Optimal analysis of the weighted matrices in LQR based on the differential evolution algorithm

The general method of determining the weighted matrices Q and R in the linear quadratic regulator (LQR) is trial and error via simulation. The way is simple but inefficient, and it is more complicated for high-dimension systems. In order to overcome these shortcomings, this paper introduces an optimization algorithm, which is the Differential Evolution Algorithm (DEA), to optimize the weighted matrices of the LQR. Through the strategy of real coded for the weighted matrices Q and R, the best weighted matrices and the feedback matrix K can be obtained. By simulation analysis of the selected system, the results showed that a better effect is obtained based on the DEA to optimize the weighed matrices of the LQR. The system's regulation time and the peak value also have a big improvement.