遗传算法研究原子团簇的玻璃转变及其最优构型

Study on Glass Transition of Atomic Cluster and Its Optimal Configuration Used Genetic Algorithm

运用遗传算法模拟方法,对SiO2原子团簇和Binary Lennard-Jonesmixture(BMLJ)液体原子团簇进行模拟研究.通过对两者的势能变化图像分析,发现SiO2原子团簇和LJ液体原子团簇随着遗传代数的增加,势能都逐渐降低到最小值,从而寻找最优构型.但是LJ液体原子团簇的势能下降趋势明显比SiO2原子团簇势能下降趋势要快.同时还对比分析了两种原子团簇所对应的最优构型.

The atomic cluster of SiO2 and Binary Lennard- Jones mixture （BMLJ） fluid have been studied by genetic algorithm simulation method. Based on the potential energy curve, as time goes on, potential energy are gradually reduced to the minimum, so as to find the optimal configuration . But LJ fluid potential energy decline more quickly than the atomic cluster of SiO2. At the same time, optimal configuration of atomic cluster is also analyzed.