Enhancing the role of multi-agent systems in the development of intelligent environments

Abstract

The development of Intelligent Environments is a complex challenge. This complexity arises, in part, from the amount of different devices that need to be seamlessly integrated in a common and homogeneous environment, despite the fact of each device having its own characteristics. This heterogeneity of devices is particularly risky when one passes from the specification to the implementation phase, where all unexpected things start to happen. Multi-agent systems are the paradigm par excellence for implementing Intelligent Environments. However, traditionally, agents are only used at the implementation phase. In this paper we propose a new 3 step approach in which agents are used during all the development process, playing undoubtedly a much more preponderant role and making the path from the specification to the implementation a much easier and controllable one, always having in mind the challenges of the development of Intelligent Environments.