A kind of steady-state target robust Optimal methods and device for model mismatch

This application discloses a kind of steady-state target robust Optimal methods and device for model mismatch.This method initially sets up the transmission function mismatch model of the relationship between prediction error, control Delta and the mismatch gain for describing controlled variable, and then recognizes to the mismatch gain in mismatch model, to determine the range set of mismatch gain.Then, according to the range set of mismatch gain and the steady-state gain of Controlling model, the range set of Controlling model overall gain is determined, so that the range set based on overall gain, the optimal steady-state target of controller is determined using default optimizing algorithm.Compared with prior art, present invention the case where there are mismatches in view of object model and controller model when to optimal steady-state target optimizing, thus improve the stability and robustness of model predictive controller.