INTELLIGENT ADAPTIVE DYNAMIC MATRIX CONTROL

Abstract

Listen

Translate article

This paper presents an approach to adapt the suppression and scaling factor from a single input single output (SISO) dynamic matrix controller (DMC) thought a multiobjective optimization algorithm. To optimize, a nonlinear neural network (NN) process model is used, combined with a multiobjective evolutionary algorithm called SPEA II (Strength Pareto Evolutionary Algorithm) to find better controller parameters for the plant each sample time. Also every sample time, a decision over the resultant pareto front from the multiobjective optimization process are taken using a simple decision approach.