蚂蚁网络中后效问题的研究

Research on after-effect problem in antnet

蚂蚁网络算法(简称蚂蚁算法)是通讯网络的一种新型路由算法。在算法中,智能体通过使用它们在探测网络时收集的信息进行间接的、不同时的交流,来修正传输数据包所需的路由表中的信息。但是,由于网络中存在了大量的智能体且与数据包分享相同的队列,使得网络的局部信息不能更快的反映到路由表中去,不能很好的解决后效问题。本文对蚂蚁算法中智能体的优先权进行了改进,并在OMNeT++仿真平台上加以实现。仿真结果表明本文的模型通过对优先权的改进,使得网络中的信息传播更加迅速,整个网络的吞吐量得到了提高。

Antnet algorithm (antnet for short) is a new routing algorithm for communications networks. In the algorithm, agents communicate indirectly and asynchronously through the collected information during the network exloration and modify the information in the routing table that is used to transmit the packets. However, as a number of agents are involved in the network and share the same queue with packets, the local network information cannot transmit to routing table quickly enough, and therefore, cannot solve the after-effect problem well enough. This paper is designed to suggest an improvement on the priority of agents in antnet and simulate on OMNeT++ platform. The simulation result shows that the improvement of priority helps the transmission of information go more quickly and the throughput in the whole network is further increased.