GENERAL THEORY FOR STACKELBERG GAMES WITH PARTIAL STATE INFORMATION.

This paper presents a general method for derivation of a tight lower bound on the Stackelberg cost of the leader in general two-person deterministic games with partial dynamic state information. The method converts the original dynamic Stackelberg problem into two open-loop optimization problems whose solutions can readily be obtained using the standard techniques of optimization and optimal control theory. When applied to the class of linear-quadratic dynamic games with partial dynamic information, defined on general Hilbert spaces, each one of these open-loop optimization problems becomes a quadratic programming problem with linear constraints, thus allowing for an explicit computation of the Stackelberg cost value.