摘要
为了解决萤火虫算法在全局寻优中收敛速度慢、易陷入局部最优的缺陷,提出了一种改进的文化萤火虫算法,利用信度空间中的规范知识引导搜索区域,自适应调整算法的搜索范围,提高算法的收敛速度和勘探能力;采用随机选择的方式进行扰动操作,增加种群的多样性,平衡算法的局部搜索和全局搜索;当算法陷入局部最优时,自适应的对种群空间进行变异更新,从而有效发挥文化算法的"双演化双促进"机制。通过对6个标准测试函数进行实验测试,结果表明改进算法在收敛速度和求解质量上均取得较好的效果。
In order to resolve such deficiencies of Firefly Algorithm, such as slow convergence and easy to fall into local optimum during global optimization search, this paper proposed a new improved culture firefly algorithm. This algorithm made use of normative knowledge in the belief space to guide the search space, adaptively adjusted the hunting zone to improve its convergence rate and exploration ability. It conducted disturbance operation through random selection so as to increase the diversity of the population and balance local search and global search of the algorithm. When the algorithm falls into local optimum, the algorithm can adaptively conduct variation and update the population space to effectively bring the ＂dual evolution and dual promotion＂ mechanism of cultural algorithm （CA） into play. The results of experimental tests on six standard test functions show that this algorithm has achieved better improvement in convergence rate and solution quality.
出处
《计算机仿真》
CSCD
北大核心
2014年第6期261-265,286,共6页
Computer Simulation
基金
国家自然科学基金(6097004)
山东省自然科学基金(ZR2012HW052)
关键词
萤火虫算法
文化算法
种群多样性
全局优化
Firefly algorithm
Cultural algorithm
Diversity of population
Global optimization
