Real-time optimization of perturbed control systems with linear control (Q2745862)

The paper discusses the need for a fast numerical method to compute real-time control approximations during the process so that optimality conditions and prescribed constraints are preserved even in the presence of disturbances. For finite-dimensional optimization problems, an approach based on suitable discretization strategies for optimal control problems with linear controls is proposed. As a first step the problem is solved by a direct optimization method. Once an optimal solution is found, the switching structure is extracted to formulate a low-dimensional switching interval problem where only switching points and discretized controls on singular subarcs are optimized. By means of this transformation, the number of optimization variables decreases enormously. As an example a batch reactor with elevent states and two inputs is considered.