An integrated approach for the optimal design of a dynamic system and its controller

Abstract: "Optimal control system design traditionally involves the adjustment of state and input variables to optimize a specified performance index. However, in practice there may exist variables, such as plant parameters, that offer extra freedom in the design of systems that can achieve even better performance. In this paper, we describe an integrated approach for dynamic system/control system design which can satisfy a variety of constraints and allows for the introduction of three additional types of variables in the optimization process. The approach we outline is based on a direct search method which can be used in conjunction with standard optimal control algorithms such as the Riccati equation for unconstrained linear systems. The method is used to generate optimal state and input trajectories for a linear mass-spring-damper system."