Modeling Decentralized Information Flow in Ambient Environments

This paper proposes a decentralized approach for modeling information flow in ambient environments. We study how query and notification mechanisms can be used to reduce the amount of information exchanged between agents. We will propose qualitative criteria which state whether querying a concept is appropriate given the logical structure of an agent’s knowledge base. Furthermore, we will propose quantitative criteria which state which concept is most likely to be most informative, given an agent’s information needs and its experience with past events.