计算机科学与探索 ›› 2014, Vol. 8 ›› Issue (3): 352-358.DOI: 10.3778/j.issn.1673-9418.1310016

• 人工智能与模式识别 • 上一篇    下一篇

混沌萤火虫优化算法的研究及应用

郁书好1,2+,苏守宝2   

  1. 1. 合肥工业大学 计算机网络系统研究所,合肥 230009
    2. 皖西学院 信息工程学院,安徽 六安 237012
  • 出版日期:2014-03-01 发布日期:2014-03-05

Research and Application of Chaotic Glowworm Swarm Optimization Algorithm

YU Shuhao1,2+, SU Shoubao2   

  1. 1. Institute of Computer Network Systems, Hefei University of Technology, Hefei 230009, China
    2. School of Information Engineering, West Anhui University, Lu’an, Anhui 237012, China
  • Online:2014-03-01 Published:2014-03-05

摘要: 针对基本萤火虫群优化算法的早熟收敛,易陷入局部最优值,求解精度不高等问题,提出了一种基于切比雪夫映射的混沌萤火虫优化算法。利用混沌系统的随机性和遍历性初始化萤火虫群,获得了质量较高且分布较均匀的初始解;同时对部分适应值低的个体进行了混沌优化,以提高种群的多样性。对4个标准测试函数进行了仿真实验,结果表明该算法的求解精度、全局搜索能力优于基本萤火虫优化算法。将改进算法应用于车辆路径问题的求解中,结果表明了改进算法的有效性。

关键词: 萤火虫优化(GSO), 早熟收敛, 混沌, 车辆路径问题(VRP), 切比雪夫映射

Abstract: To overcome the disadvantages of premature convergence, local optimum and low precision in basic glowworm swarm optimization (GSO) algorithm, this paper proposes a chaotic glowworm swarm optimization (CGSO) algorithm based on Chebyshev map. CGSO applies the features of chaotic randomness and ergodicity to initial the glowworm population. Therefore, it can achieve high quality and uniformly distributed initial solutions. Meanwhile, in order to increase the diversity of population, the proposed algorithm disturbs the partial individuals with low fitness value by Chebyshev map. The experiments on four standard test functions show that CGSO outperforms the basic GSO in precision and global searching ability. Finally, the improved algorithm is applied to vehicle routing problem (VRP), the results show that the algorithm is effective.

Key words: glowworm swarm optimization (GSO), premature convergence, chaos, vehicle routing problem (VRP), Chebyshev map