Abstract
We consider the irreducible representations each of dimension 2 of the necklace braid group \({\mathcal {N}}{\mathcal {B}}_n\) (\(n=2,3,4\)). We then consider the tensor product of the representations of \({\mathcal {N}}{\mathcal {B}}_n\) (\(n=2,3,4\)) and determine necessary and sufficient condition under which the constructed representations are irreducible. Finally, we determine conditions under which the irreducible representations of \({\mathcal {N}}{\mathcal {B}}_n\) (\(n=2,3,4\)) of degree 2 are unitary relative to a hermitian positive definite matrix.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Bullivant et al. [4] studied representations of the necklace braid group \({\mathcal {N}}{\mathcal {B}}_n\), especially those obtained as extensions of representations of the braid group \(B_n\) and the loop braid group \(LB_n\) (see [3, 5]). They showed that any irreducible \(B_n \) representation extends to \({\mathcal {N}}{\mathcal {B}}_n\) in a standard way. Moreover, they proved that any local representation of \(B_n\), coming from a braided vector space, can be extended to \({\mathcal {N}}{\mathcal {B}}_n\).
A link \(\mathcal {L}_n=K_0\cup K_1\cup \cdots \cup K_n\) is called a necklace if:
-
\(K_0\) is an Euclidean circle of center O and of radius 1
-
each \(K_i\) is an Euclidean circle whose center \(O_i\) belongs to \(K_0\) and radius \(r_i\) such that \(0<r_i<\frac{1}{2}\) for \(0<i\leqslant n\)
-
the plane of each \(K_i\) is the one containing the line \((OO_i)\) and perpendicular to the plane of \(K_0\).
-
If \(O_i=O_j\) then \(r_i\ne r_j\)
The motion group \({\mathcal {N}}{\mathcal {B}}_n\), the necklace braid group, as described in [1] is identified with the fundamental group of the configuration space \(\mathcal {L}_n\). The group \({\mathcal {N}}{\mathcal {B}}_n\) is generated by the elements \(\sigma _1,\ldots ,\sigma _n\) and \(\tau \) where \(\sigma _i\) is the motion, up to homotopy, of passing the i th circle through the \((i+1)\)th circle, while \(\tau \) corresponds to shifting each circle one position in the counterclockwise direction.
In Sect. 2, we consider the irreducible representations defined on the necklace braid group \({{\mathcal {N}}{\mathcal {B}}}_n\) (\(n=2,3,4\)) each of dimension 2. In Sect. 3, we construct the tensor product of representations on \({{\mathcal {N}}{\mathcal {B}}}_n\). Then, we discuss the irreducibility of the tensor product of the representations of \({\mathcal {N}}{\mathcal {B}}_n\) (\(n=2,3,4\)). Theorem 3.5 gives necessary and sufficient condition for the irreducibility of tensor product of the representations of \({\mathcal {N}}{\mathcal {B}}_4\). Theorem 3.17 provides necessary and sufficient conditions under which the tensor product of the representations of \({\mathcal {N}}{\mathcal {B}}_3\) is irreducible. Theorem 3.18 gives necessary and sufficient condition for the irreducibility of the tensor product of the representations of \({\mathcal {N}}{\mathcal {B}}_2\). In Sect. 4, we prove that the irreducible representations of dimension 2 of \({\mathcal {N}}{\mathcal {B}}_n\) are unitary relative to hermitian positive definite matrix in the case \(n=3,4\) (see Propositions 4.4, 4.5 and Proposition 4.6).
Definition 1.1
[1] The necklace braid group \({\mathcal {N}}{\mathcal {B}}_n\) is identified with the fundamental group of the configuration space of \(\mathcal {L}_n\).
The following theorem gives a presentation of the necklace braid group \({\mathcal {N}}{\mathcal {B}}_n\) by generators and relations.
Theorem 1.2
[1] The necklace braid group \({{\mathcal {N}}{\mathcal {B}}}_n\) has a presentation by generators \(\sigma _1, \ldots ,\sigma _n,\tau \) and relations as follows:
-
(B1)
\(\sigma _i \sigma _{i+1} \sigma _i = \sigma _{i+1} \sigma _i \sigma _{i+1}\) for \(1\leqslant i\leqslant n\);
-
(B2)
\(\sigma _i \sigma _j = \sigma _j \sigma _i \text { for } |i-j|\ne 1\);
-
(N1)
\(\tau \sigma _i\tau ^{-1}=\sigma _{i+1}\) for \(1\leqslant i \leqslant n~(\text {mod } n)\);
-
(N2)
\(\tau ^{2n}=1\)
Here, indices are taken modulo n, with \(\sigma _{n+1}=\sigma _1\) and \(\sigma _0=\sigma _n\). The relations (B1) and (B2) are those for the braid group \(B_n\) (see [2]).
2 Irreducible representations of dimension 2 of \({\mathcal {N}}{\mathcal {B}}_n\) (\(n=2,3,4\))
Consider the necklace braid group \({\mathcal {N}}{\mathcal {B}}_n\) for \(n=2,3,4\). We have the following proposition (see [4]).
Proposition 2.1
Any irreducible representation of dimension 2 of \({\mathcal {N}}{\mathcal {B}}_n (n=2,3,4)\) is isomorphic to the representation \(\rho =\rho (T,t,a,c,d)\) that is defined by:
with the following conditions:
n | T | Conditions |
---|---|---|
2 | \(T=-1\) | \(c=a^2-ad+d^2\), \(c\ne 0\) and \(a\ne d\) |
3 | \(T=e^{\pm i2\pi /3}\) | \(a=\omega d\), \(c\ne 0\), \(c\ne \omega d^2\) and \(\omega =e^{\pm i\pi /3}\) |
4 | \(T=-1\) | \(c=a^2-ad+d^2\), \(c\ne 0\) and \(a\ne d\) |
\(T=\pm i\) | \(c=-d^2\), \(a=d\ne 0\) |
Here, t is any \((2n)^{th}\) root of unity.
Proof
Let \((\rho ,V)\) be an irreducible representation of dimension 2 of \({\mathcal {N}}{\mathcal {B}}_n\) (for \(n=2,3,4\)). From the fact that \(\tau \) has order 2n, we may choose a basis for V such that \(\rho (\tau ) = \left( \begin{matrix} t_1&{}0\\ 0&{}t_2 \end{matrix}\right) \), where \(t_1\) and \(t_2\) are (2n)th roots of unity. Since \(\rho \) is irreducible representation, it follows that \(t_1\ne t_2\). Similarly we have that \(\rho (\sigma _1)\) is neither upper nor lower triangular, because (1, 0) or (0, 1) would generate an invariant subspace. Due to rescaling, we may assume that \(\rho (\sigma _1) = \left( \begin{matrix} a&{}1\\ c&{}d \end{matrix}\right) \). Since \(\rho (\sigma _1)\) is neither diagonal nor triangular, we have that \(c\ne 0\). To require invertibility, we assume that \(ad\ne c\). Using \(\rho (\sigma _1)=\left( \begin{matrix} a&{}1\\ c&{}d \end{matrix}\right) \) and the condition (N1): \(\tau \sigma _i\tau ^{-1}=\sigma _{i+1}\), we get
where \(T=t_1t_2^{-1}\). Note that \(T\ne 1\) because \(t_1\ne t_2\). Set \(t_2=t\) then \(t_1=Tt\) and \(\rho (\tau )\) has the form \(\left( \begin{matrix} Tt&{}0\\ 0&{}t \end{matrix}\right) \). Since \(\sigma _{n+1}=\sigma _1\), we have \(T^{n}=T^{-n}=1\). Thus, T is a primitive nth root of unity.
Now, we check the conditions (B1): \(\sigma _i\sigma _{i+1}\sigma _i=\sigma _{i+1}\sigma _i\sigma _{i+1}\). By Lemma 1.2 in [4], it is sufficient to check (B1) for \(i=1\) (i.e. \(\sigma _1\sigma _2\sigma _1=\sigma _2\sigma _1\sigma _2\) ).
By direct calculations, we get
For \(n=2\), we have that \(T=-1\) and \(c=a^2-ad+d^2\). Under the assumption that \(ad\ne c\), we have \(a\ne d\).
For \(n=3\), we have that \(T=e^{\pm 2\pi i/3}\). Then, \(a=\omega d\) where \(\omega =e^{\pm \pi i/3}\). Under the assumption that \(ad\ne c\), we have \(c\ne \omega d^2\).
For \(n=4\), we have that \(T=-1, \pm i\).
-
If \(T=-1\), then \(c=a^2-ad+d^2\) with \(a\ne d\).
-
If \(T=\pm i\), then \(c=-a^2+ad-d^2\).
It remains to verify the condition (B2): \(\sigma _i\sigma _j=\sigma _j\sigma _i\) for \(|i-j|>1\) when \(n=4\). By Lemma 1.2 in [4] we just check (B2) for \(i=1 \) and \(j=3\).
If \(T=-1\), then it is clear that \(\rho (\sigma _1)\rho (\sigma _3)=\rho (\sigma _3)\rho (\sigma _1)\).
If \(T=\pm i\), then
It follows that \(c=a^2+ad-d^2=-d^2\). \(\square \)
3 Representations of dimension 4 of \({\mathcal {N}}{\mathcal {B}}_n\) (\(n=2,3,4\))
Consider two irreducible representations \(\rho _1=\rho (T_1,t_1,a_1,c_1,d_1)\) and \(\rho _2=\rho (T_2,t_2,a_2,c_2,d_2)\) of \({\mathcal {N}}{\mathcal {B}}_n\) (\(n=2,3,4\)) each of dimension 2.
Definition 3.1
Consider the tensor product \(\rho _1\otimes \rho _2\) given by \(\rho _1\otimes \rho _2(\alpha )=\rho _1(\alpha )\otimes \rho _2(\alpha ),\) where \(\alpha \) is a generator of \({\mathcal {N}}{\mathcal {B}}_n\) (\(n=2,3,4\)). We get the following matrices for the generators \(\tau \) and \(\sigma _1\).
3.1 Representations of \({\mathcal {N}}{\mathcal {B}}_4\)
In this section, we study the irreducibility of the representation \(\rho \) of the necklace braid group \({\mathcal {N}}{\mathcal {B}}_4\). Actually, we have four cases: (3.1.1) \(T_1=T_2=-1\), (3.1.2) \(T_1=T_2=\pm i\), (3.1.3) \(T_1=-T_2=\pm i\) and (3.1.4) \(T_1=-1, T_2=\pm i\).
In what follows, suppose that \(a_1a_2d_1d_2\ne 0\).
3.1.1 Case \(T_1=T_2=-1\)
Direct computations show that the representation \(\rho \) is given by
and
The eigenvalues of \(\rho (\tau )\) are \(\lambda _1=t_1t_2\) and \(\lambda _2=-t_1t_2\). Both are of multiplicities 2. The corresponding eigenvectors are \(\alpha _1e_1+\alpha _4e_4=(\alpha _1,0,0,\alpha _4) \) for \(\lambda _1=t_1t_2\) and \(\alpha _2e_2+\alpha _3e_3=(0,\alpha _2,\alpha _3,0)\) for \(\lambda _2=-t_1t_2\), where \(\{\alpha _1,\alpha _2,\alpha _3,\alpha _4\}\subset {\mathbb {C}}\).
We now determine conditions under which the representation \(\rho \) is irreducible.
Proposition 3.2
The representation \(\rho :{{\mathcal {N}}{\mathcal {B}}}_4\rightarrow GL(4,{\mathbb {C}})\) has no non-trivial proper invariant subspaces of dimension 1 if and only if
Proof
The subspaces of dimension 1 that are invariant under \(\rho (\tau )\) are those spanned by one of the following vectors: \(e_1,~e_2,~e_3,~e_4,~e_1+ye_4~\text { and }e_2+x e_3~~\text { for }x\ne 0\) and \(y\ne 0\).
\(\rho (\sigma _1)(e_1)=\left( \begin{matrix} a_1a_2\\ a_1c_2\\ c_1a_2\\ c_1c_2 \end{matrix}\right) \not \in \langle e_1\rangle \) because, \(c_1c_2\ne 0\). So, \(\langle e_1\rangle \) is not invariant.
\(\rho (\sigma _1)(e_2)=\left( \begin{matrix} a_1\\ a_1d_2\\ c_1\\ c_1d_2 \end{matrix}\right) \not \in \langle e_2\rangle \) because, \(c_1\ne 0\). So, \(\langle e_2\rangle \) is not invariant.
\(\rho (\sigma _1)(e_3)=\left( \begin{matrix} a_2\\ c_2\\ d_1a_2\\ d_1c_2 \end{matrix}\right) \not \in \langle e_3\rangle \) because, \(c_2\ne 0\). Thus, \(\langle e_3\rangle \) is not invariant.
\(\rho (\sigma _1)(e_4)=\left( \begin{matrix} 1\\ d_2\\ d_1\\ d_1d_2 \end{matrix}\right) \not \in \langle e_4\rangle \). So, \(\langle e_4\rangle \) is not invariant.
Now consider the subspaces of the form \(\langle e_1+ye_4\rangle \) with \(y\ne 0\)
Since \(\rho (\sigma _i)(e_1+ye_4)=\left( \begin{matrix} a_1a_2+y\\ \pm a_1c_2\pm d_2y\\ \pm c_1a_2\pm d_1y\\ c_1c_2+d_1d_2y \end{matrix}\right) \), it follows that \(\langle e_1+ye_4\rangle \) is invariant under \(\rho (\sigma _i)\) for \(1\leqslant i\leqslant n.\) It follows that \(\left( \begin{matrix} a_1a_2+y\\ \pm a_1c_2\pm d_2y\\ \pm c_1a_2\pm d_1y\\ c_1c_2+d_1d_2y \end{matrix}\right) = \left( \begin{matrix} \alpha \\ 0\\ 0\\ \alpha y \end{matrix}\right) \) for some \(\alpha \in {\mathbb {C}}\setminus \{0\}\).
This is equivalent to the system:
We consider the subspaces of the form \(\langle e_2+xe_3\rangle \) with \(x\ne 0\).
\(\rho (\sigma _j)(e_2+xe_3)=\left( \begin{matrix} \pm a_1\pm a_2x\\ a_1d_2+c_2x\\ c_1+d_1a_2x\\ \pm c_1d_2\pm d_1c_2x \end{matrix}\right) \) for \(1\leqslant j\leqslant 4\).
If we assume that the subspace \(\langle e_2+xe_3\rangle \) is invariant, then \(\left( \begin{matrix} a_1+a_2x\\ a_1d_2+c_2x\\ c_1+d_1a_2x\\ c_1d_2+d_1c_2x \end{matrix}\right) = \left( \begin{matrix} 0\\ \alpha \\ \alpha x\\ 0 \end{matrix}\right) \) for some \(\alpha \in {\mathbb {C}}^*\). This is equivalent to:
\(\square \)
Proposition 3.3
The representation \(\rho :{{\mathcal {N}}{\mathcal {B}}}_4\rightarrow GL(4,{\mathbb {C}})\) has no non-trivial proper invariant subspaces of dimension 2 if and only if
Proof
The subspaces of dimension 2 that are possibly invariant are those spanned by the following subsets of vectors: \(\{e_1,he_2+je_3\},~\{e_4,re_2+se_3\},~\{e_2,ke_1+me_4\},~\{e_3,pe_1+qe_4\},~\{e_1+xe_4,e_2+ye_3\}\), where \(\{h,j,r,s,k,m,p,q,x,y\}\subset {\mathbb {C}}\).
\(\rho (\sigma _1)(e_1)=\left( \begin{matrix} a_1a_2\\ a_1c_2\\ c_1a_2\\ c_1c_2 \end{matrix}\right) \not \in \langle e_1,he_2+je_3\rangle \) because \(c_1c_2\ne 0\). So \(\langle e_1,he_2+je_3\rangle \) are not invariant subspaces for all \(h,j\in {\mathbb {C}}\).
\(\rho (\sigma _1)(e_4)=\left( \begin{matrix} 1\\ d_2\\ d_1\\ d_1d_2 \end{matrix}\right) \not \in \langle e_4,re_2+se_3\rangle \). So, \(\langle e_4,re_2+se_3\rangle \) are not invariant subspaces for all \(r,s\in {\mathbb {C}}\).
\(\rho (\sigma _1)(e_2)=\left( \begin{matrix} a_1\\ a_1d_2\\ c_1\\ c_1d_2 \end{matrix}\right) \not \in \langle e_2,ke_1+me_4\rangle \) because \(c_1\ne 0\). So, the subspaces \(\langle e_2,ke_1+me_4\rangle \) are not invariant for all \(k,m\in {\mathbb {C}}\).
\(\rho (\sigma _1)(e_3)=\left( \begin{matrix} a_2\\ c_2\\ d_1a_2\\ d_1c_2 \end{matrix}\right) \not \in \langle e_3,pe_1+qe_4\rangle \) because \(c_2\ne 0\). So, the subspaces \(\langle e_3,pe_1+qe_4\rangle \) are not invariant for all \(p,q\in {\mathbb {C}}\).
Now consider the subspaces of the form \(\langle e_1+xe_4,e_2+ye_3\rangle \) where \(x,y\in {\mathbb {C}}^*.\) Put \(u=e_1+xe_4\) and \(v=e_2+ye_3\) then
\(\rho (\sigma _j)(u)=\left( \begin{matrix} a_1a_2+x\\ \pm a_1c_2\pm d_2x\\ \pm c_1a_2\pm d_1x\\ c_1c_2+d_1d_2x \end{matrix}\right) \) and \(\rho (\sigma _j)(v)=\left( \begin{matrix} \pm a_1\pm a_2y\\ a_1d_2+c_2y\\ c_1+d_1a_2y\\ \pm c_1d_2\pm d_1c_2y \end{matrix}\right) .\)
The subspace \(\langle u,v\rangle \) is invariant if and only if \(\rho (\sigma _j)(u)=\alpha u+\alpha 'v\) and \(\rho (\sigma _j)(v)=\beta u+\beta 'v\) for some \(\alpha , \alpha ', \beta , \beta '\in {\mathbb {C}}\).
This is equivalent to \(\left\{ \begin{matrix} a_1a_2+x=\alpha \\ \pm (a_1c_2+d_2x)=\alpha '\\ \pm (c_1a_2+d_1x)=\alpha 'y\\ c_1c_2+d_1d_2x=\alpha x \end{matrix}\right. \) and \(\left\{ \begin{matrix} \pm (a_1+a_2y)=\beta \\ a_1d_2+c_2y=\beta '\\ c_1+d_1a_2y=\beta 'y\\ \pm (c_1d_2+d_1c_2y)=\beta x \end{matrix}\right. .\)
By eliminating \(\alpha ,\alpha ', \beta \) and \(\beta '\), we get the following system.
The Eqs. (3.1a) and (3.1d) lead to four solutions:
Substitute each of these solutions into the Eqs. (3.1b) and (3.1c) then, using Mathematica, we get the following relations:
\((a_2=d_2),~~~(a_1=d_1),~~~(d_1=e^{\pm \pi i/3}a_1), ~~~ (d_1d_2=a_1a_2),~~~ (d_1a_2=a_1d_2),~~~ (a_1=a_2=0),~~~(a_1=d_2=0),~~~(d_1=a_2=0),~~~(d_1=d_2=0),~~~(d_1=a_1,a_2=0)~~~(d_1=a_1,d_2=0),~~~(d_1=e^{\pm \pi i/3}a_1,a_2=0)\) or \((d_1=e^{\pm \pi i/3}a_1,d_2=0)\)
But we have \(a_1\ne d_1,~~ a_2\ne d_2,~~ a_1a_2\ne d_1d_2,~~ a_1d_2\ne a_2d_1~\) and \(a_1d_1a_2d_2\ne 0\). So, the only 2 relations left are \(d_1=e^{\pm \pi i/3}a_1\) which lead to \(c_1=a_1^2-a_1d_1+d_1^2=0\). This contradicts the fact that \(c_1\ne 0\).
Therefore \(\langle u,v\rangle \) is not invariant under \(\rho (\sigma _j)\) for \(1\leqslant j\leqslant 4\). \(\square \)
Proposition 3.4
\(\rho :{{\mathcal {N}}{\mathcal {B}}}_4\rightarrow GL(4,{\mathbb {C}})\) has no non-trivial proper invariant subspaces of dimension 3 if and only if
Proof
The subspaces of dimension 3 that are possibly invariant are those spanned by the following subsets of vectors: \(\{e_1,e_4,e_3\},~\{e_1,e_4,e_2\},~\{e_1,e_2,e_3\},~\{e_4,e_2,e_3\},~\{e_1,e_4,e_2+xe_3\},~\{e_1+ye_4,e_2,e_3\}\), where \(x,y\in {\mathbb {C}}^*\).
\(\rho (\sigma _1)(e_3)=\left( \begin{matrix} a_2\\ c_2\\ d_1a_2\\ d_1c_2 \end{matrix}\right) \not \in \langle e_1,e_4,e_3\rangle \) since \(c_2\ne 0\). Thus, \(\langle e_1,e_4,e_3\rangle \) is not invariant.
\(\rho (\sigma _1)(e_2)=\left( \begin{matrix} a_1\\ a_1d_2\\ c_1\\ c_1d_2 \end{matrix}\right) \not \in \langle e_1,e_4,e_2\rangle \) since \(c_1\ne 0\). So, \(\langle e_1,e_4,e_2\rangle \) is not invariant
\(\rho (\sigma _1)(e_1)=\left( \begin{matrix} a_1a_2\\ a_1c_2\\ c_1a_2\\ c_1c_2 \end{matrix}\right) \not \in \langle e_1,e_2,e_3\rangle \) since \(c_1c_2\ne 0\). So, \(\langle e_1,e_2,e_3\rangle \) is not invariant.
\(\rho (\sigma _1)(e_4)=\left( \begin{matrix} 1\\ d_2\\ d_1\\ d_1d_2 \end{matrix}\right) \not \in \langle e_4,e_2,e_3\rangle \). So, \(\langle e_4,e_2,e_3\rangle \) is not invariant.
Now, consider the subspace \(\langle e_1,e_4,e_2+xe_3\rangle \) with \(x\ne 0\).
Note that \(\rho (\sigma _j)(e_1)=\left( \begin{matrix} a_1a_2\\ \pm a_1c_2\\ \pm c_1a_2\\ c_1c_2 \end{matrix}\right) \), \(\rho (\sigma _j)(e_4)=\left( \begin{matrix} 1\\ \pm d_2\\ \pm d_1\\ d_1d_2 \end{matrix}\right) \) and \(\rho (\sigma _j)(e_2+xe_3)=\left( \begin{matrix} \pm a_1\pm a_2x\\ a_1d_2+c_2x\\ c_1+d_1a_2x\\ \pm c_1d_2\pm d_1c_2x \end{matrix}\right) \).
Assume that the subspace \(\langle e_1,e_4,e_2+xe_3\rangle \) is invariant. It follows that \(\{\rho (\sigma _1)(e_1),\rho (\sigma _1)(e_4), \rho (\sigma _1)(e_2+xe_3)\}\subset \langle e_1,e_4,e_2+xe_3\rangle \). This is equivalent to the following system.
We determine conditions for which the system above is consistent.
From Eqs. (3.2a) and (3.2b), we get \(x=\dfrac{c_1a_2}{a_1c_2}=\dfrac{d_1}{d_2}\) then substitute it into the Eq. (3.2c) to obtain \(c_2\dfrac{d_1^2}{d_2^2}+a_1d_1-\dfrac{d_1^2a_2}{d_2}-c_1=0.\)
\(\Longrightarrow a_2d_2c_1=a_1d_1c_2\text { and }c_2d_1^2-a_2d_2d_1^2=d_2^2c_1-d_2^2a_1d_1\)
\(\Longrightarrow a_2d_2(a_1^2-a_1d_1+d_1^2)=a_1d_1(a_2^2-a_2d_2+d_2^2) ~ \text { and } ~ d_1^2(c_2-a_2d_2)=d_2^2(c_1-a_1d_1)\)
\(\Longrightarrow a_2d_2(a_1^2-2a_1d_1+d_1^2)=a_1d_1(a_2^2-2a_2d_2+d_2^2) ~ \text { and } ~ d_1^2(a_2-d_2)^2=d_2^2(a_1-d_1)^2\)
\(\Longrightarrow \dfrac{a_2d_2}{a_1d_1}=\dfrac{(a_2-d_2)^2}{(a_1-d_1)^2}=\dfrac{d_2^2}{d_1^2}\)
\(\Longrightarrow \dfrac{a_2}{a_1}=\dfrac{d_2}{d_1}\)
\(\Longrightarrow a_2d_1=a_1d_2\).
This gives a contradiction.
Now, consider the subspace \(\langle e_1+ye_4,e_2,e_3\rangle \) with \(y\ne 0\). Note that \(\rho (\sigma _j)(e_2)=\left( \begin{matrix} \pm a_1\\ a_1d_2\\ c_1\\ \pm c_1d_2 \end{matrix}\right) \), \(\rho (\sigma _j)(e_3)=\left( \begin{matrix} \pm a_2\\ c_2\\ d_1a_2\\ \pm d_1c_2 \end{matrix}\right) \) and \(\rho (\sigma _j)(e_1+ye_4)=\left( \begin{matrix} a_1a_2+y\\ \pm a_1c_2\pm d_2y\\ \pm c_1a_2\pm d_1y\\ c_1c_2+d_1d_2y \end{matrix}\right) \).
Assume that the subspace \(\langle e_1+ye_4,e_2,e_3\rangle \) is invariant. It follows that the vectors \(\rho (\sigma _i)(e_2),~~\rho (\sigma _i)(e_3)\) and \(\rho (\sigma _i)(e_1+ye_4)\) are linear combinations of \(e_1+ye_4,~e_2,\) and \(e_3\).
\(\Longrightarrow \dfrac{c_1d_2}{a_1}=\dfrac{d_1c_2}{a_2}=y\) and \(y^2+(a_1a_2-d_1d_2)y-c_1c_2=0\)
\(\Longrightarrow a_2d_2c_1=a_1d_1c_2 ~ \text { and } ~ \dfrac{c_1^2d_2^2}{a_1^2}+a_2d_2c_1-\dfrac{c_1d_1d_2^2}{a_1}-c_1c_2=0\)
\(\Longrightarrow a_2d_2(a_1^2-a_1d_1+d_1^2)=a_1d_1(a_2^2-a_2d_2+d_2^2) ~ \text { and } ~ c_1d_2^2+a_2d_2a^2_1-d_1d_2^2a_1-c_2a_1^2=0\)
\(\Longrightarrow a_2d_2(a_1^2-2a_1d_1+d_1^2)=a_1d_1(a_2^2-2a_2d_2+d_2^2) ~ \text { and } ~ d_2^2(c_1-d_1a_1)=a_1^2(c_2-a_2d_2)\)
\(\Longrightarrow a_2d_2(a_1-d_1)^2=a_1d_1(a_2-d_2)^2 ~ \text { and } ~ d_2^2(a_1-d_1)^2=a_1^2(a_2-d_2)^2\)
\(\Longrightarrow \dfrac{a_2d_2}{a_1d_1}=\dfrac{d_2^2}{a_1^2}=\dfrac{(a_2-d_2)^2}{(a_1-d_1)^2}\)
\(\Longrightarrow \dfrac{a_2}{d_1}=\dfrac{d_2}{a_1}\).
This gives a contradiction. \(\square \)
We, thus, get our main theorem.
Theorem 3.5
The representation \(\rho :{\mathcal {N}}{\mathcal {B}}_4\rightarrow GL(4,{\mathbb {C}})\) is irreducible if and only if \(a_1a_2\ne d_1d_2\) and \(a_1d_2\ne a_2d_1\).
3.1.2 Case \(T_1=T_2=\pm i\)
In the case \(T_1=T_2=i\), the representation \(\rho \) of \({\mathcal {N}}{\mathcal {B}}_4\) is given by
Likewise, we write the matrices for \(T_1=T_2=-i.\)
Proposition 3.6
The representation \(\rho \) is reducible.
Proof
Let S be the subspace spanned by the vector \(v=d_2e_2-d_1e_3\). Note that \(e_2\) and \(e_3\) are eigen-vectors of \(\rho (\tau )\) corresponding the same eigenvalue. So, S is invariant under \(\rho (\tau )\). It is easy to show that \(\rho (\sigma _j)(v)=2d_1d_2v\) for \(j=1,2,3,4\). This completes the proof. \(\square \)
3.1.3 Case \(T_1=-T_2=\pm i\)
In the case \(T_1=-T_2=i\), the representation \(\rho \) of \({\mathcal {N}}{\mathcal {B}}_4\) is given by
Likewise, we write the matrices for \(T_1=-T_2=-i.\)
Proposition 3.7
The representation \(\rho \) is reducible.
Proof
Let S be the subspace spanned by the vector \(v=e_1+d_1d_2e_4\). The subspace S is invariant under \(\rho (\tau )\) because, \(e_1\) and \(e_4\) are eigenvectors of \(\rho (\tau )\) corresponding to the same eigenvalue. By direct calculations we get \(\rho (\sigma _j)(v)=2d_1d_2v\in S\) for \(j=1,2,3,4\). Hence, S is invariant under \(\rho (\sigma _j)\). \(\square \)
3.1.4 Case \(T_1=-1,T_2=\pm i\)
Without loss of generality, assume that \(T_1=-1\) and \(T_2=i\). Then, we have \(T_1T_2=-i\) and
Likewise, we write the matrices for \(T_1=-1, T_2=-i.\)
Note that the vectors \(e_j\) (\(j=1,2,3,4\)) are eigenvectors of \(\rho (\tau )\) corresponding to four distinct eigenvalues (each of multiplicity 1).
Proposition 3.8
The representation \(\rho \) is irreducible.
Proof
We have \(\rho (\sigma _1)(e_1)=\left( \begin{matrix} a_1d_2\\ -a_1d_2^2\\ c_1d_2\\ -c_1d_2^2 \end{matrix}\right) \), \(\rho (\sigma _1)(e_2)=\left( \begin{matrix} a_1\\ a_1d_2\\ c_1\\ c_1d_2 \end{matrix}\right) \), \(\rho (\sigma _1)(e_3)=\left( \begin{matrix} d_2\\ -d_2^2\\ d_1d_2\\ -d_1d_2^2 \end{matrix}\right) \) and \(\rho (\sigma _1)(e_4)=\left( \begin{matrix} 1\\ d_2\\ d_1\\ d_1d_2 \end{matrix}\right) \). Note that \(a_1\ne 0\), \(d_1\ne 0\), \(c_1\ne 0\) and \(d_2\ne 0\). This implies that none of the proper subspaces spanned by the vectors \(e_j\) are invariant under \(\rho (\sigma _1)\). Likewise, we show that there are no invariant subspaces of dimensions 2 and 3. Hence \(\rho \) is irreducible. \(\square \)
3.2 Representation of \({\mathcal {N}}{\mathcal {B}}_3\)
Consider two irreducible representations of dimension 2 of the necklace braid group \({\mathcal {N}}{\mathcal {B}}_3\) \(\rho _1=\rho (T_1,t_1,\omega _1,c_1,d_1)\) and \(\rho _2=\rho (T_2,t_2,\omega _2,c_2,d_2)\) (see Proposition 2.1).
Recall that \(T_1=e^{\pm i2\pi /3},T_2=e^{\pm i2\pi /3},\omega _1=e^{\pm i\pi /3}\) and \(\omega _2=e^{\pm i\pi /3}\).
Let \(\rho \) be the representation of \({\mathcal {N}}{\mathcal {B}}_3\), which is defined in Definition 3.1. We have
and
for \(j=1,2,3\). We have two cases (1) \(T_1=e^{2\pi i/3},T_2=e^{-2\pi i/3}\) and (2) \(T_1=T_2=e^{\pm 2\pi i/3}\).
We determine a necessary and sufficient condition under which \(\rho \) is an irreducible representation of \({\mathcal {N}}{\mathcal {B}}_3\) of dimension 4.
In what follows, suppose that \(d_1d_2\ne 0\).
3.2.1 Case \(T_1=e^{2\pi i/3}, T_2=e^{-2\pi i/3}\)
In this case, we have
Then, the eigenvalues of \(\rho (\tau )\) are \(\lambda _1=t_1t_2\) (of multiplicity 2) and \(\lambda _2=e^{2\pi i/3}t_1t_2\) and \(\lambda _3=e^{-2\pi i/3}t_1t_2\). The corresponding eigenvectors are \(x_1e_1+x_4e_4\) for \(\lambda _1\) and \(e_2\) and \(e_3\) for \(\lambda _2\) and \(\lambda _3\), respectively, where \(\{x_1,x_4\}\subset {\mathbb {C}}^*\).
Lemma 3.9
If \(T_1=e^{2\pi i/3}\), \(T_2=e^{-2\pi i/3},\omega _1=e^{\pm i\pi /3}\) and \(\omega _2=e^{\pm i\pi /3}\), then the representation \(\rho \) of \({\mathcal {N}}{\mathcal {B}}_3\) has no invariant proper subspaces of dimension 1 if and only if
Proof
By direct calculations, we check that each of the subspaces spanned by the vectors \(e_j\) (\(j=1,2,3\)) are not invariant under \(\rho (\sigma _1)\).
It remains to consider the subspace S spanned by the vector \(v=e_1+xe_4\) with \(x\in {\mathbb {C}}^*\). Since \(T_1T_2=1\), it follows that
So
Therefore the subspace S, that is spanned by \(v=e_1-\frac{\omega _1d_1c_2}{d_2}e_4\), is invariant under \(\rho (\sigma _j)\) if \(\omega _1\omega _2=1\) and \(\omega _1d_1^2c_2= \omega _2d_2^2c_1\). This gives a contradiction. \(\square \)
Lemma 3.10
If \(T_1=e^{2\pi i/3}\), \(T_2=e^{-2\pi i/3},\omega _1=e^{\pm i\pi /3}\) and \(\omega _2=e^{\pm i\pi /3}\) then the representation \(\rho \) of \({\mathcal {N}}{\mathcal {B}}_3\) has no invariant proper subspaces of dimension 2.
Proof
The possible two-dimensional invariant subspaces candidates to study are the following:
-
1.
\(\langle e_i,e_j\rangle \) for \(i\ne j\)
-
2.
\(\langle e_2,e_1+xe_4\rangle \) for \(x\ne 0\)
-
3.
\(\langle e_3,e_1+ye_4\rangle \) for \(y\ne 0\)
Since \(c_1\ne 0, c_2\ne 0\), \(d_1\ne 0\) and \(d_2\ne 0\), it follows that none of the subspaces mentioned above is invariant under \(\rho (\sigma _1)\). \(\square \)
Lemma 3.11
If \(T_1=e^{2\pi i/3}\), \(T_2=e^{-2\pi i/3}, \omega _1=e^{\pm i\pi /3}\) and \(\omega _2=e^{\pm i\pi /3}\) then the representation \(\rho \) of \({\mathcal {N}}{\mathcal {B}}_3\) has no invariant proper subspaces of dimension 3 if and only if
Proof
The possible invariant three-dimensional invariant subspaces to study are:
-
1.
\(\langle e_i,e_j,e_k\rangle \) where i, j, k are pairwise distinct
-
2.
\(\langle e_2,e_3,e_1+xe_4\rangle \) for \(x\ne 0\)
By direct computations, it is easy to check that all the subspaces in case 1 are not invariant.
In case 2, we let S be a subspace of the form \(\langle e_2,e_3,e_1+xe_4\rangle \) with \(x\ne 0\).
and
Therefore
and
This is equivalent to
Therefore, S is invariant under \(\rho (\sigma _j)\) (\(j=1,2,3\)) if and only if
and
The Eqs. (3.3) and (3.4) imply that
and
Since \(\omega _2d_2^2\ne c_2\), it follows that
Therefore the subspace S, which is spanned by \(v=e_1-\frac{c_1d_2}{\omega _1d_1}e_4\), is invariant under \(\rho (\sigma _j)\) if \(\omega _1\omega _2=1\) and \(\omega _1d_1^2c_2= \omega _2d_2^2c_1\). This gives a contradiction. \(\square \)
Now, we have the following proposition.
Proposition 3.12
If \(T_1=e^{i2\pi /3}\), \(T_2=e^{-i2\pi /3}, \omega _1=e^{\pm i\pi /3}\) and \(\omega _2=e^{\pm i\pi /3}\) then \(\rho \) is irreducible representation of \({\mathcal {N}}{\mathcal {B}}_3\) if and only if \(\omega _1\omega _2\ne 1\) or \(\omega _2d_2^2c_1\ne \omega _1d_1^2c_2\).
3.2.2 Case (\(T_1=T_2=e^{2\pi i/3}\)) and (\(T_1=T_2=e^{-2\pi i/3}\))
In this case, we have \(T_1=T_2\). Then,
and
for \(j=1,2,3\). The eigenvalues of \(\rho (\tau )\) are then \(\lambda _2=t_1t_2e^{\pm i2\pi /3}\) (of multiplicity 2) and \(\lambda _1=e^{\mp i2\pi /3}t_1t_2\) and \(\lambda _3=t_1t_2\). The corresponding eigenvectors are \(xe_2+ye_3\) for \(\lambda _2\) and \(e_1\) and \(e_4\) for \(\lambda _1\) and \(\lambda _3\), respectively.
Lemma 3.13
If \(T_1=T_2=e^{\pm 2\pi i/3},\omega _1=e^{\pm i\pi /3}\) and \(\omega _2=e^{\pm i\pi /3}\) then the representation \(\rho \) of \({\mathcal {N}}{\mathcal {B}}_3\) has no invariant proper subspaces of dimension 1 if and only if \(\omega _1\ne \omega _2\) or \(\omega _1d_1^2c_2\ne \omega _2d_2^2c_1\).
Proof
By direct calculations, we check that each of the subspaces spanned by the vectors \(e_j\) are not invariant under \(\rho (\sigma _1)\).
It remains to consider the subspace S spanned by the vector \(v=e_2+xe_3\) with \(x\in {\mathbb {C}}^*\). Since \(T_1=T_2\), it follows that
So
Therefore the subspace S, which is spanned by \(v=e_2-\frac{\omega _1d_1}{\omega _2d_2}e_3\), is invariant under \(\rho (\sigma _j)\) if \(\omega _1=\omega _2\) and \(\omega _1d_1^2c_2= \omega _2d_2^2c_1\). This gives a contradiction. \(\square \)
Lemma 3.14
If \(T_1=T_2=e^{\pm 2\pi i/3}\) and \(\omega _1=e^{\pm i\pi /3}\) and \(\omega _2=e^{\pm i\pi /3}\), then the representation \(\rho \) of \({\mathcal {N}}{\mathcal {B}}_3\) has no invariant proper subspaces of dimension 2.
Proof
The possible two-dimensional invariant subspaces candidates to study are the following:
-
1.
\(\langle e_i,e_j\rangle \) for \(i\ne j\)
-
2.
\(\langle e_1,e_2+xe_3\rangle \) for \(x\ne 0\)
-
3.
\(\langle e_4,e_2+ye_3\rangle \) for \(y\ne 0\)
Since \(c_1\ne 0, c_2\ne 0\), \(d_1\ne 0\) and \(d_2\ne 0\), it follows that none of the subspaces mentioned above is invariant under \(\rho (\sigma _1).\) \(\square \)
Lemma 3.15
If \(T_1=T_2=e^{\pm 2\pi i/3}\), \(\omega _1=e^{\pm i\pi /3}\) and \(\omega _2=e^{\pm i\pi /3}\), then the representation \(\rho \) of \({\mathcal {N}}{\mathcal {B}}_3\) has no invariant proper subspaces of dimension 3 if and only if \(\omega _1\ne \omega _2\) or \(\omega _1d_1^2c_2\ne \omega _2d_2^2c_1\).
Proof
The possible invariant three-dimensional invariant subspaces to study are:
-
1.
\(\langle e_i,e_j,e_k\rangle \) where i, j, k are pairwise distinct
-
2.
\(\langle e_1,e_4,e_2+xe_3\rangle \) for \(x\ne 0\)
By direct computations, it is easy to check that all the subspaces in case 1 are not invariant under \(\rho (\sigma _1)\).
In case 2, we let S be a subspace of the form \(\langle e_1,e_4,e_2+xe_3\rangle \) with \(x\ne 0\).
and
Therefore, we get
and
This is equivalent to
Therefore, S is invariant under \(\rho (\sigma _j)\) (\(j=1,2,3\)) if and only if:
and
The Eqs. (3.5) and (3.6) imply that
and
Since \(\omega _1d_1^2\ne c_1\), it follows that \(\omega _1=\omega _2\), which gives a contradiction. Therefore, the subspace S, that is spanned by \(v=e_2-\frac{d_1}{d_2}e_3\), is invariant under \(\rho (\sigma _j)\) if \(\omega _1=\omega _2\) and \(\omega _1d_1^2c_2= \omega _2d_2^2c_1\). \(\square \)
Proposition 3.16
If \(T_1=T_2=e^{\pm 2\pi i/3}\), \(\omega _1=e^{\pm i\pi /3}\) and \(\omega _2=e^{\pm i\pi /3}\) then \(\rho \) is an irreducible representation of \({\mathcal {N}}{\mathcal {B}}_3\) if and only if \(\omega _1\ne \omega _2\) or \(\omega _2d_2^2c_1\ne \omega _1d_1^2c_2\).
Proof
The proof follows directly from Lemmas 3.13, 3.14 and 3.15. \(\square \)
Theorem 3.17
The representation \(\rho \) of \({\mathcal {N}}{\mathcal {B}}_3\) is irreducible if and only if one of the following conditions hold.
-
1.
\(T_1=e^{2\pi i/3}, T_2=e^{-2\pi i/3}\) and \(\omega _1\omega _2\ne 1\)
-
2.
\(T_1=e^{2\pi i/3}, T_2=e^{-2\pi i/3}\) and \(\omega _1d_1^2c_2\ne \omega _2d_2^2c_1\)
-
3.
\(T_1=T_2=e^{\pm 2\pi i/3}\) and \(\omega _1\ne \omega _2\)
-
4.
\(T_1=T_2=e^{\pm 2\pi i/3}\) and \(\omega _1d_1^2c_2\ne \omega _2d_2^2c_1\)
Proof
The proof follows directly from Propositions 3.12 and 3.16\(\square \)
3.3 Representation of \({\mathcal {N}}{\mathcal {B}}_2\)
In this section, we consider the representation \(\rho \) given by Definition 3.1. In this case, we have \(T_1=T_2=-1\). A similar work to that done for \({\mathcal {N}}{\mathcal {B}}_4\) (Theorem 3.5), we get the following theorem.
Theorem 3.18
The representation \(\rho \) is irreducible if and only if \(a_1a_2\ne d_1d_2\) and \(a_1d_2\ne a_2d_1\).
4 Unitary representations of \({\mathcal {N}}{\mathcal {B}}_n\) of dimension 2 (\(n=2,3,4\))
Definition 4.1
A square matrix A is unitary relative to a matrix M if \(AMA^*=M\). Here, \(A^*\) is the conjugate transpose of A.
Definition 4.2
A representation \(\rho :G\rightarrow GL(n,{\mathbb {C}})\) of a group G is called unitary if \(\rho (g)\) is unitary for every element \(g\in G\).
We consider the irreducible representations of \({\mathcal {N}}{\mathcal {B}}_n\), given by Proposition 2.1. Then, we determine a necessary and sufficient condition under which these representations are unitary relative to some hermitian positive definite matrix.
Recall that
where t is a (2n)th root of unity and \(T\in \{-1,i,-i,e^{i2\pi /3},e^{-i2\pi /3}\}\).
Lemma 4.3
Let \(\rho \) be a complex irreducible representation of \({\mathcal {N}}{\mathcal {B}}_n\) (\(n=2,3,4\)), given by Proposition 2.1. If \(\rho \) is unitary relative to a matrix \(M=\left( \begin{matrix} x&{}y\\ z&{}u \end{matrix}\right) \), then \(y=z=0\) and
Proof
Since \(\rho \) is unitary relative to a matrix \(M=\left( \begin{matrix} x&{}y\\ z&{}u \end{matrix}\right) \), we have \(\rho (\tau )M\rho (\tau )^*=M\). Thus, we get
Then, we get
It follows that
Therefore, \(y=z=0\) because \(T\ne 1\).
We also have \(\rho (\sigma _j)M\rho (\sigma _j)^*=M\) (\(j=1,2,3,4\)).
Then, we get
So we have
Since \({\bar{T}}=T^{-1}\), it follows that
Then we get
\(\square \)
Proposition 4.4
Let \(\rho :{{\mathcal {N}}{\mathcal {B}}}_4\rightarrow GL(2,{\mathbb {C}})\) be the irreducible representation of \({{\mathcal {N}}{\mathcal {B}}}_4\) given in Proposition 2.1\((T=-1)\). Then, \(\rho \) is unitary relative to a hermitian positive definite matrix M if and only if
and
where \(\theta ,\delta ',\delta \in {\mathbb {R}}\)
Proof
Recall that
where t is an eighth root of unity and \(c=a^2-ad+d^2\) such that \(a\ne d\) and \(c\ne 0\). Suppose that \(\rho \) is unitary relative to matrix M. By Lemma 4.3, we have \(M=\left( \begin{matrix} x&{}0\\ 0&{}u \end{matrix}\right) \) and
To have a non-trivial solution of the first two equations of the system above, the determinant of the system is zero and so
The first equation yields
From the third equation of the last system, we conclude that \(ad\ne 0\) and
The Eq. (4.2) and the Eq. (4.3) imply that
Using the fact \(c=a^2-ad+d^2\) and the Eqs. (4.1) and (4.4), we get
Thus, we get
Hence, there exists \(\theta \in {\mathbb {R}}\) such that
Therefore we get
Set \(d=\delta '+i\delta \) where \(\delta ',\delta \in {\mathbb {R}}\). Then
To solve the Eq. (4.5), consider two cases: (1) \(1+e^{2i\theta }=0\), (2) \(1+e^{2i\theta }\ne 0\).
If \(1+e^{2i\theta }=0\) then \(d=\delta '\mp \dfrac{i}{2}\) and \(a=\delta '\pm \dfrac{i}{2}\).
If \(1+e^{2i\theta }\ne 0\) then \(d=-\dfrac{1}{2}\sec \theta -(\tan \theta )\delta +\delta i\).
Therefore, we have
where \(\theta ,\delta ',\delta \in {\mathbb {R}}\).
The condition \(|a|<1\) follows directly from the Eq. (4.2) and the fact that M is positive definite. The matrix M becomes \(\begin{pmatrix} 1&{}0\\ 0&{}1-|a|^2 \end{pmatrix}\) up to a constant. \(\square \)
Proposition 4.5
Let \(\rho :{{\mathcal {N}}{\mathcal {B}}}_4\rightarrow GL(2,{\mathbb {C}})\) be the irreducible representation of \({{\mathcal {N}}{\mathcal {B}}}_4\) given by Proposition 2.1\((T=\pm i)\). The representation \(\rho \) is unitary relative to a hermitian positive definite matrix M if and only if
Proof
Recall that
where t is an eighth root of unity and \(d\ne 0\).
Suppose that \(\rho \) is unitary relative to hermitian positive definite matrix M. Then, by Lemma 4.3, we have \(M=\left( \begin{matrix} x&{}0\\ 0&{}u \end{matrix}\right) \) and
Since \(a=d\) and \(c=-d^2\), it follows that
To have a non-trivial solution of the first two equations of the last system, the determinant of the system is zero and so
The first and third equations of the last system imply that
Then, we get
and the matrix M becomes \(\begin{pmatrix} 2&{}0\\ 0&{}1 \end{pmatrix}\) up to a constant. \(\square \)
Proposition 4.6
Let \(\rho :{{\mathcal {N}}{\mathcal {B}}}_3\rightarrow GL(2,{\mathbb {C}})\) be the irreducible representation of \({{\mathcal {N}}{\mathcal {B}}}_4\) that is defined, in Proposition 2.1\((T=e^{\pm 2\pi i/3})\). The representation \(\rho \) is unitary relative to a hermitian positive definite matrix M if and only if \(|d|<1\) and
Proof
Recall that
where t is a sixth root of unity and \(\omega =e^{\pm i\pi /3}\) such that \(c\ne 0\).
Suppose that \(\rho \) is unitary relative to a matrix M. Then, by Lemma 4.3, we have \(M=\left( \begin{matrix} x&{}0\\ 0&{}u \end{matrix}\right) \) and
Since \(a=\omega d\), it follows that
So
Since \(\omega ^{-1}={\bar{\omega }}\), it follows that
The condition \(|d|<1\) follows from the Eq. (4.6) and the fact that M is a positive definite. The matrix M becomes \(\left( \begin{matrix} 1&{}0\\ 0&{}1-|d|^2 \end{matrix}\right) \) up to a constant. \(\square \)
References
Bellingeri, P.; Bodin, A.: The braid group of a necklace. Math. Z. 283(3–4), 995–1010 (2016)
Birman, J.S.: Braids, Links and Mapping Class Groups. Annals of Mathematical Studies, vol. 82. Princeton University Press, Princeton (1975)
Bruillard, P.; Chang, L.; Hong, S.-M.; Plavnik, J.; Rowell, E.; Sun, M.: Low-dimensional representations of the three component loop braid group. J. Math. Phys. 56(11), 111707 (2015)
Bullivant, A.; Kimball, A.; Martin, P.; Rowell, E.: Representations of the necklace braid group: topological and combinatorial approaches. Commun. Math. Phys. 375(2), 1223–1247 (2020)
Kadar, Z.; Martin, P.; Rowell, E.; Wang, Z.: Local representations of the loop braid group. Glasg. Math. J. 59(2), 359–378 (2017)
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Mayassi, T.I., Abdulrahim, M.N. Representations of the necklace braid group \({{\mathcal {N}}{\mathcal {B}}}_n\) of dimension 4 (\(n=2,3,4\)). Arab. J. Math. 10, 423–441 (2021). https://doi.org/10.1007/s40065-021-00325-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40065-021-00325-1