Particle Swarm Optimization Algorithm Based On Human Behavior And Application

Since the particle swarm optimization algorithm is proposed,it has become a hot topic for many scholars because of its simple parameters and fast convergence.There are some shortcomings in the algorithm: convergence precocity,poor adaptive ability,easy to fall into local optimization,and the algorithm has limitations depending on specific optimization problems.In view of these shortcomings,it is of theoretical and practical significance to study the particle swarm optimization algorithm in depth,how to improve the diversity and adaptive ability of population,control the global search and local mining capacity of population and its own,reduce the sensitivity of parameters and extend the scope of application optimization.In view of these problems,the following work has been done in this dissertation:(1)A human behavior particle swarm optimization algorithm with adaptive disturbance(HPSO AD)is proposed.In HPSO AD,an adaptive disturbance probability based on global optimal stop counter is proposed.Then,the disturbance frequency and dimension are determined according to the adaptive disturbance probability.The individual and global optimal of the set are randomly searched to generate candidate solutions.Finally,a greedy selection mechanism is proposed to retain the elite in the population.The HPSO is verified by using 40 reference functions composed of simple single peak function,basic multi peak function and rotation function good performance of HPSO AD,experimental results and statistical analysis show that HPSO AD has high convergence precision and fast convergence speed.At the same time,four constrained engineering optimization problems are used to test the ability of HPSO AD to solve practical problems.The statistical optimization results show that HPSO AD greatly improves the performance of practical engineering design problems.(2)A hybrid integrated learning and dynamic multi subgroup human behavior particle swarm optimization algorithm(HCLDMS-HPSO)is proposed.In HCLDMS-HPSO,the improved CL and DMS update strategy are proposed to update the particle location information.Finally,Gaussian mutation operator is introduced as the local optimization operator in global (70) to enhance the local search ability of HCLDMS-HPSO algorithm.The performance of HCLDMS-HPSO is verified by using 29 benchmark functions of the algorithm.The results are compared with other evolutionary algorithms.The statistical analysis results show that HCLDMS-HPSO has high convergence accuracy and fast convergence speed.