摘要
蚁群算法是一种模拟进化算法,具有很强的全局搜索能力。提出了一种基于粗粒度模型的并行蚁群算法,该算法采用了一个新的信息素更新策略———Ant-proportion,这种新的更新策略是综合考虑全局和局部信息,依据蚂蚁在搜索过程中所得到的路径的优劣程度和路径中各路段对其贡献的大小来分配信息素增量;另一方面,该算法采用的粗粒度模型充分利用了蚁群算法内在的并行性,使得算法具有更快的收敛速度和更好的优化质量。最后,选用了CHN144问题对该算法进行了检验,算法求得的最优路径优于已知的最优结果。
The ant colony optimization algorithm is a novel simulated evolution algorithm featuring a robust global searching ability. A parallel ant colony optimization algorithm based on coarse-grain model is presented. This algorithm adopts a new pheromone update strategy： Ant-proportion, which incorporates global and local information and allocates the increment of pheromone to links according to the optimization route quality of ants and the link contribution to the route. On the other hand, the coarse-grain model adopted fully utilizes the parallel property of the ant colony optimization algorithm. Thus, the algorithm has higher stability, convergence speed and quality than that of classical ant colony optimization algorithm. At last, the CHN144 problem is applied to test the algorithm and the optimization result is better than the existing ones.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2006年第4期626-629,共4页
Systems Engineering and Electronics
基金
国家自然科学基金资助课题(50479055)
关键词
蚁群算法
并行
搜索能力
ant colony optimization algorithm
parallel
searching ability
