基于小生境改进的遗传算法求解多序列比对问题  

An Improved Genetic Algorithm for Solving Multiple Sequence Alignment Problem Based on Niche Technology

作　　者：张继成[1] 羊秋玲[2] 

机构地区：[1]长江大学工程技术学院,湖北荆州434020 [2]海南大学信息科学技术学院,海南海口570228

出　　处：《常州工学院学报》2016年第4期34-38,共5页Journal of Changzhou Institute of Technology

基　　金：海南省重点研发资助项目(ZDYF2016153);长江大学工程技术学院基金项目资助(2016KY13)

摘　　要：基于多核平台设计了一个求解多序列比对问题的改进遗传算法。该算法采用一致性函数作为个体的适应度函数,引进小生境技术,维持种群进化的多样性,以改善算法的整体搜索能力。考虑遗传算法本身具有较好的并行性,对其各算子针对多核平台进行了并行化设计。通过对BAliBASE中的测试例进行测试,与已有的算法相比取得了更优的结果,证明该算法是有效的。并行化设计使算法在多核平台的运行时间显著缩短,加速效果明显。An improved genetic algorithm based on the multi-core platform was presented for solving the multiple sequence alignment problem. The algorithm took consistency function as an individual fitness function,and introduced niche technology to maintain the diversity of population so as to improve the overall search capability of the algorithm. Considering the good parallelism of genetic algorithm,the operators were paralleled for the multi-core platform. BAli BASE test cases were used to test the algorithm,comparing with the existing algorithm. The better results showed that it was effective. Paralleling algorithm run significantly and reduced the running time on the multi-core platform with obvious acceleration effect.

关 键 词：多序列比对 小生境遗传算法 多核 并行 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]