应用精英反向学习的引力搜索算法  

Improved gravitational search algorithm with elite opposition-based learning

作　　者：井福荣[1] 郭肇禄[2] 罗会兰[1] 李康顺[1,3] 

机构地区：[1]江西理工大学信息工程学院,广州510642 [2]江西理工大学理学院,江西赣州341000 [3]华南农业大学信息学院,广州510642

出　　处：《计算机应用研究》2015年第12期3638-3641,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(61462035);江西省教育厅科技项目(GJJ14456)

摘　　要：引力搜索算法是近年提出的一种颇具潜力的全局优化算法,已经成功应用到了各种工程实践中,然而它在求解复杂工程优化问题时容易出现早熟收敛问题。为了在一定程度上避免早熟收敛现象,提出一种应用精英反向学习策略的引力搜索算法(EOGSA)。在演化进程中,对当前种群中的每个个体分别执行精英反向学习策略,生成一个精英反向种群,并将生成的精英反向种群与当前种群同时进行竞争,选择出下一代种群。在一系列经典函数优化测试问题上的对比实验结果表明,EOGSA算法能够提高传统引力搜索算法的性能,在一定程度上避免早熟收敛的缺点。Gravitational search algorithm （GSA） is a newly developed global optimization algorithm,which has been successfully applied in many practical applications. However,it tends to suffer from premature convergence when solving complex practical optimization problems. In order to avoid the premature convergence to some degree, this paper proposed an improved gravitational search algorithm with elite opposition-based learning （ EOGSA）. In the evolutionary process, each individual of EOGSA in the current population was undergone by the elite opposition-based learning strategy to create an elite opposition-based population. Moreover, the created elite opposition-based population competed with the current population to select the individuals for the next generation. The experiments were conducted on a set of classical test functions. The comparison results indicate that the proposed EOGSA can enhance the performance of the traditional GSA, and alleviate the premature convergence to some extent.

关 键 词：全局优化 演化算法 精英反向学习 引力搜索 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]