化工学报 ›› 2021, Vol. 72 ›› Issue (1): 27-41.DOI: 10.11949/0438-1157.20201037

• 综述与专论 • 上一篇    下一篇

遗传算法在催化体系的全局结构优化中的应用

石向成1,2,3(),赵志坚1,2(),巩金龙1,2,3   

  1. 1.天津大学化工学院,绿色合成与转化教育部重点实验室,天津 300072
    2.天津化学化工协同创新中心,天津 300072
    3.天津大学-新加坡国立大学福州联合学院,福建 福州 350207
  • 收稿日期:2020-07-25 修回日期:2020-10-05 出版日期:2021-01-05 发布日期:2021-01-05
  • 通讯作者: 赵志坚
  • 作者简介:石向成(1996—),男,博士研究生,e0444250@u.nus.edu
  • 基金资助:
    国家自然科学基金项目(21761132023)

Application of genetic algorithm in the global structure optimization of catalytic system

SHI Xiangcheng1,2,3(),ZHAO Zhijian1,2(),GONG Jinlong1,2,3   

  1. 1.Key Laboratory for Green Technology, School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China
    2.Collaborative Innovation Center of Chemical Science and Engineering(Tianjin), Tianjin 300072, China
    3.Joint School of National University of Singapore and Tianjin University, Fuzhou 350207, Fujian, China
  • Received:2020-07-25 Revised:2020-10-05 Online:2021-01-05 Published:2021-01-05
  • Contact: ZHAO Zhijian

摘要:

对催化体系进行全局结构优化,搜寻基态结构对预测催化剂结构、分析反应物的吸附特性、研究多相催化反应机理、构建实际反应路径等方面至关重要。遗传算法通过交叉、变异和选择等操作,模拟了自然淘汰进化过程,来搜索势能面上的基态结构。作为一种无偏优化算法,遗传算法的优化过程不依赖于输入结构,具有很强的全局搜索能力。对遗传算法在催化体系的全局结构优化问题中的应用进行了综述,介绍了遗传算法在实空间上进行全局结构优化的基本程序框架以及近年来结合并行计算、机器学习等技术发展的改进框架,并讨论了它们在团簇优化、负载型催化剂的结构优化问题上的相关应用,为遗传算法的进一步改进以及更广泛的应用提供理论指导。

关键词: 遗传算法, 全局优化, 催化, 势能面, 纳米结构, 分子模拟, 机器学习

Abstract:

Genetic algorithm is widely used to search for the global minimum structure that is important for analyzing the catalyst structure, the mechanism of heterogeneous catalytic reaction, and actual reaction pathway. By performing crossover, mutation and selection, genetic algorithm reflects the process of natural selection where the fittest individuals are selected for reproduction to explore the potential energy surface. As an unbiased optimization algorithm, the optimization process of genetic algorithm does not depend on the input structure and has strong global search capabilities. This review summarizes the recent progress of the design and application of genetic algorithm, as a global structure optimizer, in the catalytic system. Starting with introducing the standard genetic algorithm framework for global structure optimization, this review also includes the advanced framework developed by introducing parallel computing and machine learning technique. Finally, some examples about the reported application of genetic algorithm in catalytic structure optimization are presented, such as the optimization of metallic clusters, supported catalysts, etc. This review might provide a significant insight into the further improvement of genetic algorithm and the wider application in catalytic system.

Key words: genetic algorithm, global optimization, catalysis, potential energy surface, nanostructure, molecular simulation, machine learning

中图分类号: