摘要
针对萤火虫算法FA对于高维复杂问题,收敛速度慢、求解精度低,优化效果不理想等缺点,提出一种基于全局信息共享的自适应FA算法。分别从三个方面对FA算法进行了改进:通过引入群体距离,改进γ值的调节方式,提升算法的自适应调节能力;通过增加过程搜索信息,加强算法的精细化调节能力;通过引入基于全局平均位置信息的量子空间下的δ势阱模式,增强算法的全局搜索能力。最后对几种典型函数的测试结果表明,改进算法在收敛速度与收敛精度上,较其它算法有明显提高。
The firefly algorithm (FA) for high dimensional problems has some disadvantages, inclu ding slow convergence speed, low solving precision and unsatisfactory optimization effect. To overcome these disadvantages, we propose a novel adaptive FA algorithm based on global information sharing. Firstly, an adaptive control for gamma value is designed by the swarm distance. Secondly, the search process information of the firefly algorithm is updated to enhance its adjustment capacity of refinement. Thirdly, the global searching ability is improved by introducing the Delta potential well of the vector subspace based on the global mean location information. Simulation results show that the proposal has better convergence speed and precision than the basic FA and the PSO.
出处
《计算机工程与科学》
CSCD
北大核心
2016年第6期1164-1170,共7页
Computer Engineering & Science
基金
国家自然科学基金(61170119)
安徽省自然科学研究项目(KJ2016A514)
关键词
萤火虫算法
自适应
全局信息共享
过程信息更新
firefly algorithm
adaptive
global information sharing
process information update