材料的微观结构是深入理解其宏观物理和化学性质的最基本信息，是开展新型功能材料设计的基础。因此发展只根据材料化学组分高效的理论结构预测方法至关重要。申请人所在课题组前期基于晶体对称性的分类检索思想，引入基于粒子群优化算法的群智理论，提出并发展了CALYPSO材料结构预测方法和软件。该方法在预测中小尺度材料（胞内原子数 < 100）结构上取得较好效果。但由于第一性原理总能计算方法比较耗时，导致CALYPSO结构预测方法在预测大体系结构时还存在一些困难。本项目拟基于前期发展的CALYPSO结构预测方法，引入自适应经验势模型和结构-经验势数据库技术，发展高效的自适应CALYPSO结构预测新方法, 为预测复杂大尺度材料的结构提供技术支持。同时编写具有自主知识产权的自适应CALYPSO结构预测程序，并集成于CALYPSO软件包。后续应用该方法预测Fe和碱土金属Mg的合金结构，以期在应用中完善该方法。

Microstructure of the material is the basic information for understanding of macroscopic physical and chemical properties and design of functional materials. The development of effective method on structure prediction with the only known information of chemical composition is crucial. Our team has applied particle swarm optimization into the prediction of structures and developed the CALYPSO structural prediction method and software. This method combined with first-principles total energy calculations has successfully applied to predict the structure of medium systems (the number of atoms per cell < 100). However, the first-principles calculations are computationally so expensive that there are some difficulties in predicting structures for large systems. The current project is trying to develop the adaptive CALYPSO structure prediction method for large system based on structure-empirical potentials database and code the developed method into the CALYPSO software. Later on, the developed method will be applied to the structure prediction of the alloy of Fe and Mg.