Optimal control problems with $L^0(\Omega)$ constraints: maximum principle and proximal gradient method

TL;DR

This paper develops necessary optimality conditions and a proximal gradient algorithm for control problems constrained by the measure of the control support, extending classical maximum principles to non-convex $L^0$ constraints.

Contribution

It introduces a Pontryagin maximum principle for $L^0$ constrained controls and analyzes a proximal gradient method ensuring convergence to feasible solutions.

Findings

Maximum principle in integral and pointwise forms for $L^0$ constraints

Proximal gradient method converges to feasible solutions under certain assumptions

Sequences of iterates have strongly converging subsequences satisfying optimality conditions

Abstract

We investigate optimal control problems with $L^0$ constraints, which restrict the measure of the support of the controls. We prove necessary optimality conditions of Pontryagin maximum principle type. Here, a special control perturbation is used that respects the $L^0$ constraint. First, the maximum principle is obtained in integral form, which is then turned into a pointwise form. In addition, an optimization algorithm of proximal gradient type is analyzed. Under some assumptions, the sequence of iterates contains strongly converging subsequences, whose limits are feasible and satisfy a subset of the necessary optimality conditions.