Hierarchical FLS Design Using Multi-Agent Genetic Approach

Abstract—This paper proposes an automatic design method for Hierarchical Fuzzy Logic Systems (HFLS) using genetic algorithms parallelized under a multi-agent approach. To save onhierarchical level numbers, genetic exploration is adopted. Distributed implementation by multi-agent approach improves the genetic algorithms (GA) convergence by reducing temporal and rule complexities. Dynamic agents interact to reach the balance interpretability-precision of a HFLS model.