Computation of the Nash equilibrium selected by the tracing procedure in \(N\)-person games (Q700098): Difference between revisions

One of the purposes of the equilibrium selection theory [see \textit{J. C. Harsanyi} and \textit{R. Selten}, A general theory of equilibrium selection in games, Cambridge, MIT Press (1988; Zbl 0693.90098)] is to select equilibria since almost every non-trivial noncooperative game possesses many of them. The linear tracing procedure is a method that, given an \(n\)-person noncooperative game and the initial beliefs of all players about the strategies played by the others, selects Nash equilibria of the game. When the linear tracing procedure yields a unique outcome, the procedure is said to be well defined. Given the prior beliefs of the players, the linear tracing procedure is well defined for almost all \(n\)-person noncooperative games [see \textit{J. C. Harsanyi}, Int. J. Game Theory 4, 61-94 (1975; Zbl 0319.90078)]. When the number of players is greater than two the Nash equilibrium equations are nonlinear and one has to use numerical approaches. Different algorithms to approximate Nash equilibrium have been proposed. This paper presents an homotopy algorithm that computes the Nash equilibrium selected by the linear tracing procedure, but one of the important advantages with respect to others is the game-theoretic interpretation of the entire path of the strategy vectors that the algorithm generates. Another feature is that the starting point is endogenously determined by the game's payoffs. Besides, even when the linear tracing procedure is not well defined, the algorithm computes an approximation of a Nash equilibrium arbitrarily closed to the set of strategies generated by the linear tracing procedure.