Computer science Computational complexity: resources and classes Theorem
Schaefer's dichotomy theorem
Builds on Constraint satisfaction problemBoolean satisfiability problemPNP-completenessConjunctive normal formLadner's theorem
Statement
Let be a finite set of nonempty relations on . Let be the constraint satisfaction problem with domain . Then lies in P if at least one of the following holds.
- Every relation in contains the all-zero tuple.
- Every relation in contains the all-one tuple.
- Every relation in is the set of models of a conjunctive normal form whose clauses each have at most one positive literal.
- Every relation in is the set of models of a conjunctive normal form whose clauses each have at most one negative literal.
- Every relation in is the set of models of a conjunctive normal form whose clauses each have at most two literals.
- Every relation in is the set of solutions of a system of linear equations over the field .
Otherwise is NP-complete.
Remarks
The classes are called -valid, -valid, Horn, dual Horn, bijunctive and affine. In cases 1 and 2 the constant assignment works. In cases 3 and 4 unit propagation decides the problem, in case 5 the method for 2-SAT does, and in case 6 Gaussian elimination does. For the hardness, assume that none of the six conditions holds. Schaefer showed that the relations of then express every Boolean relation by existentially quantified conjunctions, possibly with the help of constants. So SAT reduces to . Thomas Schaefer proved the theorem in 1978.
Example: 3-SAT is for the eight relations given by clauses of three literals. The relation of the clause fails conditions 1, 3, 5 and 6, and the relation of fails conditions 2 and 4. So 3-SAT is NP-complete, in agreement with the Cook–Levin theorem.
Going deeper
For a student
The theorem is the two-element case of the dichotomy for constraint satisfaction problems on finite domains. That general dichotomy, conjectured by Feder and Vardi, was proved by Bulatov and by Zhuk in 2017. By Ladner's theorem, NP contains intermediate problems if , so a dichotomy for a whole family is a strong structural fact. The six tractable cases are described by closure under polymorphisms: constants, the binary minimum, the binary maximum, the majority operation, and the operation modulo .
Reading path
To understand this page from the ground up, read these 18 pages in order. The foundations come first.
Show the path as a graph
- no evidence level
- Conjunctive normal form concept
- Graph homomorphism concept
- NP concept
- NP-completeness concept
- P concept
- Polynomial-time reduction concept
- Proposition concept
- Satisfiability concept
- Constraint satisfaction problem problem
- Ladner's theorem statement
- Conjunction definition
- Disjunction definition
- Implication definition
- Negation definition
- Biconditional definition
- Propositional logic concept
- Boolean satisfiability problem problem
- Schaefer's dichotomy theorem statement
Further reading
About this page
| Tag | HFW5G8EG99 |
|---|---|
| URI | m2://HFW5G8EG99 |
| Kind | statement, theorem |
| Writing | W0 stub |
| MSC | 68Q17, 03B05 |
| Status | This copy of the site has no build data. Run m2 status HFW5G8EG99. |
| About | Constraint satisfaction problem |
| Uses | Constraint satisfaction problem, Boolean satisfiability problem, P, NP-completeness, Conjunctive normal form, Ladner's theorem |
| Same as | |
| Changed | 3 November 2026 |
Markdown · JSON · Edit on GitHub
What this page needs for W1
- an inventory row: the row of the inventory that the node accounts for (0 of 1)
- a source with a locator (0 of 1)
The writing levels say how complete the text of a page is. They do not measure its truth.
What links here
No page links to this page yet.