THE USE OF EVOLUTIONARY OPTIMIZATION IN FUZZY TSK MODEL IDENTIFICATION

This paper is focused on identification of fuzzy TSK model using evolutionary algorithms. The use of genetic algorithm with newly developed genetic operators is described. Proposed solution gives a possibility to have fully automatic and homogeneous algorithm to identify all structure and parameters of TSK model in one run. There are also given some remarks about the use another evolutionary algorithms like Particle Swarm optimization are also given. Quality index, which takes into account TSK model specific and allows to proper estimation of consequent coefficients for all rules is proposed. An example of dynamic model identification for pneumatic drive is included.