中文摘要：

本文应用灵敏度分析方法从BaTiO3壳模型众多参数中找到对结构和性能影响最为明显的主要参数,将次要参数常数化,降低优化问题的维数;然后应用遗传算法对主要参数进行优化.优化结果表明,势参数可以较准确地模拟结构、力学性能以及相变过程.

英文摘要：

Shell model potential is widely applied to molecular dynamics simulations of ionic crystal,and the parameters of shell model potential function are crucial to the simulation veracity. The parameters of shell model of multi-oxide are numerous,and the optimization is a challenging task. In this paper,the sensitivity analysis is applied to find the key parameters which affect structure and properties mostly in all the shell model potential parameters of BaTiO3. Whereafter, the genetic algorithm is applied to optimize the key parameters,and the insignificant parameters are kept constant in optimization. The results show that the structures,physical properties and phase transition of BaTiO3 simulated by the optimized potential agree well with the experimental data.