基于适应度地形分析的进化计算研究综述  

A survey of evolutionary computing based on fitness landscape analysis

作　　者：李亚欣 梁静[1] 岳彩通 李珂[1] LI Yaxin;LIANG Jing;YUE Caitong;LI Ke(School of Electrical Engineering,Zhengzhou University,Zhengzhou 450001,Henan,China)

机构地区：[1]郑州大学电气工程学院,河南郑州450001

出　　处：《陕西师范大学学报（自然科学版）》2021年第5期39-53,共15页Journal of Shaanxi Normal University：Natural Science Edition

基　　金：国家自然科学基金(61922072,61876169)。

摘　　要：适应度地形分析关注优化问题本身,用于提取与适应度地形特性有关的描述性或数值度量,在评估算法相对性能、对优化问题分类、指导算法设计和算法推荐等方面具有重要的研究意义和应用价值。本文介绍了进化计算的特点、分类、理论基础研究和适应度地形的定义与基本特性,总结了进化计算在适应度地形方面的设计研究,包括地形特征的设计与改进,利用地形特征对优化问题分类、指导算法设计、评估算法相对性能、指导算法选择等,并针对进化计算在适应度地形分析中存在的共性关键问题进行了讨论与展望。Fitness landscape analysis focuses on the optimization problem itself and is used to extract descriptive or numerical measures related to fitness landscape characteristics.It has important research significance and application value in evaluating the relative performance of the algorithm,classifying optimization problems,guiding algorithm design and algorithm recommendation.The characteristics,classification,theoretical basic research of evolutionary computing and the definition and basic characteristics of fitness landscape are introduced,and the design and research of evolutionary computing in fitness landscape are summarized,including the design and improvement of landscape features,the classification of optimization problems by using landscape features,guiding algorithm design,evaluating the relative performance of algorithms,guiding algorithm selection,etc.The common key problems of evolutionary computing in fitness landscape analysis are discussed and prospected.

关 键 词：进化计算 适应度地形 优化 特征提取 

分 类 号：O156.4[理学—数学]