Robert N. Lass, Evan A. Sultanik, William C. Regli
What local action can agents take, without the benefit of global knowledge, to produce the best global solution? Many dynamic distributed systems can be modeled using techniques from distributed constraint reasoning, however, existing work in the distributed constraint reasoning community does not address the true dynamism inherent in many real-world systems. This paper describes a formal model for dynamic distributed constraint reasoning, presents an example of a dynamic distributed system, and shows how the model could be applied to the example. Finally, a new algorithm and a classification scheme for algorithms solving these problems are proposed.
Subjects: 7.1 Multi-Agent Systems; 15.2 Constraint Satisfaction
Submitted: Apr 15, 2008