Towards global solutions of optimal discrete-valued control problems

Abstract

This paper proposes a new heuristic approach for solving optimal discrete-valued control problems. We illustrate the approach with an existing hybrid power system model. The problem of choosing an operating schedule to minimize generator, battery, and switching costs is first posed as a mixed discrete dynamic optimization problem. Then, a discrete filled function method is employed in conjunction with a computational optimal control technique to solve this problem. Computational results indicate that this approach is robust, efficient, and can successfully identify a near-global solution for this complex applied optimization problem despite the presence of multiple local optima.