基于模糊推理的带收缩因子的粒子群算法

An Adaptive Constriction Factor Particle Swarm Optimization Algorithm Using Fuzzy Inference Engine

为了提高粒子群算法的搜索性能,本文提出了一种基于模糊推理的改进方法。通过模糊推理调整算法参数:对学习因子c1和c2进行自适应调整,平衡粒子向自身经验和向群体经验学习的能力;收缩因子χ也通过模糊规则随之调整,从而平衡粒子的全局搜索能力和局部搜索能力。通过对多个基准测试函数的进行仿真实验,并与标准的粒子群算法和带收缩因子的粒子群算法相比较,结果表明改进后的算法的性能更好,尤其是对具有多个局部极值点且极值相差不大的多峰函数的优化更有效。该算法具有一定的实际应用价值。

In order to improve the performance of Particle Swarm Optimization （PSO） in the search process, an adaptive fuzzy method is proposed. In this paper, fuzzy inference engine is used to adjust the learning factors. On the one hand, it can adjust the ability of particles to learn from their own experience and learn from group experience- On the other hand, the constriction factor also changes with the values of C1 and C2 , so as to hal ance the global search capability and the local development capacity of particles. By simulating multiple bench mark functions and comparing with the standard PSO and the PSO with constriction factor, a better conver gence effect is achieved. This algorithm has a certain practical value.