学术文献库

The algebraic approach to the Constraint Satisfaction Problem (CSP) uses high order symmetries of relational structures -- polymorphisms -- to study the complexity of the CSP. In this paper we further develop one of the methods the algebraic approach can be implemented, and apply it to some kinds of the CSP. This method was introduced in our LICS 2004 paper and involves the study of the local structure of finite algebras and relational structures. It associates with an algebra A or a relational structure S a graph, whose vertices are the elements of A (or S), the edges represent subsets of A such that the restriction of some term operation of A is `good' on the subset, that is, act as an operation of one of the 3 types: semilattice, majority, or affine. In this paper we use this theory and consider algebras with edges from a restricted set of types. We prove type restrictions are preserved under the standard algebraic constructions. Then we show that if the types edges in a relational structure are restricted, then the corresponding CSP can be solved in polynomial time by specific algorithms. In particular, we give a new, somewhat more intuitive proof of the Bounded Width Theorem: the CSP over algebra A has bounded width if and only if A does not contain affine edges. Actually, this result shows that bounded width implies width (2,3). Finally, we prove that algebras without semilattice edges have few subalgebras of powers, that is, the CSP over such algebras is also polynomial time. The methods and results obtained in this paper are important ingredients of the 2017 proof of the Dichotomy Conjecture by the author. The Dichotomy Conjecture was also proved independently by Zhuk.