Stackelburg solution for two-person games with biased information patterns

Abstract

A strategy suggested by Stackelberg for static economic competition is considered and extended to the case of dynamic games with biased information pattern. This strategy is reasonable when one of the players knows only his own cost function but the other player knows both cost functions. As with Nash strategies for nonzero-sum dynamic games open-loop and feedback Stackelberg strategies for dynamic games could lead to different solutions, a phenomenon which does not occur in optimum control problems. Necessary conditions for open-loop Stackelberg strategies are presented. Dynamic programming is used to define feedback Stackelberg strategies for discrete-time games. A simple resource allocation example illustrates the solution concept.