【摘要】 为提高不同粒子群之间的协作性能,减小粒子群局部收敛现象,提出了一种"逆群"协作粒子群方法,利用不同动作特性的两个群之间的相互协作,来提高粒群算法的性能。在此方法中,一个粒子群追随较好的位置,另一个群是远离较差的位置,与传统的粒子协作方法不同,在粒子飞翔的过程中,两群不仅交换较好的信息,同时也交换较差的信息,通过充分利用这两种信息,两群共同完成任务的求解。在完成算法设计的基础上,对两个群的运动特性和算法的收敛性进行分析,通过函数的优化实验验证了方法的有效性。更多还原

【Abstract】 In order to improve the cooperation performance and reduce local convergence probability of particle swarms, a novel inverse cooperation particle swarm optimization was proposed, and the performance of particle swarm optimization was improved by cooperation of the different swarms. In the method, one swarm tracked the best position and the other moved away from the worst position, different from conventional cooperation particle swarms, not only the best information was exchanged but the worst information was flowed between them in the flying process, and a corporate task was performed by the two kinds information. The movement characters and convergent character were analyzed based on the designing of the algorithm, and the effectiveness of the method is indicated by simulation experiments of function optimization.更多还原