ECAI 2004 Conference Paper

[full paper]

Role Swapping in Multi-Agent Sensor Webs

Ronan Mac Ruairi, Mark T. Keane

This paper reports on the application of multi-agent systems to the design of autonomic sensor networks. In advance of the current hardware/software capabilities of these sensor networks, we simulate an autonomic sensing network handling the detection of a contaminant, in a hostile environment, where the effective life of the network has to be optimised (through power handling). By casting each sensor as an agent that can swap its role (Produce or Integrator, both of which have different power usages) when the conditions in the network change we show that networks using role-swapping, as opposed to fixed roles, can be much more effective over time. This emergent behaviour, that arises from economical local interactions, also becomes more effective as the environment for the network becomes progressively harsher.

Keywords: Multi-Agent Systems, Sensor Web, Autonomous Agents, Distributed AI

Citation: Ronan Mac Ruairi, Mark T. Keane: Role Swapping in Multi-Agent Sensor Webs. In R.López de Mántaras and L.Saitta (eds.): ECAI2004, Proceedings of the 16th European Conference on Artificial Intelligence, IOS Press, Amsterdam, 2004, pp.1055-1056.