Resource allocation in communication networks using market-based agents

This work describes a system that allocates end-to-end bandwidth, in a switched meshed communications network. The solution makes use of market-based software agents that compete in a number of decentralised marketplaces to buy and sell bandwidth resources. Agents perform a distributed depth first search with decentralised markets in order to allocate routes for calls. The approach relies on a resource reservation and commit mechanism in the network. Initial results show that under a light network load, the system sets up a high percentage of calls which is comparable to the optimum value and that, under all network loads, it performs significantly better than a random strategy.