面向动态约束优化问题的进化算法:设计、分析与应用

项目名称： 面向动态约束优化问题的进化算法:设计、分析与应用

项目编号： No.61273314

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 王勇

作者单位： 中南大学

项目金额： 81万元

中文摘要： 动态约束优化问题在实际应用中十分常见，但在进化计算领域，目前此类问题很少受到研究人员的关注。本项目拟对面向动态约束优化问题的进化算法展开全面研究，以突破当前研究的局限和不足。本项目将围绕设计、分析与应用这一主线展开，主要包含以下五个方面：1）目前针对动态约束优化问题的测试函数还十分匮乏，拟设计能够体现实际动态约束优化问题特征的标准测试函数集；2）设计能够有效求解动态约束优化问题的进化算法；3）为了客观、全面地比较各种求解动态约束优化问题的进化算法，将设计合理的性能比较准则；4）分析进化算法求解动态约束优化问题的收敛性和收敛速度，完善其理论基础；5）将所设计的进化算法应用于求解电信系统和软件工程中的动态约束优化问题。本项目的研究成果将拓展进化计算的研究领域，进一步促进进化计算研究的发展。由于实际工程应用中存在着大量的动态约束优化问题，所以本项目还具有十分重要的实际意义。

中文关键词： 动态约束优化；约束处理技术；多目标优化技术；进化算法；时间复杂度

英文摘要： Dynamic constrained optimization problems (DCOPs) are frequently encountered in the real-world applications, however, at present very few attempts have been made to investigate this kind of problems in the field of evolutionary computation. This project aims at studying on evolutionary algorithms (EAs) for DCOPs extensively, in order to overcome the limitations and deficiencies of the current research. This project will mainly focus on the design, analysis, and applications of EAs for DCOPs, and include the following five aspects. Firstly, since there are few test instances for DCOPs, we will design a set of standardized dynamic constrained optimization test instances, which can resemble real-world problems. Secondly, we will design some novel and effective EAs to solve DCOPs. Thirdly, some suitable performance indicators will be designed, in order to guarantee a fair and all-around comparison between different EAs when solving DCOPs. Fourthly, we will analyze the convergence property and convergence speed of EAs for DCOPs, with the aim of establishing the theoretical foundation of this field. Finally, applying the developed EAs to deal with the DCOPs in telecommunication systems and software engineering is also one part of this project. The fruits of this project will enlarge the field of evolutionary computati

英文关键词： dynamic constrained optimization；constraint-handling technique；multiobjective optimization technique；evolutionary algorithms；computational time complexity