Core-Guided and Core-Boosted Search for CP

Core-guided search has proven to be the state-of-the-art in finding optimal solutions for maximum Boolean satisfiability and these techniques have recently been successfully imported in constraint programming While effective on a wide range of problems, the methods are direct translations of their propositional logic counterparts. We propose two reformulation techniques that take advantage of the rich formalism offered by constraint programming rather than relying on propositional logic strategies, and generalise two existing techniques to improve core-extraction and the overall performance. Our experiments demonstrate the effectiveness of our approaches over the conventional (core-guided) CP methods, both in terms of proving optimality and quickly computing high-quality solutions.