Abstract:
The paper considers three basic phases of fuzzy systems construction: expert evaluation, structure identification, parameter estimation. Expert evaluation includes: selection of fuzzy model type; choice of $t$-normal functions to set the fuzzy logic operations; choice of a fuzzy logic inference. For structure identification the fuzzy clustering method and iterative algorithm are offered. For parameters optimization the following methods have been chosen: genetic algorithm, ant colony algorithm, particle swarm optimization, simulated annealing.
Keywords:fuzzy system identification, metaheuristics, simulated annealing, genetic algorithm, ant colony algorithm, particle swarm techniques.