采用动态知识因子的基于知识共享的优化算法  

Gaining-sharing Knowledge Based on Algorithm Using Dynamic Knowledge-factor

作　　者：袁磊 王勇[1,2] 赵艳玲 YUAN Lei;WANG Yong;ZHAO Yanling(College of Artificial Intelligence,Guangxi Minzu University,Nanning 530006,China;Guangxi Key Laboratory of Hybrid Computation&IC Design Analysis,Nanning 530006,China)

机构地区：[1]广西民族大学人工智能学院,广西南宁530006 [2]广西混杂计算与集成电路设计分析重点实验室,广西南宁530006

出　　处：《广西民族大学学报（自然科学版）》2022年第4期58-67,共10页Journal of Guangxi Minzu University ：Natural Science Edition

基　　金：国家自然科学基金项目(61662005);广西自然科学基金项目(2021GXNSFAA220068)。

摘　　要：针对基于知识共享的优化算法(GSK)之不足,提出一种采用动态知识因子的基于知识共享的优化算法(DKGSK)。首先,个体通过自适应权重来调节其搜索,在算法前期以较大步长进行全局探索,增强了算法全局探索能力;在算法后期以较小步长进行局部搜索,提升了算法局部开发能力。其次,个体使用动态知识因子来调控其搜索步长,使搜索步长更具灵活性和随机性,提升了个体的局部搜索能力,从而增强了算法的局部开发能力。再次,个体借助列维飞行增强其跳出局部最优的能力,从而增强了算法规避陷入局部最优的能力。通过12个基准测试函数的数值实验与仿真,结果表明DKGSK的全局收敛速度和优化精度方面均得到了明显的改善,规避陷入局部最优的能力得到了增强。Aiming at the shortcomings of normal gaining-sharing knowledge based algorithm(GSK),this paper proposes a new gaining-sharing knowledge based algorithm using dynamic knowledge-factor(DKGSK).In the algorithm,individuals adjust their search by adaptive weight first,and carry out global exploration with a large step in the early stage of the algorithm,which enhances the global exploration ability of the algorithm;In the later stage of the algorithm,individuals carry out local search with a small step length,which improves the local exploitation ability of the algorithm.Secondly,individuals use dynamic knowledge factors to regulate their search step size,which makes the search step size more flexible and random,and improves the local search ability of individuals,thus enhancing the local development ability of the algorithm.Secondly,individuals use dynamic knowledge factors to regulate their search step size,which makes their search step size more flexible and random,and then improves the ability of individual fine search,thus enhancing the local exploitation ability of the algorithm.Thirdly,individuals enhance their ability to jump out of the local optimum by using Levi's flight,which enhances the ability of the algorithm to avoid falling into the local optimum.Through the numerical experiments and simulations of 12 benchmark functions,the results show that the global convergence rate and optimization accuracy of DKGSK have been significantly improved,and the ability to avoid falling into local optimization has been enhanced.

关 键 词：基于知识共享的优化算法 自适应权重 动态知识因子 智能优化 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]