计算机工程 ›› 2021, Vol. 47 ›› Issue (1): 275-283.doi: 10.19678/j.issn.1000-3428.0056800

• 开发研究与工程应用 • 上一篇    下一篇

基于改进离散差分进化算法的桁架优化

王烁1, 谷正气1,2, 韩征彤1, 马晓骙1   

  1. 1. 湖南大学 汽车车身先进设计制造国家重点实验室, 长沙 410082;
    2. 湖南文理学院 洞庭湖生态经济区建设与发展协同创新中心, 湖南 常德 415000
  • 收稿日期:2019-12-04 修回日期:2020-01-14 发布日期:2020-01-17
  • 作者简介:王烁(1994-),男,硕士研究生,主研方向为结构优化设计、智能算法;谷正气,教授;韩征彤、马晓骙,博士研究生。
  • 基金项目:
    湖南省重点研究发展计划(2017GK2203);湖南大学汽车车身先进设计制造国家重点实验室自主课题项目(734215002);常德市重大科技攻关项目(CD201701)。

Truss Optimization Based on Improved Discrete Differential Evolution Algorithm

WANG Shuo1, GU Zhengqi1,2, HAN Zhengtong1, MA Xiaokui1   

  1. 1. State Key Laboratory of Advanced Design and Manufacture for Vehicle Body, Hunan University, Changsha 410082, China;
    2. Cooperative Innovation Center for the Construction and Development of Dongting Lake Ecological Economic Zone, Hunan University of Arts and Science, Changde, Hunan 415000, China
  • Received:2019-12-04 Revised:2020-01-14 Published:2020-01-17

摘要: 为提高离散桁架优化问题的计算效率,提出一种改进的离散差分进化算法。基于种群多样性自适应地选择变异策略以平衡探索和收敛能力,根据个体差异度和种群多样性缩减种群规模以减少计算量,在进行结构分析前舍弃较大的实验个体规避无用计算,并引入精英选择技术解决选择阶段目标个体和实验个体数量不等的问题,在此基础上,给出一种将数值之间的距离转化为概率的离散化方法,处理离散变量问题。实验结果表明,与IGA、DE等算法相比,该算法在保证最优解质量的同时,能够大幅减少结构分析次数。

关键词: 离散差分进化算法, 自适应变异策略, 自适应种群规模, 结构分析次数, 桁架优化

Abstract: In order to improve the calculation efficiency of the discrete truss optimization problem,this paper proposes an improved discrete differential evolution algorithm.The mutation strategies is adaptively selected based on population diversity to balance exploration and convergence capabilities,and the population size is adaptively reduced based on individual differences and population diversity to reduce calculations.Heavier trial individuals are discarded before structural analysis to avoid useless calculations,and the elite selection technique is introduced to solve the problem of unequal numbers of target individuals and trial individuals in the selection phase.On this basis,a discretization method that converts the distance between the values into a probability is proposed to solve the problem of discrete variables. Experimental results show that compared with algorithms such as IGA and DE,this algorithm can greatly reduce the structural analysis times while ensuring the optimal solution quality.

Key words: discrete differential evolution algorithm, adaptive mutation strategy, adaptive population size, number of structural analysis, truss optimization

中图分类号: