元启发式荧光粒子群优化算法与理论分析研究

项目名称： 元启发式荧光粒子群优化算法与理论分析研究

项目编号： No.61463007

项目类型： 地区科学基金项目

立项/批准年度： 2015

项目学科： 其他

项目作者： 周永权

作者单位： 广西民族大学

项目金额： 47万元

中文摘要： 元启发式优化方法作为智能科学的热点研究领域，受到国内外研究者的广泛关注。本项目受自然界荧科族昆虫通过发光求偶或觅食行为的启发，基于荧科族的生物学原理，以荧科族中的萤火虫群算法为主，以粒子群算法为辅，嵌入随机自适应步长、全局优化吸引子、强搜索、弱搜索、自学习能力等演化算子来缩短个体之间信息交流的成本，利用集成演化策略，设计出一种新型元启发式萤光粒子群全局优化算法，理论上证明该算法以概率1收敛于全局最优。最后，采用经典的测试函数和工程实例进行仿真实验，与代表性元启发式优化算法ACO、PSO、GSO等算法的性能进行数值实验对比分析，期望通过实验结果表明所提出的算法能达到节约群体规模，减少迭代次数，鲁棒性强，捕峰数目全，收敛精度高，在高维空间多目标函数优化领域有着极其重要的应用。本项目所采用的研究方法，对于提高和改进其它元启发式算法的整体性能和效率提供了一新途径和方法。

中文关键词： 元启发式；荧光粒子群优化；全局收敛；多模态优化；智能计算

英文摘要： Meta heuristic optimization method as a hot research field of intelligent science，it's wide attention by the researchers both at home and abroad Fluorescent particle swarm optimization is affected by the nature lampyridae family of insects by luminescence mating or inspired by the foraging behavior of emerging evolution method, this method is given priority to with fluorspar family of glowworm swarm algorithm, particle swarm optimization (pso) algorithm as the auxiliary pole, embedded stochastic adaptive step length, weak global attractor, strong search and search, self-learning ability evolution strategy to shorten the cost of information exchange between individuals, using integrated evolution technology, designed a new type of fluorescent particle swarm optimization algorithm, theoretically proved that the algorithm to converge to global optimal probability 1.Finally, the classic functions and an engineering example is adopted to improve the testing, algorithm performance contrast experiment, the experimental results show that the proposed algorithm can save the group scale, reduce the number of iterations, strong robustness, catching peak number,and high precision and convergence in multi-modal function optimization is very important in the field of application. This project adopted the research methods, to enhance and improve the group of the overall performance and efficiency of intelligent optimization algorithm provides a new way and method.

英文关键词： Meta-heuristic;bioluminescent particle swarm optimization;global convergence;multi-modal function optimization;intelligent computing