Adaptive Switching Structure Detection for the Solution of Dynamic Optimization Problems

Abstract

Recent research in the field of dynamic optimization is striving for increased robustness and efficiency of the numerical solution methods. In addition, the physical interpretation of the optimal solution, with respect to the control switching structure, has been gaining increased interest, because it provides valuable insight that is useful for applying the optimization results on the process. In a previous publication, we have addressed this issue by proposing an automatic switching structure detection method. This paper presents an extension of this idea by combining the structure detection approach with a control profile adaptation technique. The resulting adaptive structure detection approach enables a faster numerical solution, which is more likely to contain the correct switching structure. This further enhances the numerical solution of dynamic optimization problems. The concept is illustrated by means of a small example problem and a large-scale industrial case study.