摘要
为多目标优化问题设计有效算法相当困难,且得到的算法缺乏复用性。本文提出一个求解框架,对问题特征分类,特征类型对应框架中的模块,具体特征由模块中的策略处理,策略成为可复用的。模块和策略的分析与总结使为问题选择合适策略组成算法变得简单。
Designing efficient algorithms for multi-objective optimization problems is always hard and the obtained algorithms are not reusable. A framework is thus proposed,through classifying problem features,the category corresponds to the module of the frame-work while the concrete feature is processed by the strategy in the module,and strategies become reusable. The analysis and summarization of modules and strategies makes it easier to choose proper strategies to compose the algorithm for the problem.
出处
《微计算机信息》
2009年第36期164-165,168,共3页
Control & Automation
关键词
多目标优化问题
框架
模块
策略
粒子群优化
Multi-objective Optimization Problem
Framework
Module
Strategy
Particle Swarm Optimization