An Improved Genetic Algorithm for the Stable Structures of(C60)N Clusters
Genetic algorithm was employed to optimize thestructures of(C60)N molecular clusters with the lowestenergy.Aiming at an effective solution of stablestructures,some improvements are made to traditionalgenetic algorithm.Firstly,gene is coded by the mixedmethod(real number and integral number)forinitialized population quality.Secondly,a newselection mechanism based on roulette wheel andhamming distance is introduced Thirdly,in order toenhance the chromosome diversity,a new self-adaptivecrossover method is developed which combines l-pointcrossover with uniform crossover.Finally,for the sakeof improving the global searching capacity andavoiding the premature convergence,a feedbackmutation based on dynamic encoding is put forward.The experiment results show that the improved geneticalgorithm is good at rapidity and convergence as wellas can search for the stable structure when N variesfrom 3 to 25.
Guifang Shao Yuhua Wen Yaohua Chen
Department of Pattern Recognition and Intelligent System,Xiamen University,Xiamen,China Department of Physics,Xiamen University,Xiamen,China
国际会议
厦门
英文
476-480
2008-11-17（万方平台首次上网日期，不代表论文的发表时间）