An incomplete and asynchronous algorithm based on localization for Distributed Constraint Optimization

A Distributed Constraint Optimization Problem (DCOP) is a fundamental problem that can formalize various applications related to multi-agent cooperation and is a popular framework for cooperative multi-agent decision making. Since it is NP-hard, considering faster incomplete algorithms is necessary for large-scale applications. Most incomplete algorithms generally do not provide any guarantees on the quality of solutions. Some notable exceptions are DALO, the bounded max-sum algorithm, and ADPOP. One of the incomplete algorithms, that working based on bounds on solution quality is k-size optimality, which creates local optimality groups of neighbor agents. Another incomplete algorithm is t-distance optimality which departs from k-size optimality by using graph distance for selecting groups of deviating agents. In t-distance optimality, quality bounds improve rather than bounds for k-size optimality. But unfortunately, both of these algorithms, has problem in their initialization phase, and initialize the variables´ value in random or by zero. In this paper, we develop a new efficient asynchronous local search algorithm for finding both k-size and t-distance optimal solutions, that uses an intelligent initialization process. Indeed, empirical results show that this algorithm relatively outperforms the only two existing algorithms for finding general k-size and t-distance optimal solutions.