Online tuning scheme for generalized predictive controller via simulation-optimization

From the process control point of view, it is difficult to find out the optimal parameters for the control system based on the single quadratic performance index, which is used in the standard predictive control algorithm. The fuzzy decision-making function is investigated in this paper. First, M control actions are obtained by an unconstrained predictive control algorithm, and fuzzy goals and fuzzy constraints can then be calculated and the global satisfactory degree is obtained by fuzzy inference. Moreover, the weighting coefficient /spl lambda/ in the cost function is tuned using simulation-optimization according to the fuzzy criteria.