Homotopy methods to compute equilibria in game theory

P.J.J. Herings*, Ronald Peeters

*Corresponding author for this work

    Research output: Contribution to journalArticleScientificpeer-review

    61 Downloads (Pure)

    Abstract

    This paper presents a survey of the use of homotopy methods in game theory. Homotopies allow for a robust computation of game-theoretic equilibria and their refinements. Homotopies are also suitable to compute equilibria that are selected by various selection theories. We present the relevant techniques underlying homotopy algorithms. We give detailed expositions of the Lemke-Howson algorithm and the van den Elzen-Talman algorithm to compute Nash equilibria in 2-person games, and the Herings-van den Elzen, Herings-Peeters, and McKelvey-Palfrey algorithms to compute Nash equilibria in general n-person games. We explain how the main ideas can be extended to compute equilibria in extensive form and dynamic games, and how homotopies can be used to compute all Nash equilibria.

    Original languageEnglish
    Pages (from-to)119-156
    Number of pages38
    JournalEconomic Theory
    Volume42
    Issue number1
    DOIs
    Publication statusPublished - Jan 2010

    Keywords

    • Homotopy
    • Equilibrium computation
    • Non-cooperative games
    • Nash equilibrium
    • NASH EQUILIBRIA
    • DIFFERENTIABLE HOMOTOPY
    • EFFICIENT COMPUTATION
    • TRACING PROCEDURE
    • 2-PERSON GAMES
    • ALGORITHM

    Fingerprint

    Dive into the research topics of 'Homotopy methods to compute equilibria in game theory'. Together they form a unique fingerprint.

    Cite this