You are here

Solving the Quorumcast Routing Problem by Constraint Programming

The quorumcast routing problem is a generalization of multicasting which arises in many distributed applications. It consists of finding a minimum cost tree that spans the source node r and any q out of m specified nodes on a given undirected weighted graph. This paper proposes a complete and an incomplete approach, both based on the same Constraint Programming (CP) model, but with two different specific search heuristics based on shortest paths. Experimental results show the efficiency of the two proposed approaches. Our complete approach (CP model + complete search) is better than the state of the art complete algorithm and our incomplete approach (CP model + incomplete search) is better than the state of the art incomplete algorithm. Moreover, the proposed complete search is better than the standard First-Fail search in the same CP model.

The benchmark experimented in the paper can be found here. The source code of the algorithms can be requested by sending email to