Training ANFIS Parameters with a Quantum-behaved Particle Swarm Optimization Algorithm

摘要

This paper proposes a novel method for training the parameters of an adaptive network based fuzzy inference system (ANFIS). Different from previous approaches, which emphasized on the use of gradient descent (GD) methods, we employ a method based on. Quantum-behaved Particle Swarm Optimization (QPSO) for training the parameters of an ANFIS. The ANFIS trained by the proposed method is applied to nonlinear system modeling and chaotic prediction. The simulation results show that the ANFIS-QPSO method performs much better than the original ANFIS and the ANFIS-PSO method.