摘要： 针对进化算法收敛效率和搜索能力不能有效兼顾的缺点，提出一种搜索机制与进化机制相分离的方法——搜索空间划分移动策略，并介绍其相关理论。该策略在搜索区域划分理论基础上，加入区域移动机制，使各极值能在不同环境下相互竞争，提高其搜索能力。改进的处理机制无须考虑搜索问题，能让算法迅速收敛。记录器能有效地完成各小区域间的通信，并根据要求控制区域移动。实验结果表明该策略具有较好的准确性和效率。

关键词: 空间划分, 区域移动, 并行遗传算法, 进化规划

Abstract: Considering the contradiction between efficiency of convergency and search capability, a novel approach, search space division and movement strategy, which will separate the search mechanism from evolution mechanism, is proposed. The corresponding theory is put forward based on a deep study of both traditional and improved evolutionary algorithms. The extremums will compete with each other in different circumstances, for using the region movement mechanism which is based on the theory space division strategy. And the search capability of system is improved a lot. The improved region processor without any concern of the search problem in each region is efficient. The record is designed not only to enable the region processor to exchange information effectively, but also control the region movement mechanism in terms of various requests as well. Experimental results show that this strategy has remarkable advantage in accuracy and efficiency.

Key words: space division, region movement, parallel genetic algorithm, evolutionary programming

中图分类号: 

• TP273