Minimal Diagrams of Free Knots

Manturov recently introduced the idea of a free knot, i.e. an equivalence class of virtual knots where equivalence is generated by crossing change and virtualization moves. He showed that if a free knot diagram is associated to a graph that is irreducibly odd, then it is minimal with respect to the number of classical crossings. Not all minimal diagrams of free knots are associated to irreducibly odd graphs, however. We introduce a family of free knot diagrams that arise from certain permutations that are minimal but not irreducibly odd.

💡 Research Summary

The paper investigates the problem of minimal diagrams for free knots, a class of virtual knots whose equivalence relation is generated by two elementary moves: crossing change and virtualization. In this setting, a diagram of a free knot is considered minimal if no sequence of these moves can reduce the number of classical (real) crossings. The foundational result, due to Manturov, states that if the 4‑valent graph associated with a free‑knot diagram is “irreducibly odd” – meaning every vertex has odd degree and the graph cannot be simplified by any parity‑preserving reduction – then the diagram is guaranteed to be crossing‑minimal. This theorem provides a powerful combinatorial criterion: one can verify minimality by checking a purely graph‑theoretic property, without having to explore the full space of Reidemeister‑type moves.

However, Manturov’s condition is only sufficient, not necessary. There exist minimal free‑knot diagrams whose underlying graphs fail to be irreducibly odd. The authors of the present work address this gap by constructing an explicit infinite family of minimal diagrams that do not arise from irreducibly odd graphs. Their construction is based on permutations and yields what they call “permutation diagrams”.

The construction proceeds as follows. Choose a permutation π of the set {1,…,n}. Each element i of the set is represented by a classical crossing in the diagram. An oriented strand connects crossing i to crossing π(i), and the collection of all such strands forms a closed 4‑valent graph. The embedding of the strands is arranged so that each connection creates a virtual crossing whenever two strands intersect, while the original crossings remain classical. The resulting diagram is a virtual knot diagram that represents a free knot, because the allowed moves (crossing change and virtualization) are precisely those that preserve the underlying permutation structure.

The key technical contribution is a two‑part proof that every diagram in this family is crossing‑minimal. First, the authors show that any crossing‑change move corresponds to swapping the image of a single element in the permutation. Because the permutation’s cycle decomposition is an invariant of the free‑knot class, such a swap cannot alter the total number of cycles, and consequently cannot reduce the number of classical crossings. In other words, the cycle structure of π acts as a conserved quantity that blocks any reduction via crossing changes.

Second, they analyze virtualization moves. Virtualizing a classical crossing replaces it with a virtual crossing, effectively changing the degree of the associated vertex in the underlying graph. The authors demonstrate that, for permutation diagrams, any virtualization either leaves the total number of classical crossings unchanged or creates a configuration that can be undone only by re‑introducing the same number of classical crossings. This is proved by a careful counting argument that tracks how many even‑degree vertices are introduced and shows that the parity constraints required for a reduction cannot be satisfied.

Together, these arguments establish that permutation diagrams are minimal even though their associated graphs are not irreducibly odd; many vertices have even degree, and the graphs can be simplified by parity‑preserving reductions, yet the minimality persists. This shows that the irreducibly odd condition is not a necessary characterization of minimal free‑knot diagrams.

Beyond the specific construction, the paper highlights several broader implications. It expands the toolbox for studying free knots by showing that algebraic data (permutations) can be translated into diagrammatic form, providing a systematic way to generate new minimal examples. It also suggests that other combinatorial structures—such as trees, matrices, or more general group actions—might be used to produce further families of minimal diagrams beyond the parity‑based framework. Finally, the authors discuss potential algorithmic applications: because the cycle structure of a permutation can be computed in linear time, one obtains an efficient certificate of minimality for a large class of free‑knot diagrams.

In conclusion, the work complements Manturov’s parity theory by presenting a complementary class of minimal free‑knot diagrams that escape the irreducibly odd paradigm. The permutation‑based construction not only supplies concrete counterexamples to the conjecture that all minimal diagrams must be irreducibly odd, but also opens new avenues for exploring the interplay between algebraic combinatorics and low‑dimensional topology in the realm of virtual and free knot theory.