基于排序法和分解的高维多目标演化算法研究

项目名称： 基于排序法和分解的高维多目标演化算法研究

项目编号： No.61502290

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 自动化技术、计算机技术

项目作者： 代才

作者单位： 陕西师范大学

项目金额： 21万元

中文摘要： 科学和工程领域中存在着许多的高维多目标优化问题（目标个数大于4）。本项目对高维多目标优化问题的求解方法展开了深入研究，探索了新的研究方法，克服了当前研究的局限和缺点。本项目的主要研究内容包括两个方面：一方面，从本质上分析了基于Pareto最优概念的排序法的缺陷，提出了高维多目标问题的排序新方法；另一方面，分析了现有的演化算法求解高维多目标优化问题的缺陷,提出了出了一种基于分解和排序方法的演化模型，设计了基于自动学习机的交叉算子来提高算法的搜索效率，设计了新的基于分解的更新策略来更好地维持解的多样性，最后将设计的交叉算子、更新策略与所提出的演化模型结合起来用于求解高维多目标优化问题，开发出具有很强通用性和鲁棒性的高维多目标演化算法。本项目的研究成果对高维多目标优化问题的求解方法起到了积极的推动作用。同时，因为工程领域中存在着许多的高维多目标优化问题，因此这个项目也具有很大的实际意义。

中文关键词： 高维优化；多目标优化；排序法

英文摘要： There are many many-objective optimization problems (the number of objectives is more than four) in science and engineering applications. This project is to further study the solving methods of many-objective optimization problems, explore the new methods of research, overcome the limitation and disadvantages of current research. The main research content of this project includes two aspects: on the one hand , through the analysis of the nature of drawback of sorting method based on Pareto dominance, propose new methods of sorting candidate solutions so as to overcome the drawback that Pareto optimal solution based sorting method will result in too many Pareto optimal solutions; on the other hand, analysis of the drawbacks of the existing many-objective evolutionary algorithms, propose a evolving model based sorting and decomposition, design new crossover operator based on learning automata to improve the search performance of algorithms, design new update strategies based on decomposition to well maintain the diversity of solutions, and base on above methods, some many-objective evolutionary algorithms which have good versatility and robustness are designed for many-objective optimization problems are designed. This project will promote the development and the practical application of many-objective evolutionary algorithms, and has a great important meaning in the science and engineering application.

英文关键词： many-objective optimization;multi-objective optimization;ranking method