A multi-objective predictive control method for model mismatch

In this paper, a multi-objective predictive control method was presented for model mismatch of nonlinear system. Predictive control was restored to multi-objective control, then more Pareto solutions were obtained and pseudo-weighted vector were computed by using multi-objective particle swarm optimization algorithm. Finally, the model error probability distribution was used to select the optimal solution as a control input for controlling object. Proposed three control performance index, that smoothness index, the average error index and quadratic performance index, based on three index a nonlinear system was simulated. Results show that when the nonlinear system model mismatch, the multi-objective predictive control performance better, and a good control effect was achieved.