摘要
本文算法体现"分而治之"的思想,首先采用动态K均值聚类快速邻域分解,其次应用蚁群算法同时对分区并行优化计算,最后基于分区重心进行邻域全局连接,得到大规模TSP问题的满意解.
For a large number of TSP of combinatorial\|explode NP\|Hard problems, Any only one algorithm will be face challenge in optimization characteristic and CPU run\|time. The paper thinking is 'divide and rule'. First, we fast nearest neighbor classify used to \$k\$\|means cluster. Second, we are parallel computing used to ant colonies optimization for every group. Finally, we get good results used to globally optimizing use of between\|group linkage and between\|group Centro baric distance.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2003年第9期105-110,共6页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(60174021)
天津自然科学基金重点项目(013800711)
河南科技攻关项目
