Agent-based organizational structures for ambient intelligence scenarios

In a sociesty where Information and Communication Technology (ICT) becomes a key component to improve the quality of our daily-life, systems' complexity dramatically increases to better accommodate to an inherent complexity of users' requirements, especially when dealing with pervasive computing. This poses demanding architectural requirements such as massive decentralization and disintermediation along with self-organizing properties. This higher level of system complexity has triggered the birth of several software engineering methodologies that adopt the agent paradigm. Design patterns are a well known approach for capturing and reusing knowledge related to known solution for recurrent architectural problems. Even though the importance of patterns is growing as systems become more and more complex, their development does not keep up with the evolving requirements, as traditional agent patterns are unable to support the majority of current complex social scenarios. This work motivates the need for new agent-based organizational structures that are more flexible than traditional agent patterns such as, Broker and Matchmaker. The newly suggested structure supports the dynamic nature of "Active Environment" where groups of agents are formed for service delivery and then dissolved with no central coordinating mechanism - i.e., gaining the disintermediation property. Serving as a case study, the PEACH and PIL projects have paved the way for experimenting with the new agent-based organizational structures that are more flexible and more suitable than traditional agent patterns for coping with ambient intelligence scenarios.