Using Multi-agent Systems to Manage Community Care

摘要

This paper discusses the evolution of the INCA demonstrator through a number of re-implementations that have investigated the applicability of various aspects of multi-agent technology to the management of Community Care. The latest experiences are described, making full use of the latest developments in semantic agents to provide a richer, more rigorous and highly scalable implementation than the previous demonstrators. This is presented in the context of a simulated real-world environment, based on knowledge of the actual operational environment within which the fully deployed agents would be expected to work. In particular the grouping of different communities of agents so that scalable solutions can be fully implemented and rigorously tested. So for example the agents that one would normally expect within a single household, including the home unit and the associated sensors, alarms etc. are treated as one group, and a care provider and the associated carers also. Not only does this reduce the communications overhead but also leads to simplifications in implementation as each class of agent only needs to be implemented once and individual instances are characterized by initial configurations and their interactions with their peers.