【摘要】 针对基本微粒群优化算法(PSO)存在陷入局部最优的问题,提出一种基于排列的改进微粒群算法(RPSO)。该算法对每次迭代过程中的个体历史最优解按照适应值的优劣顺序排列,然后选择若干个较优的个体历史最优解作为候选解,再以概率方式在候选解中确定群体历史最优解的位置。RPSO算法使基本PSO算法易于陷入局部最优的问题,得到有效的缓解。为了分析算法的性能,对几种典型的非线性函数进行了测试。实验结果表明,RPSO算法比基本PSO算法具有更好的寻优能力。更多还原

【Abstract】 A rank-based particle swarm optimization (RPSO) algorithm is proposed to overcome the shortcoming of particle swarm op- timization (PSO) algorithm, which is easy to fall into local optima. In the proposed algorithm, the best previous solutions of all particles from each iteration are selected and ranked according to their fitness. And the solutions with higher fitness are chosen as candidates for the possible solutions. The optimum solution, and thus its location, was then determined by choosing among the candidates with a probabilistic method. The performance of the new algorithm is compared with the PSO method when applied to benchmark non- linear programming problems. The new approach is seen to perform better in finding the global optimum than PSO.更多还原