基于数学规划的解高维多目标优化问题的异步并行进化算法

项目名称： 基于数学规划的解高维多目标优化问题的异步并行进化算法

项目编号： No.61070007

项目类型： 面上项目

立项/批准年度： 2011

项目学科： 建筑科学

项目作者： 邹秀芬

作者单位： 武汉大学

项目金额： 11万元

中文摘要： 高维多目标优化问题是当前大规模科学与工程计算中面临的一个挑战性问题。本项目旨在采用传统的非线性规划方法与现代的进化计算技术相结合的方法，探索求解这类问题的高效算法及其理论基础。具体研究内容包括：通过理论分析与数值模拟相结合的模式，研究在进化计算的框架下如何引入梯度法等传统的数学规划方法从而产生新的个体评价、选择等机制，进而得到高性能的求解算法；探讨算法的收敛性、时间复杂性等，为高维多目标优化问题的有效处理提供新思路和新方法。

中文关键词： 高维多目标优化问题；进化算法；收敛性；时间复杂性

英文摘要： Multi-objective optimization problems with a large number of objectives (denoted as "Many-objective optimization problems", generally the number of objectives is equal to or greater than five) are the challenge problems in large-scale scientific and engineering computing at present. The aim of this project is to search the high-performance algorithms and theoretic foundations for solving these kinds of problems by combining the conventional nonlinear programming methods and modern evolutionary algorithms. The main contents include: study how to introduce the gradient approaches to present new mechanisms for evaluating and selecting individuals in order to obtain the high-performance algorithms under the framework of evolutionary computation, and analyze the convergence and time complexity of algorithms. It can provide the new ideas and the new methods for effective handling Many-objective optimization problems.

英文关键词： Many-objective optimization problems; evolutionary algorithms; convergence; time complexity