An agent-based knowledge sharing model for information retrieval on the Internet

With the proliferation of electronically available information, and its diverse nature, a critical problem has arisen which is how to locate, retrieve and process relevant information effectively. It is impractical to build either a unified peer-to-peer system or a centralised architecture that combines all of these information resources. A more promising approach is to program specialised informative agents within an organisational structure that provide public access to information resources on the Internet. These agents can co-operate with other agents to share their knowledge when appropriate, This paper describes an integrated approach to developing this type of agent-based information retrieval system model, which employs a user-modelling strategy, co-operation methods and matchmaking middle-agents. The feasibility of the resultant knowledge-sharing environment is demonstrated by experimental results from a simulation based on requirements from the UK National Health Service.