compare our results with existing ones in a classical framework, we restrain here the study to cases corresponding to the prisoner’s dilemma framework. The obtained results exhibit variability about convergence times as well as emergent social behaviours depending on the implementation choices for the learning classifier systems (LCS) and on the LCS parameters. We conclude by analysing the sources of this variability and giving perspectives about the use of such a model in broader cases.

Keywords

Global Satisfaction Learn Agent Mutual Cooperation Action Pair Learning Classifier System 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Axelrod, R.: The evolution of cooperation. Basic Books, New York, (1984).Google Scholar
  2. Conte, R., Paolucci, M.: Intelligent Social Learning. JASSS vol. 4, no. 1, (2001).Google Scholar
  3. Crozier, M., Friedberg, E.: L’acteur et le système: les contraintes de l’action collective. Seuil (1977).Google Scholar
  4. Delahaye, J. P., L’altruisme récompensé ? Pour La Science, 181:150–156, (1992).Google Scholar
  5. Dugatkin, L.A.: Cooperation among Animals: An Evolutionary Perspective. Oxford University Press, (1997).Google Scholar
  6. Flache A., Macy, M.W.: Stochastic colusion and the power law learning. Journal of Conflict Resolution, (2002).Google Scholar
  7. Hoffmann, R.: Twenty Years on: The Evolution of Cooperation Revisited, Journal of Artificial Societies and Social Simulation (JASSS) vol. 3, no. 2, (2000).Google Scholar
  8. Holland, J, All.: What Is a Learning Classifier System? LCS’99, LNAI 1813, 3–32 (2000).Google Scholar
  9. Macy, M.., Flache. A.. Learning Dynamics in Social Dilemmas. Proceedings of the National Academy of Sciences U.S.A. May 14;99(10):7229–7236,(2002).CrossRefGoogle Scholar
  10. Mailliard, M., Amblard, F., Sibertin-Blanc C.: Modélisation multi-agents pour la formalisation de théories sociologiques: Le cas de la sociologie de l’action organisée appliquée à l’étude de la dynamique du pays Quercy-Rouergue. In Proceedings of the SMAGET, France (2005).Google Scholar
  11. Mailliard, M., Audras, S., Casula, M.: Multi Agents Systems based on Classifiers for the Simulation of Concrete Action Systems. In Proceedings of the 1st EUropean Workshop on Multi-Agent Systems (EUMAS), Oxford University, (2003).Google Scholar
  12. Molm, L.: Affect and social exchange: satisfaction in power-dependence relations. American Sociological Review, vol. 56, (1991).Google Scholar
  13. Sibertin-Blanc, C., Amblard, F., Mailliard, M.: A coordination framework based on the Sociology of the Organized Action, In Proceedings of the From Organization to Organization Oriented Programming in MAS, AAMAS, Utrecht University, (2005).Google Scholar
  14. Simon, H.: The sciences of the artificial, MIT Press, 3rd edition (1996).Google Scholar
  15. Takadama, K., and al.: Cross-Element Validation in Multiagent-based Simulation: Switching Learning Mechanisms in Agents. JASSS vol. 6, no. 4, (2003).Google Scholar

Copyright information

© Springer 2007

Authors and Affiliations

  • M. Mailliard
    • 1
  • F. Amblard
    • 1
  • C. Sibertin-Blanc
    • 1
  • P. Roggero
    • 1
  1. 1.IRIT, LEREPS-CIRESSUniversité de Toulouse 1Toulouse CedexFrance

Personalised recommendations