Fuzzy Systems Based On Particle Swarm Optimization Algorithm

Fuzzy systems is one of the most famous application of fuzzy logic and fuzzy set theory,they can approximate different sorts of nonliner relationship using the expert knowledge as well as observed data.its heart is consisted of "if-then" fuzzy rules.Fuzzy systems can approxiimate different sorts of nonlinear relationship using the expert knowledge as well as observed data.But fuzzy systems are interpretable,their structure and parameters can be explained by fuzzy rules.Then fuzzy systems often suffer from several problems such as local optimum,Slow convergence speed,the curse of dimensionality.The concept Particle swarm optimizationis is simple with a fast convergence rate, it is easy to be implemented.The purpose of this study is to use particle swarm and fuzzy clustering to design fuzzy system.Particle Swarm Optimization is a popution-based random optimization algorithm,comes from the study on the movement behavior of birds. It as a new intelligent optimization algorithm.which is used to seek the optimal solution of complex problems.But PSO has some inherent defects,such as the weak local searching abilility of the algorithm in the later. The initial parameters of Mamdani Fuzzy Neural Network(FNN) are generated by Fuzzy C-means clustering based on PSO, then are optimizated by using PSO. Finally, Gradient descent method is adopted for further optimizating the parameters so that the fuzzy rules can be automatically adjusted, modified and improved. Numerical experiments show that the presented algorithm improves the approximation ability of Mamdani FNN.