Exact penalty functions for generalized Nash problems

Abstract

We propose the use exact penalty functions for the solution of generalized Nash equilibrium problems (GNEPs). We show that by this approach it is possible to reduce the solution of a GNEP to that of a usual Nash problem. This paves the way to the development of numerical methods for the solution of GNEPs. We also introduce the notion of generalized stationary point of a GNEP and argue that convergence to generalized stationary points is an appropriate aim for solution algorithms.