基于并行免疫遗传算法基因表达数据动态模糊聚类.pdfVIP

47 1 ( ) Vo.l 47 No. 1 2009 1 Journal of Jilin Un ivers ity ( Science Edition) Jan 2009 , , , , , ( , 1 0012) : 提出一种改进的并行免疫遗传算法, 通过在群体规模上引入岛的概念, 实现了 变 的群体规模; 通过在适应度函数内引入免疫算子, 避免了算法过早收敛. 因此, 解决了寻优 算法中局部收敛的困扰, 提高了获得全局最优解的几率. 把此算法应用于斯坦福大学酵母细 胞周期表达数据库的数据进行共表达聚类, 并将实验结果与 Spellman按照功能基因组学进行 聚类所得结果进行了对比, 证明了所给算法在功能基因组学聚类上的有效性. : 动态模糊聚类; 并行免疫遗传算法; 免疫算子; 岛屿 : TP18 : A : 1671-5489( 2009) 01-006 -06 Dynam ic Fuzzy Clustering for AnalyzingM icroarray Gene Expression Data Based on a Parallel Immune Genetic A lgorithm ZHENG M ing, L IU Gu-i xia, ZHOU Chun-guang, WANG H an, ZHENG X iao-hong, L IYan-w en (College of Compu ter S cience and Technology, J ilin University, Changchun 1 0012, Ch ina) Ab stract: W e proposed an mi proved parallel mi mune genetic algorithm. The introduced island concept on the popu lation size can make the sizes of popu lations variable; The mi mune operator introduced in the fitness function can avoid the algorithm prem ature convergence. H ence, th ismethod can prevent local convergence in the optmi al algorithm in a great extent, and make the probability of approach ing the global convergence bigger. The algorithm w as used via the Yeast Saccharomyces cerevisiae cell cycle gene expression profile from SGD to co-express clusters. Compared to the expermi ent analyzed by Spellman for functional genom ics, the efficiency of th is algorithm in the functional genom ics can be proved. Key wo rds: dynam ic fuzzy clustering; parallel mi mune genetic algorithm; mi mune operator; island , [ 1] , . , , . , , , , , , , . : : 2008-07-22. : ( 198 ~ ), , , , , E-m ail: zheng _m th@ 16 . com. : ( 1947 ~ ), , , , , , E-m ail: cgzhou@ jlu. edu. cn. : (: 6