Supervised and Reinforcement Group-based Hybrid Learning Algorithms for TSK-type Fuzzy Cerebellar Model Articulation Controller

Abstract

In this study, a TSK-type fuzzy cerebellar model articulation controller (T-FCMAC) model with a group-based hybrid learning algorithm (GHLA) is proposed for solving various problems. The proposed T-FCMAC model is mainly derived from the traditional CMAC model and the Takagi-Sugeno-Kang (TSK)-type fuzzy model. In the supervised learning, the proposed GHLA combines the improved quantum particle swarm optimization (IQPSO) and the Nelder-Mead (NM) method for adjusting the parameters of a T-FCMAC model. The fuzzy C-mean clustering technique is adopted to improve the performance of quantum particle swarm optimization. The grouping concept is also used to reform the search ability and greatly increase the convergence speed. In addition, when exact training data may be expensive or even impossible to obtain in some real-world applications, a reinforcement GHLA (R-GHLA) is proposed. Experimental results have been conducted to illustrate the performance and applicability of the proposed GHLA and R-GHLA.