State Difference Sensitivity-based Control Vector Parameterization Refinement Approach for Optimal Control Problems

Abstract

To improve the optimization performance of control vector parameterization (CVP) method for optimal control problems, a refinement CVP method based on the state difference sensitivity is proposed. According to the influence of the state vector on the control vector, the state difference of the adjacent time grid nodes makes the control vector sensitive, so new time grid nodes can be selected. Finally, the control parameters obtained with the non-uniform grid are analyzed accordingly. A new NLP problem is obtained with the refined time grid and simulation tests are carried out on a classical optimal control problem. The results show that the proposed method can effectively shorten the calculation time and increase the calculation accuracy compared with the traditional CVP method.