Reinforcement Hybrid Evolutionary Learning for TSK-type Neuro-Fuzzy Controller Design

Abstract

This paper proposes a recurrent TSK-type neuro-fuzzy controller (TNFC) with reinforcement hybrid evolutionary learning algorithm (R-HELA). The proposed R-HELA combines the compact genetic algorithm (CGA) and the modified variable-length genetic algorithm (MVGA) to perform the structure/parameter learning for constructing the TNFC dynamically. The evolution of a population consists of three major operations: group reproduction using the compact genetic algorithm, variable two-part crossover, and variable two-part mutation. Illustrative example is conducted to show the performance and applicability of the proposed R-HELA method.