Search combinators

The ability to model search in a constraint solver can be an essential asset for solving combinatorial problems. However, existing infrastructure for defining search heuristics is often inadequate. Either modeling capabilities are extremely limited or users are faced with a low-level programming language and modeling search becomes unwieldy. As a result, major improvements in perfor- mance may remain unexplored. This paper introduces search combinators, a lightweight and solver-independent method that bridges the gap between a conceptually simple search language (high- level, functional and naturally compositional) and an efficient implementation (low-level, imperative and highly non-modular). Search combinators allow one to define application-tailored strategies from a small set of primitives, resulting in a rich search language for the user and a low implementation cost for the de- veloper of a constraint solver. The paper discusses two modular implementation approaches and shows, by empirical evaluation, that search combinators can be implemented without overhead compared to a native, direct implementation in a constraint solver