Click here for full text:
Distributed, Interleaved, Parallel and Cooperative Search in Constraint Satisfaction Networks
Keyword(s): distributed AI; constraint satisfaction; search
Abstract: In this work, we extend the efficiency of distributed search in constraint satisfaction networks. Our method adds interleaving and parallelism into distributed backtrack search. Moreover, it has a filtering capacity that makes it open to cooperative work. Experimentations show that 1) the shape of phase transition with random problem can be characterized, 2) important speed-up can be achieved when the distribution of solutions is non uniform.
Back to Index