Nash equilibrium in games with incomplete preferences (Q2486960)

This paper considers Nash equilibria in games where agents have incomplete preferences representable by vector-valued utility functions via the usual partial order in \(\mathbb{R}^n\). For example, a firm may be interested in profits \textit{and} sales, and may not be able to rank two vectors \((\pi,s)\) and \((\pi+h,s+k)\) if \(hk<0\). If a preference is represented by \(u=(u_1, u_2,\dots, u_n)\), then the preference represented by the (real-valued) function \(\sum \beta_j u_j\) with positive \(\beta\) is a completion of the original one. The paper proves and exploits a result linking the Nash equilibria of the original game and those of the games obtained via these `linear' completions. The author shows that allowing oligopolistic firms to have the above mentioned multi-criteria kind of preferences and applying the incomplete preferences approach to the resulting strategic games yields novel and interesting solutions to the well-known delicate existence problems in oligopoly theory. On the negative side, the author observes that given the large multiplicity of equilibria in incomplete-preference games (given any profile, preferred deviations are harder to find), it would be desirable to have clear-cut results on the most common refinements, e.g.\ perfection; but as she hints at, such direction seems to be less promising than one would hope.