基于改进萤火虫算法的多模函数优化

MULTIMODAL FUNCTION OPTIMISATION BASED ON IMPROVED GLOWWORM SWARM OPTIMISATION

为了提高萤火虫算法GSO(Glowworm Swarm Optimization algorithm)多模函数优化性能,针对GSO峰值发现率低、收敛速度慢和求解精度不高的缺点,提出萤火虫个体可自适应搜索峰值且移动步长可变的改进萤火虫算法IGSO(Improved Glowworm Swarm Optimization algorithm)。IGSO引入尝试性移动策略以增强算法的搜索能力,同时,以邻域平均距离为参考,对个体移动步长进行调整。采用典型多模函数进行测试,实验结果表明,IGSO峰值发现率高,收敛速度快且求解精度高,比GSO具有更优的多模函数优化性能。

In order to improve the performance of multimodal function optimisation with glowworm swarm optimisation （ GSO）, and to solve the problems of GSO in low peaks discovery rate, slow convergence speed and low computational accuracy, we propose an improved glowworm swarm optimisation （IGSO）, in which the individual glowworm （agent） can adaptively search the peaks, and its moving step is variable. The IGSO introduces the tentative moving strategy to enhance the searching ability of the algorithm, and meantime it uses average neighbourhood distance as the reference to adjust agent＇ s moving step. The results of experiment on typical multimodal functions indicate that the IGSO is superior to GSO in multimodal function optimisation with high peaks discovery rate, fast convergence speed and high computational accuracy.