基于对立学习的PSO算法研究

A Research into PSO Algorithm Based on Stochastic Opposition Learning

为进一步提高PSO算法的优化效率,加速寻优过程,提出基于随机对立策略的PSO算法,包括QOP-SO和QRPSO。这两种算法在种群初始化阶段采用随机对立学习方法,并在进化过程中用随机对立学习进行种群动态跳跃,以提高产生解的质量。利用6个测试函数对算法的效率进行检验,将其与标准PSO和OPSO算法进行对比,结果表明,新算法具有更快的收敛速度和更高的求解精度。

In order to improve the performance of Particle Swarm Optimization（PSO） and accelerate the convergence speed,two improved PSOs named as QOPSO and QRPSO are proposed based on stochastic opposition.The two different kinds of stochastic opposition learning are employed in population initialization and generation jumping,which can improve the quality of solutions.Simulation results on six benchmark functions show that the two proposed algorithms are superior to the SPSO and the OPSO.