Empirical Evaluation of DisCOP algorithms Academic Article uri icon

abstract

  • All experiments were performed on a simulator in which agents are simulated by threads which communicate only through message passing. The distributed optimization problems used in all of the presented experiments are random MaxDisCSPs. The network of constraints in each of the experiments is generated randomly by selecting the probability p1 of a constraint among any pair of variables and the probability p2, for the occurrence of a violation

publication date

  • January 1, 2008