一种移动代理巡行规划算法

An algorithm for mobile agents itinerary planning

在基于移动代理的分布式信息系统中,单一代理巡行节点数量过多会延长系统响应时间并造成任务完成速度缓慢,一般通过派遣多个代理来加快任务的完成。提出了一种在已知节点间带宽和延迟的条件下,将较大待巡行节点集聚类划分为若干子集并优化子集内巡行路径的启发式算法。在不同网络环境下对算法的效果进行了仿真测试,结果显示在相同的分组数量下与平均随机划分相比,该算法能明显降低任务完成时间,且广域网环境下效果优于局域网。

In the distributed information systems based on mobile agent technology,using single mobile agent to travel all nodes in the systems will prolong the system responding time.One solution is dispatching multi agents simultaneously.A heuristic algorithm is proposed in which the set of itinerated nodes is divided into several independent subsets and the itinerary path in each subset is optimized.The simulation in different network environment verifies that the algorithm can evidently reduce the execution time of itinerary task in contrast to random grouping algorithm.Furthermore,the effect in WAN is better than its in LAN.