Project B6:
Learning, Evolution and Local Interaction

Director: Prof. Avner Shaked, Ph.D.

We analyse learning rules in populations with a Local Interaction structure - where an individual meets his neighbours only. We study the population behaviour for a given learning rule and study the optimality and the evolution of such learning rules. The Local Interaction structure affects the information the individuals have and, therefore, the speed and direction the evolutionary process takes.

In particular, we study the following topics:

  1. Partnership for Survival (Shaked)
    Individuals may tend to be more cooperative if their neighbours may die and disappear when not helped.

  2. Neighbourhood & Kinship (Shaked)
    Individuals may learn to treat their neighbours as their kin when behaviour spreads by imitation.

  3. Optimal Imitation under Imperfect Information (Schlag)
    We will look for optimal learning rules for boundedly rational agents when only partial information is observable, e.g. play in an extensive form game.

  4. Evolution in Networks (Schlag)
    Competition among learning rules and consequences for information processing and asymptotic learning will be analysed.

  5. Evolution and Equilibrium Selection (Schlag)
    Predictions of evolutionary learning processes in economic settings, e.g., commitment games, pricing and product differentiation, will be derived.

  6. The Evolution of Social Insurance in Social Networks (Pollock)
    How social networks which distribute support (monetary or otherwise) in times of crisis are formed and maintained.

  7. Punishment and Cheating in Coordination Mechanisms (Pollock)
    Analysis of a biological system (ant queen foundresses) involving an "altruistic" coordinated role, with experiments and simulations.

  8. Market Organization and Searching (Herreiner)
    Concentrates on questions like how markets organize, how buyers and sellers meet, how they learn from past experience.

  9. Learning, Evolution and Local Interaction (Herreiner)
    Analyzes how local learning and interaction rules determine the evolution of a population.

Address:

University of Bonn
Institut für Gesellschafts- und Wirtschaftswissenschaften
Wirtschaftstheoretische Abteilung III
Adenauerallee 24-26
D - 53113 Bonn / Germany

Phone: +49+228+73 7996
Fax: +49+228+73 7900

Researchers:

08.09.2000, © Webmaster