Optimization of weight selection for H_∞ control using evolutionary algorithms

In the process of the H_∞ control design, the selection of appropriate weighting matrices directly reflects the stability and performance of a control system. This paper presents a new method for optimizing the selection of weighting matrices in the H_∞ control design using evolutionary algorithms. The design objectives are converted to the inequality constraints and the selection of weighting functions is expressed as a multi-objective optimization problem which is solved using evolutionary algorithms. The simulation results show that, the satisfactory weighting matrices can be effectively searched out and thus the H_∞ controller satisfying given frequency-and time-domain performance indexes can be found by using the proposed approach.