Abstract: For improving the performance of particle swarm optimization (PSO) in optimization simulation, a multi-strategy cooperative evolutionary PSO based on Cauchy mutation strategy is proposed. The new algorithm divides the whole swarm into three sub-swarms. A part of particles is selected to Cauchy mutation with a certain probability, and the rest of particles adjust their exploitation and exploration by different evolutionary strategies (large-scale search strategy, local search strategy, and adaptive velocity updating strategy). The sub-swarms share their information to achieve cooperation. Three strategies are used to optimize three test functions, and the result shows the advantages of three strategies. The simulation experiment uses the soft lunar landing problem as the simulation model to optimize the trajectory. Simulation results indicate that the performance of improved PSO is superior to other PSO. The simulation uses OpenMP to parallelization optimization, which improves the efficiency of the algorithm.

Key words: particle swarm optimization, multi-strategy, exploitation, exploration, Cauchy mutation strategy, OpenMP parallelization