动态优化问题中的演化膜算法  

Evolutionary membrane algorithms in dynamic environments

作　　者：刘闯[1] 韩敏[1] 王心哲[1] 

机构地区：[1]大连理工大学电子信息与电气工程学部,辽宁大连116024

出　　处：《哈尔滨工程大学学报》2013年第11期1397-1401,共5页Journal of Harbin Engineering University

基　　金：国家自然科学基金资助项目(61074096)

摘　　要：针对现有动态优化算法易陷入局部极值和多样性差等问题,提出了一种动态演化膜算法.依据膜计算理论,所提算法引入膜结构、多重集和反应规则来求解动态优化问题.为了增强在动态环境下的适应能力,所提算法使用了网格策略对搜索空间进行划分,同时设计了4个反应规则来保持算法在动态寻优过程中解的多样性.仿真实验采用标准移动峰测试问题验证算法的求解性能,并分别与3种动态优化算法的求解结果进行比较.仿真结果表明:所提算法提高了搜索过程中解的多样性,且求得的近似最优解更接近于问题的全局最优解,说明所提算法求解动态优化问题是可行的和有效的.Considering that the existing dynamic optimization algorithms can easily poor diversity, a novel dynamic evolutionary membrane algorithm is proposed. T~ three elements of membrane computing, including membrane structure, muhiset and reaction rules, to solve dynamic optimization problems. To enhance the adaptive ability of the proposed algorithm under dynamic environments, the al- gorithm employs the grid to divide the search space. Furthermore, the four kinds of reaction rules are introduced to maintain the diversity of solutions found by the algorithm during a dynamic optimization process. In simulation experi- ments, the standard moving peaks benchmark was used to validate the solving performance of the algorithm. Moreo- ver, the performance of the proposed algorithm was compared with three state-of-the-art dynamic optimization algo- rithms. The simulation results indicate that the proposed algorithm improves the diversity of the candidate solutions, and the approximate optimal solution found by the algorithm is closer to the global optimal solution. Therefore, the proposed algorithm is feasible and effective in solving dynamic optimization problems.

关 键 词：动态演化膜算法 动态优化 膜计算 演化膜算法 多样性 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]