摘要
蚁群算法是优化领域中新出现的一种仿生进化算法,基于蚁群算法的聚类算法已经在当前的数据挖掘研究中得到应用。文中针对早期蚁群聚类算法的缺点,提出动态调整的蚁群聚类算法,通过加入运动速度不同的蚁群、半径自适应调整、短期记忆、强行放下等策略,来指导蚁群的移动行为,降低蚁群移动的随意性,减少了蚂蚁的搜索时间,提高聚类性能。仿真实验表明:改进算法能有效地提高算法效率且取得较好的聚类结果。
Ant colony algorithm is a novel category of bionic algorithm, the ant - based clustering algorithm has currently applications in the data mining community. Based the disadvantage of the classical algorithm, presents a dynamic adjustive ant - clustering algorithm, including different speed ant colony, adaptive radius adjustment, short memory, force drop action which guide the ant＇s movement. The algorithm can lower the randomness of ant＇s moving and reduce the time of ant＇s searching to improve the performance. Experiment shows that the new algorithm effectively advances the efficiency of algorithm and the result of clustering.
出处
《计算机技术与发展》
2009年第2期145-147,共3页
Computer Technology and Development
基金
安徽省自然科学基金项目(KJ2008B092)
