关键词: 粒子群优化算法, 人工生命, 基于仿真的优化, 参数选取, 早熟, 基于主体的建模和仿真

Abstract: Particle Swarm Optimization（PSO） has the characteristics that the parameter number is few and the basic operation is simple，so it has been the hotspot of the intelligent optimization.However，PSO possesses the defects that it can be premature easily，its searching precision is not high and the performance is affected by the parameters deeply.Aiming at the deficiencies，within the framework of simulation based optimization，this paper models and simulates the artificial life model using multi-agent that integrates the traditional global PSO with local PSO，and the hybrid intelligent optimization algorithms function is used as the optimization engine.The parameter selection in PSO is discussed with emphasis.A series of benchmark functions are tested and analyzed，and high performance is obtained，which is supposed to demonstrate the feasibility and creditability of the above methodology.

Key words: particle swarm optimization, artificial life, simulation based optimization, parameter selection, premature, agent based modeling and simulation