改进的遗传算法用于氩原子簇的结构优化

The Optimization of the Argon Atom Cluster Structure Using a Modified Genetic Algorithm

目的 :研究一种改进的遗传算法对氩原子簇的结构进行优化 ;方法 :描述了一种采用实数串编码 ,非一致性变异和数值交叉算子改进的遗传算法 ,同时结合了局部搜索技术来进一步优化遗传算法得到的最佳解 ;结论 :将改进的遗传算法用于氩原子簇 (Ar) N (N =2 14)的结构优化后 ,其结果表明MGALM虽算法简单 ,但对较小原子簇体系 (N≤ 14) ,可以得到其已知的全局最优结构。

Purpose: To find minimum energy structures for Argon clusters of two to fifteen atoms using a modified genetic algorithm(MGALM); Method: A modified genetic algorithm (MGALM) which a real coded, non coherence and numerical value crossover operator were described and a local research was brought into to find global minimum energy structure of argon clusters; Result:When this modified genetic algorithm was used to optimize the structure of argon clusters (the argon atom number N ≤14), it indicated that although the MGALM was simple, the optimized lowest energies were identical with the global optimized energies taken from the published literature. By comparing with the parallel genetic algorithm, this model is simple and feasible.