Semiexhaustive search for solving nonlinear optimal-control problems

It is difficult to obtain optima, especially global optima, for optimal control problems that are highly nonlinear and involve state constraints. The recommended practice in the literature is to use different methods and then to implement the best solution found. To aid this process, a method that uses a semiexhaustive search to solve such problems is proposed in this paper. Its performance is presented on six example problems from the chemical engineering literature. As reported in the literature, iterative dynamic programming (IDP) found better optimums for several nonlinear problems. The proposed method not only found the IDP’s optimums, but also, in some cases, it found even better optimums.