Adaptive On-line Co-ordination of Ubiquitous Computing Devices with Multiple Objectives and Constraints

This paper presents a novel embedded agent controller architecture that uses a unique multi-objective and multi-constraint evolutionary system to co-ordinate real world ubiquitous computing devices. The devices operate in real world environments and should be able to learn the users’ preferences without having the users to preconfigure them. The system of embedded agents should be fault-tolerant in the sense that it should be able to handle the agents breaking down or being added into the system. The experiments presented in this paper will test the embedded agent architecture’s performance and how it adapts to sudden changes on-line. We will also show how the evolutionary system copes as the number of devices and hence the complexity of the problem increases.