[Paper Review] Notes on CSPs and Polymorphisms

Zarathustra Brady|arXiv (Cornell University)|Oct 13, 2022

Constraint Satisfaction and Optimization4 citations

TL;DR

This paper presents a comprehensive algebraic framework for understanding the complexity of Constraint Satisfaction Problems (CSPs) through the lens of polymorphisms, varieties, and term operations. It establishes that finite Taylor algebras are affine, proves bounded-width CSPs are solvable via cycle consistency, and develops algorithms for conservative and few-subpowers CSPs using absorption, Mal'cev terms, and compact representations—culminating in a complete algebraic characterization of tractable CSPs.

ABSTRACT

These are notes from a multi-year learning seminar on the algebraic approach to Constraint Satisfaction Problems (CSPs). The main topics covered are the theory of algebraic structures with few subpowers, the theory of absorbing subalgebras and its applications to studying CSP templates which can be solved by local consistency methods, and the dichotomy theorem for conservative CSP templates. Subsections and appendices cover supplementary material.

Motivation & Objective

• To develop a unifying algebraic framework for classifying the tractability of CSPs using polymorphisms and term operations.
• To characterize finite Taylor algebras as affine, resolving a key open question in universal algebra and CSP theory.
• To establish that bounded-width CSPs are solvable via cycle consistency, extending previous results on width-1 and bounded strict width.
• To provide algorithmic solutions for conservative and few-subpowers CSPs using absorption, compact representations, and Mal'cev operations.
• To resolve the meta-problem for conservative CSP templates by constructing a polynomial-time algorithm based on absorption and term operations.

Proposed method

• Uses the Inv-Pol Galois connection to link relational clones and algebras, enabling structural analysis of CSP templates.
• Applies Birkhoff’s HSP theorem and variety theory to analyze the equational properties of algebras underlying CSPs.
• Introduces compact representations of relations via Mal'cev and parallelogram terms to enable efficient algorithmic manipulation.
• Employs absorption and Jónsson absorption to reduce CSP instances to smaller, equivalent subinstances using local criteria.
• Develops cycle-consistency and arc-consistency algorithms as solvers for bounded-width and ancestral CSPs.
• Leverages Zhuk’s centers and semidefinite programming to robustly solve bounded-width CSPs, with a focus on stability and reversibility.

Experimental results

Research questions

• RQ1 Under what algebraic conditions is a CSP tractable, and how do polymorphisms and term operations determine this?
• RQ2 Can finite Taylor algebras be shown to be affine, and what does this imply for the complexity of their CSPs?
• RQ3 To what extent can cycle consistency solve bounded-width CSPs, and what role do Mal'cev and parallelogram terms play?
• RQ4 How can absorption and compact representations be used to design efficient algorithms for conservative and few-subpowers CSPs?
• RQ5 What is the computational complexity of the meta-problem for conservative CSP templates, and can it be solved in polynomial time?

Key findings

• Finite Taylor algebras are affine, which implies that their CSPs are solvable via linear programming and cycle consistency.
• Bounded-width CSPs are solvable by cycle consistency, and this holds robustly even under perturbations, as shown via semidefinite programming.
• Algebras with few subpowers are finitely related, and their relations admit compact representations via generalized majority-minority operations.
• The presence of a partial semilattice term with no neutral element allows for elimination of such algebras in conservative CSPs, enabling recursive algorithm design.
• For conservative CSPs, the meta-problem is solvable in polynomial time using a recursive strategy based on absorption and retract operations.
• The algorithm for conservative CSPs is based on Maróti’s reduction and Bulatov’s framework, combining few-subpowers and bounded-width techniques to achieve full tractability.