摘要
围绕TSP问题研究了基本蚁群算法。在此基础上,研究了串行蚁群算法的并行策略,使用C++语言调用MPI接口函数实现了并行蚁群算法。最后,分析研究了影响并行蚁群算法的因素,采用了更高效的信息素更新和变参数机制对并行蚁群算法进行了改进。通过仿真实验分析表明,改进的并行蚁群算法有较广泛的适用性,与基本蚁群算法相比,具有更高的精度和更短的收敛时间。
On the basis of studying ant colony algorithm on TSP, the parallel strategy of serial ant colony algorithm is studied. The parallel ant colony algorithm is completed, using the C＋＋ programming language and calling MPI interface timetions. Finally, the element is analyzed that impact the parallel ant colony algorithm, and the parallel ant colony algorithm is improved by adopting more efficient method of the pheromone update and change parameters mechanism. The simulation experiments shows that the improved parallel ant algorithrn has the broader applicability, higher precision and shorter time of convergence, compared with the basic ant algorithm.
出处
《长春理工大学学报（自然科学版）》
2008年第4期109-112,共4页
Journal of Changchun University of Science and Technology(Natural Science Edition)
基金
国家863计划项目
关键词
蚁群算法
并行计算
集群
优化
ant colony algorithm
parallel computing
cluster
optimization
