摘要
针对传统遗传算法在复杂高维优化问题中适应度计算代价较高的问题,提出一种基于流形学习与多元线性回归的改进遗传算法Gamma.Gamma算法通过流形学习对种群数据进行降维,并结合AP聚类(affinity propagation clustering)与多元线性回归模型,减少适应度函数的计算次数,提高算法优化效率.实验结果表明,Gamma算法在桁架穹顶结构优化等复杂工程及多个经典Benchmark函数上,均以较少的适应度调用次数达到了与传统方法相近的优化效果,在处理高维优化问题上应用前景良好,能有效提高计算效率,降低时间成本.
Aiming at the problem of the high computational cost of fitness in traditional genetic algorithms for complex high-dimensional optimization problems,we proposed an improved genetic algorithm Gamma based on manifold learning and multiple linear regression.The Gamma algorithm reduced the dimensionality of the population data through manifold learning,and combined AP clustering with a multiple linear regression model to reduce the calculation times of fitness function and improve algorithm optimization efficiency.Experimental results show that the Gamma algorithm achieves optimization results similar to traditional methods with fewer fitness calls in complex engineering such as the optimization of truss dome structures and multiple classic Benchmark functions.It has a promising application prospect in handling with complex high-dimensional optimization problems,effectively enhancing computational efficiency and reducing time costs.
作者
张蕾
仲洋
曹梦萱
卢婧
韩霄松
ZHANG Lei;ZHONG Yang;CAO Mengxuan;LU Jing;HAN Xiaosong(Division of Development and Strategic Planning,Jilin University,Changchun 130012,China;Northeast Electric Power Design Institute Co.,Ltd.,China Electric Power Engineering Consulting Group,Changchun 130021,China;College of Software,Jilin University,Changchun 130012,China;Gradute School,Jilin University,Changchun 130012,China)
出处
《吉林大学学报(理学版)》
北大核心
2025年第5期1387-1396,共10页
Journal of Jilin University:Science Edition
基金
国家自然科学基金(批准号:62372494,62372209)
吉林省科技发展计划重点研发项目(批准号:20220201145GX)
吉林省高教学会科研项目(批准号:JGJX2023C9)
吉林大学研究生教育教学改革项目(批准号:2023JGY026).
关键词
遗传算法
流形学习
代理模型
复杂优化问题
genetic algorithm
manifold learning
surrogate model
complex optimization problem
