Pseudo-Tree Based Hybrid Algorithm for Distributed Constraint Optimization

A Distributed Constraint Optimization Problem (DCOP) is a fundamental problem that can formalize various applications related to multi-agent cooperation. Considering pseudo-tree based search algorithms is important in DCOPs, since their memory requirements are polynomial in the number of agents. However, these algorithms require a large run time. Thus, how to speed up pseudo-tree based search algorithms is one of the major issues in DCOPs. In this paper, we propose a novel hybrid algorithm which combines a complete algorithm with an incomplete algorithm. Specifically, we use a state-of-the-art complete search algorithm (BnB-ADOPT) and utilize the bounds obtained by an approximate algorithm (p-optimal algorithm) in preprocessing. In the evaluations, we show that this hybrid algorithm outperforms the stateof-the-art DCOP search algorithm. Furthermore, we verify experimentally that a pseudo-tree based approximate algorithm is well-suited with a pseudo-tree based search algorithm.