解决非静态优化问题的MEAP算法  

The MEAP Algorithm to Solve the Non-Static Optimization Problems

作　　者：吴漫川[1] 李元香[1] 郑波尽[1] 

机构地区：[1]武汉大学软件工程国家重点实验室,湖北武汉430072

出　　处：《计算机工程与科学》2005年第8期73-75,80,共4页Computer Engineering & Science

基　　金：教育部博士点基金资助项目(20030486049)

摘　　要：演化算法已在传统的静态优化领域显示了惊人的能力,但非静态优化问题更接近于我们的生活。如何将演化算法应用于非静态优化是当前的一个研究热点。本文讨论了几种算法,并提出了一种基于传统演化算法的新算法(MEAP)。这种新算法可以及时得知环境的改变并进行预处理,从而让种群有更多的机会产生优解。测试结果表明,该算法有优良的性能。Evolutionary algorithms have proved to be powerful in the static optimization field. But how to apply it to the non-static optimization problems that are more common in the real world is still open. This paper discusses some algorithms that can get acceptable results, and proposes a new algorithm （MEAP） that is based on classic EA. The new algorithm can learn about environmental changes instantaneously and offer preprocessing that can lead the population out of swamp and migrate to more hopeful areas if a change occurs, The experimental results show that this algorithm has better performance.

关 键 词：演化算法 非静态优化 MEAP 函数优化 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]