Research On Adaptive Particle Swarm Algorithm Based On Fuzzy Inference

Intelligent optimization algorithm is one of the research focuses in recent years.How to improve the search performance of such algorithms in complex environments is a key issue in this field.Particle swarm optimization(PSO)is a swarm intelligence optimization algorithm.Compared with other optimization algorithms,PSO has so many advantages.For example,it has fewer parameters and is simple to implement.These advantages make it attracting the attention of many scholars.At present,PSO has been widely used in fuzzy control,vehicle routing,pattern recognition,product operation investment,advertising optimization,supply chain systems and so on.However,it also has some shortcomings such as being prone to premature convergence,slow convergence in the later periods of iteration,and poor robustness,especially when dealing with high-dimensional complex problems.In order to further improve the search performance of PSO,based on deeply studies on the PSO with contraction factor,the thesis proposed an adaptive PSO using fuzzy reasoning method in which algorithm's parameters are adjusted adaptively through fuzzy inference.Firstly,the learning factors are adaptively adjusted in order to balance the ability of particles to learn from their own experience and learn from group experience.Secondly,the contraction factor is also adjusted by fuzzy rules to balance the global search ability and local search ability of particles.By simulation experiments on multiple benchmark functions and comparing them with both the standard PSO and the original PSO with contraction factor,the results show that the improved algorithm performs better.In addition,an adaptive quantum PSO is proposed.According to the characteristics of the algorithm,the formula of potential well center is improved,and the method of fuzzy reasoning is used to update the algorithm by introducing both of health degree and particle diversity.The computer simulation results and statistical analysis show that the improved algorithm has a good performance.