Schaefers theorem for graphs

Schaefer’s theorem is a complexity classification result for so-called Boolean constraint satisfaction problems: it states that every Boolean constraint satisfaction problem is either contained in one out of six classes and can be solved in polynomial time, or is NP-complete. We present an analog of this dichotomy result for the propositional logic of graphs instead of Boolean logic. In this generalization of Schaefer’s result, the input consists of a set W of variables and a conjunction \Phi\ of statements (“constraints”) about these variables in the language of graphs, where each statement is taken from a fixed finite set \Psi\ of allowed quantifier-free first-order formulas; the question is whether \Phi\ is satisfiable in a graph. We prove that either \Psi\ is contained in one out of 17 classes of graph formulas and the corresponding problem can be solved in polynomial time, or the problem is NP-complete. This is achieved by a universal-algebraic approach, which in turn allows us to use structural Ramsey theory. To apply the universal-algebraic approach, we formulate the computational problems under consideration as constraint satisfaction problems (CSPs) whose templates are first-order definable in the countably infinite random graph. Our method to classify the computational complexity of those CSPs is based on a Ramsey-theoretic analysis of functions acting on the random graph, and we develop general tools suitable for such an analysis which are of independent mathematical interest.

💡 Research Summary

The paper presents a dichotomy theorem for constraint satisfaction problems (CSPs) formulated in the language of graphs, extending Schaefer’s classic dichotomy for Boolean CSPs. In the graph setting, an instance consists of a finite set of variables W and a conjunction Φ of constraints, each of which is a quantifier‑free first‑order formula describing adjacency relations among the variables. All allowed constraint types belong to a fixed finite set Ψ. The decision problem asks whether there exists a (finite or infinite) graph that satisfies all constraints in Φ.

To analyse this problem the authors embed it into the framework of CSPs over a countably infinite template that is first‑order definable in the random (Rado) graph G(ℵ₀). The random graph enjoys ultrahomogeneity, universality, and a strong structural Ramsey property, making it an ideal host for a universal‑algebraic approach. In this approach the computational complexity of a CSP is determined by the clone of polymorphisms (operations preserving all relations of the template). If the clone contains only “trivial” operations (e.g., projections, constant functions, or simple Boolean‑like operations) the CSP is tractable; otherwise the presence of more complex operations typically yields NP‑completeness.

The authors develop a Ramsey‑theoretic analysis of functions acting on the random graph. By exploiting the Ramsey property they can canonically decompose any polymorphism into a finite set of normal forms. This leads to the identification of 17 distinct classes of graph formulas, each characterized by a specific algebraic property of its polymorphism clone (for example, preservation of equality, preservation of non‑equality, existence of a majority‑type operation, or invariance under a graph complement).

The main theorem states that for any finite set Ψ of allowed graph formulas, exactly one of the following holds:

1. Tractable case – Ψ is contained entirely within one of the 17 identified classes. In this situation the associated CSP can be solved in polynomial time. The authors give concrete algorithms based on tree‑width‑bounded decompositions, homomorphism‑testing, or simple propagation procedures that exploit the regular structure guaranteed by the corresponding class.

2. Hard case – Ψ does not lie in any of the 17 classes. Then the CSP is NP‑complete. The proof proceeds by a reduction from classic NP‑complete problems such as 3‑SAT or graph 3‑colorability, using the expressive power that appears once the polymorphism clone escapes the tractable regimes.

The paper’s methodology intertwines three major ingredients: (i) the model‑theoretic description of the random graph, (ii) universal algebraic classification of polymorphism clones, and (iii) structural Ramsey theory that supplies the canonical forms needed for the algebraic analysis. This combination not only yields the dichotomy for graph‑logic CSPs but also provides a template for extending Schaefer‑type classifications to other ω‑categorical structures (e.g., the dense linear order, infinite lattices).

In addition to the dichotomy, the authors develop several auxiliary tools of independent interest: a systematic way to translate graph‑logic constraints into relational structures definable in the random graph, a Ramsey‑based normal‑form theorem for functions on homogeneous graphs, and a set of reduction techniques that respect the algebraic properties of polymorphisms. These contributions deepen the connection between finite‑domain CSP complexity, infinite‑domain model theory, and combinatorial Ramsey theory.

Overall, the work establishes a complete complexity classification for propositional logic of graphs, mirroring Schaefer’s theorem in a richer relational setting, and opens new avenues for applying universal‑algebraic and Ramsey‑theoretic methods to the study of infinite‑domain CSPs.