Closing the Open Shop: Contradicting Conventional Wisdom

This paper describes a new approach for solving disjunctive temporal problems such as the open shop and job shop scheduling domains. Much previous research in systematic search approaches for these problems has focused on developing problem specific constraint propagators and ordering heuristics. Indeed, the common belief is that many of these problems are too difficult to solve without such domain specific models. We introduce a simple constraint model that combines a generic adaptive heuristic with naive propagation, and show that it often outperforms state-of-the-art solvers for both open shop and job shop problems.