Article Search
eISSN 0454-8124
pISSN 1225-6951

Article

Kyungpook Mathematical Journal 2021; 61(1): 205-212

Published online March 31, 2021

Forbidden Detour Number on Virtual Knot

Shun Yoshiike, Kazuhiro Ichihara*

Nihon University Buzan Junior & Senior High School, 5-40-10 Otsuka, Bunkyo-ku, Tokyo 112-0012, Japan
e-mail : s6115m15@math.chs.nihon-u.ac.jp

College of Humanities and Sciences, Nihon University, 3-25-40 Sakurajosui, Setagaya-ku, Tokyo 156-8550, Japan
e-mail : ichihara.kazuhiro@nihon-u.ac.jp

Received: August 30, 2019; Revised: June 2, 2020; Accepted: June 4, 2020

We show that the forbidden detour move, essentially introduced by Kanenobu and Nelson, is an unknotting operation for virtual knots. Then we define the forbidden detour number of a virtual knot to be the minimal number of forbidden detour moves necessary to transform a diagram of the virtual knot into the trivial knot diagram. Some upper and lower bounds on the forbidden detour number are given in terms of the minimal number of real crossings or the coefficients of the affine index polynomial of the virtual knot.

Keywords: virtual knot, forbidden move, forbidden number.

As a generalization of (classical) knots in 3-space, Kauffman introduced virtual knots in [4]. Since then various studies on virtual knots have been done. For example, relations of virtual knots and Gauss diagrams were studied by Goussarov, Polyak, and Viro in [2]. In their research [2], a kind of local move on virtual knots was introduced, where they call it the forbidden move. See Figure 1 (left). Then it was shown by Kanenobu [3] and Nelson [8] independently that the forbidden move is an unknotting operation for virtual knots. That is, for any diagram D of a virtual knot, there exists a finite sequence of Reidemeister moves, virtual Reidemeister moves and forbidden moves that takes D to the trivial knot diagram.

Figure 1. Forbidden moves F and forbidden detour move Fd

In the studies of forbidden moves in [3], Kanenobu introduced and used several moves for virtual knot diagrams. Two of them, called F2-move and F'2-move, which are essentially equivalent, played a key role in his arguments. Actually, they were also considered and used by Nelson in [8]. Later, the F'2-move is treated by Crans, Ganzell, and Mellor in [1], where they call it the forbidden detour move. See Figure 1 (right). In this paper, we use the notation Fd for the forbidden detour move, which was denoted by FD in [1].

In this paper, we study this move, and obtain the following.

Theorem 1.1.

Let D be a virtual knot diagram of a virtual knot. Then, D can be transformed to the trivial knot diagram by using Reidemeister moves, virtual Reidemeister moves, and forbidden detour moves. Moreover, if D has c real crossings, then the number of the forbidden detour moves is at most (c1)(2c2+11c3)/24 if c is odd and c(2c2+9c14)/24 if c is even.

Remark 1.2.

We note that the F2-move in [3] (depicted in Figure 2), which is equivalent to the forbidden detour move, can be regarded as a variation of the delta move on (classical) knots, which was introduced by Matveev in [6] and by Murakami and Nakanishi in [7], independently. They showed that the delta move is an unknotting operation for classical knots, but it is known that it is not an unknotting operation for virtual knots. See [10, Theorem 1.6] for example.

Figure 2. F2 move

In virtue of the theorem above, we can introduce the following notion.

Definition 1.3.

Let K be a virtual knot. The forbidden detour number Fd(K) of K is defined as the minimal number of forbidden detour moves necessary to transform a virtual knot diagram of K into the trivial knot diagram.

We next consider lower bounds on the forbidden detour numbers of virtual knots. To obtain lower bounds, the variation of an invariant, called the affine index polynomial, under a forbidden detour move, plays a key role. In fact, we have the following.

Theorem 1.4.

Let K be a virtual knot, and PK denote the affine index polynomial of K. Suppose that PK is expressed as (t1) nantn. Then, the following holds.

Fd(K) n|an|2

In the following, our terminology about virtual knot and Gauss diagram follows from those in [1].

2. Forbidden Detour Number

A virtual knot is defined as an equivalent class of virtual knot diagrams under the (classical) Reidemeister moves and the virtual Reidemeister moves. Also, virtual knots correspond bijectively to the equivalence classes of Gauss diagrams under the moves corresponding to Reidemeister moves. That is, Reidemeister moves can modify the virtual knot diagrams, but do not change the virtual knot represented by the diagrams.

On the other hand, forbidden moves and forbidden detour moves can change virtual knots by modifying Gauss diagrams. In fact, as claimed in [1, Section 2], the forbidden detour move gives the effect on Gauss diagrams of switching the head of one arrow with the tail of an adjacent arrow. See Figure 3.

Figure 3. The effect of an Fd-move on Gauss diagrams

In the following, we call the move on Gauss diagrams corresponding to a forbidden detour move also a forbidden detour move on Gauss diagrams.

Proof of Theorem 1.1. Let D be a virtual knot diagram with c real crossings of a virtual knot, and G the Gauss diagram associated to D. We consider an arrow A of G, and assume that a arrow-heads and b arrow-tails exist on one of the arcs, say α, bounded by the endpoints of A. Then we can suppose that a+b is smaller than or equal to c-1.

Let us remove A by using forbidden detour moves as follows. We first consider the first arrow head encountered as we travel from the tail of A to its head on the arc α. For the first such arrowhead, we need at most b+1 forbidden detour moves to move it outside of α. Similarly for the next arrow head and hence each of the arrow heads, we need at most b+ 1 moves to sweep them to outside of α. We repeat this procedure a times until no arrow-heads exist on α. See Figure 4. Now we can remove A by using forbidden detour moves b times, and a single R1-move. It follows that the number of forbidden detour moves to remove A is at most a(b+1)+b = a+b+ab.

Figure 4. Sequences of forbidden detour moves (Fd-moves)

Here, since a0 and b0, we see that 2aba+bc1, and so 4ab(c1)2. Then, we get

a+b+ab(c1)+(c1)24.

Let ac=(c1)24. When n=2l with some l, we have the following.

c=1nac=s=1l(a 2s1+a 2s)  =s=1l (2s2)24+ (2s1)24  =s=1l(2s23s+1)  =16l(4l+1)(l1).

Then, since l=n/2, we get the following.

c=1n(c1)+ (c1)24=124n(2n2+9n14).

On the other hand, when n=2l1 with some l, we have the following.

c=1nac=c=12laca2l  =16l(4l+1)(l1)(l2l)  =16l(l1)(4l5).

Then, since l=(n+1)/2, we have following.

c=1n{(c1)+ (c1)24}=124(n+1)(n1)(2n3)+12n(n+1)n=124(n1)(2n2+11n3).

Consequently, D can be transformed to the trivial knot diagram by using Reidemeister moves, virtual Reidemeister moves, forbidden detour moves, and if D has c real crossings, the number of necessary forbidden detour moves is at most (c1)(2c2+11c3)/24 if c is odd and c(2c2+9c14)/24 if c is even.

3. Lower Bound of Forbidden Detour Number

In this section, we consider the lower bound for the forbidden detour number of a virtual knot. Our argument bases on the following result of Sakurai given in [9] for the forbidden move.

Let K and K' be two virtual knots which can be transformed into each other by a single forbidden move.

Then

PKP K=(t1)(±tl±tm)

holds for some integers l and m. Here PK denotes affine index polynomial, which we will define later. By imitating the argument in [9], we have the following.

Theorem 3.1.

Let K and K' be two virtual knots which can be transformed into each other by a single forbidden detour move. Then we have

PKP K=(t1)(±tltm)

for some integers ℓ and m, where PK denotes affine index polynomial.

To prove this, we recall some definitions about the affine index polynomial used in [9].

First, we define virtual knot invariants by indexes of arrows for a Gauss diagram. Let G be a Gauss diagram of a virtual knot K, and γ=PQ an arrow oriented from P to Q with sign ε(γ) in G. We give the signs to the endpoints P and Q, denoted by ε(P) and ε(Q), respectively, such that ε(P)=ε(γ) and ε(Q)=ε(γ).

For an arrow γ=PQ in a Gauss diagram G, the specified arc of γ is the arc β in the outer circle S1 with endpoints P and Q oriented from P to Q with respect to the orientation of S1.

The index of γ is the sum of the signs of all the endpoints of arrows on β other than P and Q, and denoted by i(γ).

Then the n-writhe Jn(K) of a virtual knot K is defined as

Jn(K)= i(γ)=nε(γ)

and, we define the affine index polynomial PK of K as

PK= nJn(K)(tn1).

We remark that this is different from the original definition by Kauffman in [5]. However Sakurai showed in [9, Proposition 3.2] that this gives an alternative definition of the affine index polynomial.

Proof of Theorem 3.1. Suppose that virtual knots K and K' are represented by Gauss diagrams G and G' respectively, and that G is transformed into G' by a single forbidden detour move. Then there are two cases, Cases (I) and (II) of Figure 5, for the orientations of the arcs in G and G', where the forbidden detour move is applied. We here only consider Case (I) since the other case can be treated similarly.

Figure 5. .

Let γi and γi (i = 1, 2) be the two arrows of G and G', respectively, where the forbidden detour move is applied. For arrows γi and γi, by Figure 5, we have

i(γ1)=i(γ1)+ε(γ2),  i(γ 2)=i(γ2)ε(γ1),

and ε(γi)=ε(γ i) for i = 1, 2.

Also note that this moves preserves the indexes and the signs of all the other arrows.

Therefore, we obtain the following.

PKP K=ε(γ1)(ti(γ1)1)+ε(γ2)(ti(γ2)1)  ε(γ 1)(ti(γ 1)1)ε(γ 2)(ti(γ 2)1)  =ε(γ1)ti(γ1)+ε(γ2)ti(γ2)ε(γ 1)ti(γ 1)ε(γ 2)ti(γ 2)  ε(γ1)ε(γ2)+ε(γ 1)+ε(γ 2)  =ε(γ1)ti(γ1)+ε(γ2)ti(γ2)ε(γ1)ti(γ1)ε(γ2)ε(γ2)ti(γ2)+ε(γ1)  =ε(γ1)ti(γ1)(1tε(γ2))+ε(γ2)ti(γ2)(1tε(γ1))  =(t1)(ti(γ1 )1ti(γ2 )) ifε(γi )=1, (t1)(ti(γ1 )+ti(γ2 )) ifε(γ1 )=1,ε(γ2 )=1, (t1)(ti(γ1 )1+ti(γ2 )1) ifε(γ1 )=1,ε(γ2 )=1, (t1)(ti(γ1 )ti(γ2 )1) ifε(γi )=1.

It concludes that

PKP K=(t1)(±tltm)

holds for some integers ℓ and m.

Proof of Theorem 1.4. Let K be a virtual knot with a virtual knot diagram D which can be transformed into the trivial knot diagram 0 by using forbidden detour moves s times. That is, we suppose that there exists a sequence of virtual knot diagrams D0,D1,,Ds such that D=D0, Di is obtained from Di1 by a single forbidden detour move together with Reidemeister moves (1is), and Ds = O. We denote by Ki the virtual knot represented by Di (1is). Suppose that the affine index polynomial of Ki is expressed as Pt(Ki)=(t1) nanitn for 1is.

Then, for each i with 1is, by Theorem 3.1, there exist some integers li,mi such that

Pt(Ki)=Pt(Ki1)+(t1)(±tlitmi)  =(t1) nani1tn+(t1)(±tlitmi)  =(t1)(+(alii1±1)tli++(amii11)tmi+).

It follows that the coefficients ali, ali1, ami, ami1 satisfy the the following conditions.

|alii|=|alii1±1||alii1|1,  |amii|=|amii1±1||amii1|1.

Thus, for 1is, we obtain

n|ani|n|ani1|2.

It concludes that

s  n|an0|2,

which completes the proof.

1. A. S. Crans, B. Mellor, and S. Ganzell. The forbidden number of a knot, Kyungpook Math. J. 55(2) (2015), 485-506.
2. M. Goussarov, M. Polyak, and O. Viro. Finite-type invariants of classical and virtual knots, Topology 39(5) (2000), 1045-1068.
3. T. Kanenobu. Forbidden moves unknot a virtual knot, J. Knot Theory Ramifications 10(1) (2001), 89-96.
4. L. H. Kauﬀman. Virtual knot theory, European J. Combin. 20(7) (1999), 663-690.
5. L. H. Kauﬀman. An affine index polynomial invariant of virtual knots, J. Knot Theory Ramifications 22(4) (2013), 1340007. 30 pp.
6. S. V. Matveev. Generalized surgeries of three-dimensional manifolds and representations of homology spheres, Mat. Zametki 42(2) (1987), 268-278.
7. H. Murakami and Y. Nakanishi. On a certain move generating link-homology, Math. Ann. 284(1) (1989), 75-89.
8. S. Nelson. Unknotting virtual knots with Gauss diagram forbidden moves, J. Knot Theory Ramifications 10(6) (2001), 931-935.
9. M. Sakurai. An affine index polynomial invariant and the forbidden move of virtual knots, J. Knot Theory Ramifications 25(7)(2016), 1650040. 13 pp.
10. S. Satoh and K. Taniguchi. The writhes of a virtual knot, Fund. Math. 225(1) (2014), 327-342.