Logic and Constraint Logic Programming for Distributed Constraint Optimization
share
The field of Distributed Constraint Optimization Problems (DCOPs) has gained momentum, thanks to its suitability in capturing complex problems (e.g., multi-agent coordination and resource allocation problems) that are naturally distributed and cannot be realistically addressed in a centralized manner. The state of the art in solving DCOPs relies on the use of ad-hoc infrastructures and ad-hoc constraint solving procedures. This paper investigates an infrastructure for solving DCOPs that is completely built on logic programming technologies. In particular, the paper explores the use of a general constraint solver (a constraint logic programming system in this context) to handle the agent-level constraint solving. The preliminary experiments show that logic programming provides benefits over a state-of-the-art DCOP system, in terms of performance and scalability, opening the doors to the use of more advanced technology (e.g., search strategies and complex constraints) for solving DCOPs.
READ FULL TEXT