Distributed Cooperative Optimization on Cluster Trees

Resource allocation problems on resource supply networks are formalized with Distributed Constraint Optimization Problems. In previous studies, solution methods based on pseudo trees have been proposed. However, when the pseudo trees contain nodes of high degree and a large number of cycles, those methods are not applicable due to the large size of local problems in agents. Here, we employ cluster trees that hierarchically divide the problem into sub-problems. With optimistic approximation, solution methods on the cluster tree are applied to several large problems.