Volume 5, Number 2, May 2009, pp. 237-259

Ryoichi Nishimura, Shunsuke Hayashi and Masao Fukushima
Key words:
N-person non-cooperative game, incomplete information, robust Nash equilibrium, robust optimization, complementarity problem
Mathematices Subject Classification: 91A06, 91A10, 90C33
ONLINE SUBSCRIPTION (Institutional Subscription Only)
Copyright© 2009 Yokohama Publishers
Back

Abstract:
In this paper we propose a general framework of distribution-free models for N-person non-cooperative games with uncertain information. In the model, we assume that each player's cost function and/or the opponents' strategies belong to some uncertainty sets, and each player chooses his/her strategy according to the robust optimization policy. Under such assumptions, we define the robust Nash equilibrium for N-person games by extending some existing definitions. We present sufficient conditions for existence and uniqueness of a robust Nash equilibrium. In order to compute robust Nash equilibria, we reformulate certain classes of robust Nash equilibrium problems to second-order cone complementarity problems. We finally show some numerical results to discuss the behavior of robust Nash equilibria.
Robust Nash equilibria in N-person non-cooperative games: Uniqueness and reformulation