一种基于膜计算的改进人工鱼群算法  

An Improved Artificial Fish Swarm Algorithm Based on Membrane Computing

作　　者：廖孝勇[1] 李尚键 孙棣华[1] 何伟[1] 余楚中[1] 

机构地区：[1]重庆大学自动化学院,重庆400044

出　　处：《小型微型计算机系统》2014年第5期1142-1146,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金计划项目(61203135)资助;教育部博士点基金项目(20120191110047)资助;重庆市自然科学基金重点项目(2012JJB40002)资助;重庆市科委工程中心研究计项目(2011pt-gc30005)资助

摘　　要：针对人工鱼群算法运算速度慢,收敛精度低,易陷入局部最优等问题,基于膜计算思想,通过引入差异因子,提出一种改进的自适应人工鱼群算法.算法采用膜计算的层次结构和交流规则,以保持鱼群的多样性,并克服其易陷入局部最优的缺陷.此外通过简化觅食行为,并根据种群中不同个体与种群规模的比例定义差异因子,对算法的视距、步长、拥挤度因子、尝试次数等进行自适应调整,改善算法的收敛精度和运算速度.实验证明,本文所提算法能够有效提高计算效率和收敛精度.In order to overcome the shortcomings that the artificial fish swarm algorithm （ AFSA ） traps into local optima easily and has slow computational speed and low convergence accuracy, an improved adaptive AFSA algorithm based on differential factor and mem- brane computing （ MC ） is proposed. The algorithm keeps the diversity of fish swarm and overcomes the problem of trapping into lo- cal optima easily by using the framework of MC and rules. In addition, the algorithm simplifies prey behavior and enhances the per- formance of speed and accuracy by using differential factor to adjust visual, step, delta and attempts. Experimental results show that this algorithm can improve the calculation efficiency and accuracy effectively.

关 键 词：人工鱼群算法 膜计算 差异因子 自适应 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]