A system which will use networks of computerised agents to cope with disaster scenarios such as outbreaks of fires, will be outlined by an ECS researcher on Guy Fawkes night.
Professor Nick Jennings from the University's School of Electronics and Computer Science (ECS) has picked Wednesday 5 November to tell an audience comprised of businesses and academia at Park Centre, Farnborough, about ALADDIN, an ambitious £5.5 million five-year research programme aimed at developing computerised agents for use in disaster recovery or terrorist attacks, which has just reached its half-way point.
ALADDIN stands for Autonomous Learning Agents in Decentralized Data and Information Networks.
Professor Jennings believes that the UK Fire and Rescue Service is a likely potential client for some of the applications already developed by ALADDIN.
"We are developing decentralised information systems that can continue to operate effectively when there is a fire or in other extremely difficult circumstances," said Professor Jennings.
"We use computerised agents which can sense, act and interact in order to achieve individual and collective aims. Central to this endeavour is the effective coordination of the different actors and, to this end, we've developed a rich series of algorithms for inter-agent co-operation and negotiation."
ALADDIN is one of a number of programmes that bring leading academic groups to key industrial challenges through the BAE Systems/EPSRC Strategic Partnership, with the objective of delivering high-quality research and enhancing industrial capabilities.
The ALADDIN team comprises internationally-leading researchers in complex adaptive systems from the Universities of Southampton, Bristol and Imperial College; in fusion, inference and learning from the University of Oxford and Imperial; and in decentralised architectures from BAE Systems.
One of the challenges facing the researchers is to bring together work from a number of hitherto distinct fields, such as information fusion, inference, decision-making, and machine-learning. This work then needs to be combined with work from multi-agent systems, game theory, mechanism design and mathematical modelling of collective behaviour in order to give a collective view on behaviour.
"As ever more information sources become available - through environmental sensors, intranets, and so on - the problem of obtaining and fusing the right information when making decisions and taking actions is becoming increasingly pressing," said Professor Jennings.