Inverse-model predictive control based on INFUMO-BB-BC optimization

Abstract

Abstract This paper presents a method of advanced predictive control combining a contemporary optimization procedure Big Bang - Big Crunch (BB-BC) and the use of an interval fuzzy model (INFUMO). The method is founded on the inverse-fuzzy-model-based controller where adaptation of the model to the process is done in each calculation step. The INFUMO was used to identify a band of possible process parameter values and hence to limit the optimization search space, which resulted in a considerable reduction of computational load. The benefits of the method were illustrated on a simulation model with high-grade uncertainties and rapid changes of parameters.