A polynomial upper bound on Reidemeister moves for each link type

Marc LackenbyarXiv 2026 · DOI · arXiv

Abstract

For each link type $K$ in the 3-sphere, we show that there is a polynomial $p_K$ such that any two diagrams of $K$ with $c_1$ and $c_2$ crossings differ by at most $p_K(c_1) + p_K(c_2)$ Reidemeister moves. As a consequence, the problem of recognising whether a given link diagram represents $K$ is in the complexity class NP and hence can be completed deterministically in exponential time. We calculate this polynomial $p_K$ explicitly for various classes of links.

Introduction

In knot theory, a fundamental challenge is determining if two distinct diagrams represent the same link type. While it has long been known that equivalent diagrams can be transformed into one another via a sequence of Reidemeister moves, the number of moves required remained elusive. Previous upper bounds were "towers of exponentials," rendering them practically uncomputable. This research addresses the "link recognition problem" by establishing a polynomial upper bound on the number of Reidemeister moves required for a fixed link type $K$, effectively placing the problem within the complexity class NP.

Methodology

The proof employs a sophisticated combination of 3-manifold topology and geometric complexity theory. Central to the approach is Normal Surface Theory, which studies surfaces within a 3-manifold by triangulating the space into tetrahedra. A normal surface intersects these tetrahedra in a standard way, forming specific elementary disks.

Figure 3: Normal surface disks: triangles and squares within a tetrahedron.

Key methodological components include:

  1. Haken Hierarchies: Sequences of surfaces that progressively simplify a 3-manifold (such as a link exterior) into 3-balls.
  2. Arc Presentations: The use of rectangular diagrams, where links consist of horizontal and vertical segments, providing a combinatorial framework to measure complexity.
  3. Handle Structures: Controlling the isotopy of a link through a handle decomposition of the 3-sphere to count the resulting Reidemeister moves.

Key Insights: Managing Topological Complexity

A critical insight of this work is that for a fixed link type, the complexity of the 3-manifold exterior remains constant. This allows the crossing number of the input diagram to serve as the primary variable in the polynomial bound. To prevent exponential growth during the hierarchy process, the author utilizes "normal parallelism"—identifying identical layers in complex surfaces to perform "big moves" that reduce complexity efficiently.

Figure 4: Examples of "almost clean" and "necklace" surfaces used to manage topological complexity.

Additionally, the paper introduces "wedge insertion," a topological move designed to handle boundary vertices during the hierarchy process, ensuring the link remains properly embedded without inflating the arc index.

Results and Computational Complexity

The primary result is Theorem 1.1, which guarantees the existence of a polynomial $p_K$ such that diagrams with $c_1$ and $c_2$ crossings differ by at most $p_K(c_1) + p_K(c_2)$ Reidemeister moves. This places the recognition of a specific link $K$ in NP, meaning a proposed sequence of moves can be verified in polynomial time.

The author provides explicit polynomial calculations for specific cases:

  • Figure-eight knot: $p_K(c) = (10^{108}c)^{15,460,896}$
  • Torus knots: $p_K(c) = (10^{11}c)^{299,666}$

While these exponents are large, the transition from exponential towers to polynomial bounds represents a major theoretical milestone in low-dimensional topology.

Conclusion and Future Directions

While this work provides a polynomial bound for each fixed link type, the existence of a universal polynomial $p$ applicable to all links remains an open question. Such a discovery would place the general Link Equivalence Problem in NP. Future research will likely focus on tightening these bounds to improve the practicality of topological algorithms and extending these results to more general classes of links.

Summary updated

Abstract

For each link type $K$ in the 3-sphere, we show that there is a polynomial $p_K$ such that any two diagrams of $K$ with $c_1$ and $c_2$ crossings differ by at most $p_K(c_1) + p_K(c_2)$ Reidemeister moves. As a consequence, the problem of recognising whether a given link diagram represents $K$ is in the complexity class NP and hence can be completed deterministically in exponential time. We calculate this polynomial $p_K$ explicitly for various classes of links.
BibTeX
@misc{lackenby2026polynomial,
  author = {Marc Lackenby},
  title = {{A polynomial upper bound on Reidemeister moves for each link type}},
  year = {2026},
  eprint = {2602.09923},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2602.09923},
}

Loading related content...

Figure