Automated Trading in Agents-based Markets for Communication Bandwidth

Automated agents for electronic markets are very valuable when a large amount of information must be processed quickly or employing human traders is not cost-effective. In markets for communication bandwidth, rights to transmit data are traded through a network. Since demand fluctuates considerably every few seconds, agent-based spot markets provide extra liquidity. This paper considers the design of agents for use in a double auction market for communication bandwidth. The suggested criteria and framework for building adaptive agents are based on statistical decision theory. The proposed agents can differentiate stable from unstable market conditions and respond appropriately.