改进反向粒子群算法及其在噪声中的应用

Opposition-based particle swarm optimization with random factor for noisy problems

粒子群优化算法规则简单,收敛速度较快,但易陷入局部最优值,在噪声问题中也显示出较差的寻优能力.针对算法存在的不足,本文结合反向学习机制较快的学习速度及优化能力,并在算法进化过程中引入交叉因子,提出一种新的改进的反向粒子群算法(COPSO).实验表明,该算法在噪声问题中要优于反向粒子群算法.

Particle swarm optimization was proved to have the fast convergence and simple regulatioza. However itwas easily trapped into the local optima and shown the poor ability in the noisy problems. An improved algorithm is presented in this paper. Opposition-based learning which accelerate the learning and searching process and corm-factor are employed in the algorithm. Experimental results on the test functions are shown that the algorithm is superior to OPSO in the noise problems.