A maximum principle for nonsmooth optimal-control problems with state constraints

Abstract Optimality conditions are derived in the form of a maximum principle governing solutions to an optimal control problem which involves state constraints. The conditions, which apply in the absence of differentiability assumptions on the data, are stated in terms of Clarke's generalized Jacobians. Although not the most general available, the conditions are derived by a novel method: this involves removal of the state constraints by introduction of a penalty term and application of Ekeland's variational principle.