Click here for full text:
Conflicting Agents in Distributed Search
Keyword(s): distributed constraint satisfaction; distributed AI; collaborative software agents; search
Abstract: We extend here the work on interleaved distributed graph based backjumping, IDIBT/GBJ by considering conflicting variables. The resulting method IDIBT/CBJ combines distributed and parallel exploration with more efficient backward phases. We take advantage of our asynchronous framework to implement a more refined and efficient update of detected conflicts. Finally, directed k-consistency is added to our new method which gives IDIBT/CBJ-DkC. We show that this last method extends cooperation in the system.
Back to Index