Abstract: For the question that standard particle swarm optimization algorithm is easy to fall into local optimal solution, poor in global search ability because of the lost of diversity of particles. This article introduces a new particle swarm optimization algorithm based on campaign leader strategy. It starts the campaign mechanism when the leadership of the global optimal particle loses. It selects a new leader which has more leadership skills to enhance the global search ability of the algorithm after optimizing the elite particles with bacterial foraging algorithm chemotaxis operator. Experimental studies are carried out on four classical functions, and the computational results show?that the algorithm prevents the premature phenomenon effectively and improves the convergence precision without affecting the convergence speed.

Key words: particle swarm optimization, campaign leader, bacterial foraging algorithm, Metropolis criterion

摘要： 针对标准粒子群算法由于粒子多样性的大量丧失而导致的算法易陷入局部最优解，收敛精度不高的问题，提出一种基于竞选领导策略的改进粒子群算法，该算法在全局最优粒子的领导能力丧失时，通过引进细菌觅食算法的趋化算子对精英粒子进行优化，然后选出更具领导能力的粒子作为新的领导粒子来带领种群跳出局部最优解，以增强算法的全局搜索能力。通过四个典型函数的测试，结果表明改进算法在较好保留了标准粒子群算法快速收敛优点的前提下，有效地预防了早熟现象的产生，提高了收敛精度。

关键词: 粒子群优化算法, 竞选领导, 细菌觅食算法, Metropolis准则