随着网格规模扩大,其工作效率明显降低,为此提出了一种多代理方法,对网格资源信息进行空间排序和发现.受生物学现象(如传染病传播、蚁群觅食)启发,该方法通过代理对资源描述符进行复制和散发,从而促进资源发现.仿真分析证实了基于代理的重组和发现协议的高效性.“,”Along with the extension of a grid system,its work efficiency decreases obviously. Inspired by biological phenomena (e.g. spreading of infectious diseases and feeding of ant colonies),a multi-agent approach to spatially sort and discover information about the resources in a grid system is proposed in this paper. This method can facilitate resource discovery by the operation of the agents' replicating and distributing resource descriptors. Simulation analysis proved the effectiveness of the reorganization and discovery protocols based on agent.