Packet Routing with Genetically Programmed Mobile Agents
Jon Schuringa, Günter Remsak · 2000
This paper presents novel mobile agent routing techniques. We use genetic programming to build mobile agents that monitor the network status and set the routing tables in the network nodes in such a way that it maximizes network throughput and minimizes the overall packet delay. Performance is measured by simulation using a realistic network and traffic model that is capable of generating fractal traffic. The result is a high performance, self-configuring routing method, irrespective of the network topology and traffic.