基于群能量恒定的粒子群优化算法

Particle swarm optimization based on swarm energy conservation

针对标准粒子群优化(PSO)算法在寻优过程中容易出现早熟的情况,提出一种群能量恒定的粒子群优化(SEC-PSO)算法.算法根据粒子内能进行动态分群,对较优群体采取引入最差粒子的速度更新策略,对较差群体采取带有惩罚机制的速度更新策略,由其分担由于较优群体速度降低而产生的整群能量损失,从而有效地避免了PSO算法的早熟.典型优化问题的仿真结果表明,该算法具有较强的全局搜索能力和较快的收敛速度,优化性能得到显著的提高.

To the problem of premature convergence frequently appeared in standard particle swarm optimization（PSO） algorithm,an improved algorithm,swarm energy conservation particle swarm optimization （SEC-PSO）,is proposed. The population is partitioned into two sub-swarms according to the energy of the particles. The good particles update their velocity according to the strategy with the worst particle. The bad particles update their velocity according to the strategy with penalty mechanism,and bear the swarm energy loss which is generated by speed reduction of the good population. Thus,the problem of premature convergence of the PSO algorithm is prevented. Simulations results for several typical test functions show that SEC-PSO possesses more powerful global search capabilities,better convergence rate and better performance of optimization.