Constraint Satisfaction Problem Dichotomy for Finite Templates: a Proof Via Consistency Checks (1706.03451v6)

Abstract: One of the central problems in the study of parametrized constraint satisfaction problems is the Dichotomy Conjecture by T. Feder and M. Vardi stating that the constraint satisfaction problem (CSP) over a fixed, finite constraint language is either solvable in polynomial time or \textsc{NP}-complete. The conjecture was verified in certain special cases (domains with a relatively small number of elements, constraint languages containing all unary relations, etc.) In this article, we present a proof of the Dichotomy Conjecture via local consistency and AF- consistency checks. In fact, we show that, for every Taylor domain, which is $(2\lceil\frac{K}{2}\rceil,3\lceil\frac{K}{2}\rceil)$-consistent, where $K$ is the largest arity of a relation in the constraint language, we can define polynomially many proper subinstances such that, the original instance of the CSP is solvable if, and only if, the problem has a solution in one of those subinstances.. Finally, a solution is constructed using the combination of SLAC (Singleton Linear Arc Consistency), introduced by M. Kozik, and AF-consistency.