Abstract: A grid resource discovery model which combines ACO and mobile agent is presented based on grid cluster theory. The mobile agent technology is employed to collect information of grid resources and store the abstracts of grid service information in clusters’ cache. The cached service information abstracts are used as the heuristic factor of ACO, so that the accuracy of the pheromone can be improved in ACO. The improvement of the accuracy and the speed of the grid service discovery can increase the efficiency of the grid. Simulations have run and the result of the simulations indicate that the designed service discovery mechanism is available and efficient.

Key words: grid, service discovery, mobile agent, Ant Colony Algorithm（ACO）

摘要： 在对网格进行分簇的基础上，设计了一种基于蚁群算法及移动Agent的网格服务发现机制模型。通过使用移动Agent技术收集网格服务信息，并将收集到的网格服务信息摘要进行跨簇缓存，以提高蚁群算法中信息素信息的准确性，并且在进行服务查找时采用之前缓存的服务摘要信息作为蚁群算法的启发因子，从而提高网格服务发现的准确性以及发现速度，进而提高网格使用效率。通过仿真实验验证了所述服务发现机制的有效性。

关键词: 网格, 服务发现, 移动代理, 蚁群算法