[Paper Review] A Critique of a Polynomial-time SAT Solver Devised by Sergey Gubin

Ian Christopher, Dennis Huo+1 more|ArXiv.org|Apr 16, 2008

Logic, programming, and type systems3 references3 citations

TL;DR

This paper critiques Sergey Gubin's proposed polynomial-time algorithm for 3-SAT, which claimed to solve NP-complete problems in polynomial time, implying P = NP. The authors demonstrate that Gubin's matrix-compatibility and depletion method fails to preserve logical equivalence across clause combinations, and his reduction of SAT to 2-SAT is invalid because it can produce satisfiable 1-SAT instances from unsatisfiable 3-SAT formulas, thus refuting the algorithm's correctness.

ABSTRACT

This paper refutes the validity of the polynomial-time algorithm for solving satisfiability proposed by Sergey Gubin. Gubin introduces the algorithm using 3-SAT and eventually expands it to accept a broad range of forms of the Boolean satisfiability problem. Because 3-SAT is NP-complete, the algorithm would have implied P = NP, had it been correct. Additionally, this paper refutes the correctness of his polynomial-time reduction of SAT to 2-SAT.

Motivation & Objective

• To rigorously evaluate the validity of Sergey Gubin's polynomial-time algorithm for 3-SAT, which claims to resolve an NP-complete problem in polynomial time.
• To identify and expose logical flaws in Gubin's matrix-compatibility and depletion mechanism for combining clauses.
• To demonstrate that the algorithm's final compatibility matrix does not preserve logical equivalence to the original formula, leading to incorrect satisfiability assessments.
• To refute Gubin's claimed reduction of SAT to 2-SAT, showing it can yield satisfiable 1-SAT instances from unsatisfiable 3-SAT formulas.
• To uphold the current theoretical consensus that P ≠ NP by proving the algorithm's fundamental flaws

Proposed method

• The authors analyze Gubin's algorithm by reconstructing his matrix-compatibility method, where each clause pair is represented as an 8×8 matrix of truth table row combinations.
• They define compatibility as mutual satisfaction of truth table rows with no conflicting variable assignments across shared variables.
• The algorithm performs iterative 'depletion' by eliminating matrix entries where no consistent assignment exists across three clauses, using a triangular matrix structure.
• The method relies on the assumption that the final bottom-right matrix preserves full satisfiability information, which the authors challenge.
• The authors test the algorithm on counterexamples, showing that depletion order affects results and that the final matrix may contain non-zero entries even when the formula is unsatisfiable.
• They refute Gubin's reduction of SAT to 2-SAT by showing that the resulting 1-SAT instance is always satisfiable, regardless of the original 3-SAT formula’s status.

Experimental results

Research questions

• RQ1 Does Gubin’s matrix-compatibility and depletion method correctly preserve the satisfiability of the original 3-SAT formula across all clause combinations?
• RQ2 Can the final compatibility matrix after depletion accurately represent the full conjunction of all clauses, or is information lost due to transitivity and ordering effects?
• RQ3 Is Gubin’s reduction of SAT to 2-SAT valid, particularly when it maps unsatisfiable 3-SAT instances to satisfiable 1-SAT instances?
• RQ4 Does the algorithm’s polynomial-time complexity hold under correct logical semantics, or is it compromised by incorrect assumptions about clause compatibility?
• RQ5 Can the algorithm be salvaged by tracking all possible assignments, and would such a fix result in exponential time complexity?

Key findings

• Gubin’s algorithm fails to correctly determine satisfiability because the final compatibility matrix may contain non-zero entries even when the original 3-SAT formula is unsatisfiable.
• The depletion process is order-dependent, and different clause ordering sequences can lead to different results, invalidating the algorithm’s deterministic claim.
• The algorithm’s claim that the bottom-right matrix is logically equivalent to the full conjunction of all clauses is false, as compatibility is not transitive under Gubin’s method.
• Gubin’s reduction of SAT to 2-SAT is invalid because it generates a 1-SAT instance that is always satisfiable, even when the original 3-SAT instance is unsatisfiable.
• The authors demonstrate that the algorithm’s core mechanism—using pairwise compatibility matrices with depletion—cannot preserve full logical equivalence across multiple clauses.
• The paper concludes that Gubin’s algorithm does not prove P = NP, as it is fundamentally flawed in its treatment of clause interactions and satisfiability propagation.