Journal of Applied Mathematics and Physics, 2019, 7, 1718-1731

http://www.scirp.org/journal/jamp
ISSN Online: 2327-4379
ISSN Print: 2327-4352

Algebraic Techniques for Linear Equations

over Quaternions and Split Quaternions: A
Unified Approach in Quaternionic and Split
Quaternionic Mechanics

Gang Wang1,2, Zhenwei Guo1, Dong Zhang3, Tongsong Jiang1,2,3*

1
School of Mathematical Science, Liaocheng University, Liaocheng, China
2
School of Mathematics and Statistics, Heze University, Heze, China
3
College of Mathematics and Systems Science, Shandong University of Science and Technology, Qingdao, China

How to cite this paper: Wang, G., Guo, Abstract

Z.W., Zhang, D. and Jiang, T.S. (2019)
Algebraic Techniques for Linear Equations This paper aims to present, in a unified manner, algebraic techniques for li-
over Quaternions and Split Quaternions: A near equations which are valid on both the algebras of quaternions and split
Unified Approach in Quaternionic and
quaternions. This paper, introduces a concept of v-quaternion, studies the
Split Quaternionic Mechanics. Journal of
Applied Mathematics and Physics, 7, problem of v-quaternionic linear equations by means of a complex represen-
1718-1731. tation and a real representation of v-quaternion matrices, and gives two alge-
https://doi.org/10.4236/jamp.2019.78118 braic methods for solving v-quaternionic linear equations. This paper also
Received: July 18, 2019
gives a unification of algebraic techniques for quaternionic and split quater-
Accepted: August 13, 2019 nionic linear equations in quaternionic and split quaternionic mechanics.
Published: August 16, 2019
Keywords
Copyright © 2019 by author(s) and
Scientific Research Publishing Inc. V-Quaternion, Complex Representation, Real Representation,
This work is licensed under the Creative Linear Equations, Quaternion, Split Quaternion
Commons Attribution International
License (CC BY 4.0).
http://creativecommons.org/licenses/by/4.0/
Open Access
1. Introduction
A quaternion, which was found in 1840 by William Rowan Hamilton [1], is in
the form of q =q1 + q2 i + q3 j + q4 k , i 2 = j2 = k 2 = −1 , ijk = −1 , where
q1 , q2 , q3 , q4 ∈ R , and ij =− ji =k , jk =−kj =i , ki =−ik =j . Quaternion al-
gebra has been playing a significant role recently in geometric and physical ap-
plications, many geometric problems can be represented by quaternions. In pa-
per [2], the authors showed that a unit timelike quaternion represents a rotation
in the Minkowski 3 space, and expressed Lorentzian rotation matrix generated

DOI: 10.4236/jamp.2019.78118 Aug. 16, 2019 1718 Journal of Applied Mathematics and Physics
 G. Wang et al.

with a time like quaternion. In paper [3], the authors studied the problem of us-
ing quaternions in unconstrained nonlinear optimization of 3-D rotations, and
gave an easy and accurate method for applying the quaternion representation of
3-D rotations.
A split quaternion (or coquaternion), which was found in 1849 by James
Cockle [4], is in the form of q =q1 + q2 i + q3 j + q4 k , i 2 = −1 , = j2 k=
2
1,
ijk = 1 , where q1 , q2 , q3 , q4 ∈ R , and ij =− ji =k , jk =−kj =−i , ki =−ik =j
and denotes the sets of quaternions and split quaternions respectively by H
and H s . The quaternion ring H and the split quaternion ring H s are two
associative and noncommutative 4-dimensional Clifford algebras, and the split
quaternion ring H s contains zero divisors, nilpotent elements and nontrivial
idempotents. In paper [5], the authors stated the rotations in Minkowski 3 space
by split quaternions. In paper [6], the authors studied dual split quaternions and
screw motion in 3-dimensional Lorentzian space, and obtained the components
of a dual split quaternion by replacing the L-Euler parameters with their split
dual versions. In paper [7], the authors studied eigenvalue problem of a rotation
matrix in Minkowski 3 space by using split quaternions, and gave the characte-
rizations of eigenvalues of a rotation matrix in Minkowski 3 space according to
only first component of the corresponding quaternion. Quaternions and split
quaternions in the study of geometry and physic are more than those, e.g.
[8]-[13].
A v-quaternion is in the form of
q =q1 + q2 i + q3 j + q4 k, i 2 =−1, j2 =v, ij =− ji =k, (1.1)

in which 0 ≠ v ∈ R , q1 , q2 , q3 , q4 ∈ R , and k=2

= v , jk =
ijk −kj =−vi ,
ik =−ki =− j . Let H v denote the set of v-quaternion. Obviously, the set of all
v-quaternion is also a noncommutative 4-dimensional Clifford algebra. Specially,
when v = −1 , the ring of the v-quaternion H v is the ring of the quaternion H ;
when v = 1 , the ring of the v-quaternion H v is the ring of the split quaternion
Hs .
In the geometry research and physical application of quaternion and split qu-
aternion, the problems of solving quaternionic and split quaternionic equations
are often encountered. In paper [14], by means of a complex representation of
quaternion matrices, the authors studied the problems of quaternionic linear
equations, and gave an algorithm for quaternionic linear equations. In paper
[15], by means of a complex representation and a real representation of split qu-
aternion matrices, the authors studied the split quaternionic least squares prob-
lem, and derived two algebraic methods for finding solutions of the problems in
split quaternionic mechanics. For the problems of quaternions and split quater-
nions, the scholars need to discuss by classification. However, as two special cas-
es of four-dimensional algebra, it is of theoretical and practical significance to
solve them in a unified way. This paper aims to present, in a unified manner, al-
gebraic techniques for linear equations which are valid on both the algebras of
quaternions and split quaternions. This paper, by means of a complex represen-

DOI: 10.4236/jamp.2019.78118 1719 Journal of Applied Mathematics and Physics

G. Wang et al.

tation and a real representation of v-quaternion matrices, studies the problem of

v-quaternionic linear equations, and gives two algebraic methods for solving
v-quaternionic linear equations. This paper also gives a unification of algebraic
techniques for quaternionic and split quaternionic linear equations in quater-
nionic and split quaternionic mechanics.
Let R be the real number field, C= R ⊕ Ri the complex number field. If
q =q1 + q2 i + q3 j + q4 k ∈ Η v , q =q1 − q2 i − q3 j − q4 k is the conjugate of q. For
any matrix=A ( ast ) ∈ H vm×n , A = ( ast ) , A = ( ats ) , A = ( ats ) , A−1 denote
T *

the conjugate, the transpose, the conjugate transpose and the inverse of the ma-
trix A, respectively.
This paper is organized as follows. In Section 2, we give two new matrix re-
presentations of v-quaternion matrix, and discuss some properties and conclu-
sions of complex representation and real representation of v-quaternion matric-
es. In Section 3, we present the complex representation method for solving
v-quaternionic linear equations and some numerical examples. In Section 4, we
present the real representation method for solving v-quaternionic linear equa-
tions and some numerical examples. In Section 5, we summarize this paper.

2. Complex Representation and Real Representation of

V-Quaternion Matrices
For any v-quaternion matrix
A = A1 + A2 i + A3 j + A4 k =( A1 + A2 i ) + ( A3 + A4 i ) j =B1 + B2 j ∈ H vm×n ,
A1 , A2 , A3 , A4 ∈ R m×n , B1 , B2 ∈ Cm×n , the complex representation AC of the
v-quaternion matrix A is defined to be
B vB2 
AC =  1 (2.1)
B1 
,
 B2
and the real representation AR of the v-quaternion matrix A is defined to be
 A1 − A2 vA3 vA4 
A A1 vA4 −vA3 
AR =  2 . (2.2)
 A3 A4 A1 − A2 
 
 A4 − A3 A2 A1 

For any v-quaternion matrix A, B ∈ H vm×n , C ∈ H vn× p , a ∈ R , for

σ ∈ {C , R} , it is easy to prove the following equalities by direct calculation.
( A + B) =Aσ + Bσ , ( aA ) =aAσ , ( AC ) =Aσ C σ ,
σ σ σ
(2.3)

and

Qm−1 AC Qn = AC , (2.4)
0 vI t 
where Qt =  .
 It 0 
Similarly, by direct calculation we get the following results.
Pm−1 AR Pn A=
= R
, Rm−1 AR Rn A=
R
, Sm−1 AR Sn AR , (2.5)

DOI: 10.4236/jamp.2019.78118 1720 Journal of Applied Mathematics and Physics

 G. Wang et al.

0 − It 0 0 0 0 vI t 0
I 0 0 0  0 0 0 vI t 
where Pt =  t , Rt =  ,
0 0 0 It   It 0 0 0
   
0 0 − It 0 0 It 0 0
0 0 0 vI t 
0 0 −vI t 0  1 1
St =  , and Pt −1 = − Pt , Rt−1 = Rt , St−1 = St .
 0 − It 0 0 v v
 
 It 0 0 0
Lemma 2.1 For two special cases of quaternion (v = −1) and split quater-
nion ( v = 1) matrices, clearly by (2.1) and (2.2) the complex representation and
the real representation are respectively to be
 A1 − A2 − A3 − A4 
− A − A4 A3 
 1
B B2  2 A1
=AC =  , A R
, A ∈ H m× n . (2.6a)
 2
B B1 
 A3 A4 A1 − A2 
 
 A4 − A3 A2 A1 

 A1 − A2 A3 A4 
 − A3 
 B1 B2  R  A2 A1 A4
=AC
= , A , A ∈ H ms ×n . (2.6b)
 B2 B1   A3 A4 A1 − A2 
 
 A4 − A3 A2 A1 

For any v-quaternion matrix A ∈ H vm×n , the rank rank ( A ) of the matrix A is
defined to be

rank ( A ) ≡
1
2
( )
rank AC , (2.7)

or

rank ( A ) ≡
1
4
rank AR . ( ) (2.8)

By the definition of rank and (2.3), it is easy to get the following results by di-
rect calculation. If A, B ∈ Η vm×n , C ∈ Η vn× p , then
rank ( A + B ) ≤ rank ( A ) + rank ( B ) and rank ( AC ) ≤ min {rank ( A ) , rank ( C )} .

3. Algebraic Method of Complex Representation

If A ∈ H vm×n , B ∈ H vm× p , then by the definition of complex representation and
(2.3), AX = B if and only if AC X C = B C . That is AX = B has a solution X if
and only if AC Y = B C has a solution Y = X C .
Theorem 3.1 For A ∈ H vm×n , B ∈ H vm× p . Then
1) V-quaternionic linear equations AX = B have a solution if and only if
rank ( A ) = rank ( A, B ) , i.e. AX = B has a solution if and only if AC Y = B C
rank ( A ) rank
has a solution, and if= = ( A, B ) n , then v-quaternionic linear eq-
uations AX = B have a unique solution.
2) If Y is a solution to AC Y = B C , then the following v-quaternion matrix is a
solution to AX = B ,

DOI: 10.4236/jamp.2019.78118 1721 Journal of Applied Mathematics and Physics

G. Wang et al.

 Ip 
=X
1
4
(
[ I n , I n j] Y + Qn−1YQp  1  ,
Ip j
) (3.1)
 v 
 0 vI t 
in which Qt =  .
 It 0 
Proof: If Y is a solution of AC Y = B C , by (2.4),

(
B C ⇔ AC Qn−1YQ p =
AC Y = )
B C ⇔ AC Qn−1YQ p =
BC , (3.2) ( )
i.e. Qn−1YQ p is a solution of AC Y = B C , therefore

=
Yˆ
1
2
(
Y + Qn−1YQ p ) (3.3)

is also a solution of AC Y = B C . Let

z z12 
Y =  11 ∈ C2 n×2 p , zts ∈ Cn× p , s, t = 1, 2. (3.4)
 z21 z22 

It is easy to get, by direct calculation,

 zˆ vzˆ2 
=Yˆ  1 ∈ C 2 n× 2 p , (3.5)
 zˆ2 zˆ1 

in which
1 11 
zˆ1 = ( z11 + z22 ) , zˆ2 =  z12 + z21  . (3.6)
2 2v 
By (3.5), we construct a v-quaternion matrix.
 Ip 
1
X =zˆ1 + zˆ2 j = [ I n , I n j] Yˆ  1  . (3.7)
2  I p j
 v 

Clearly X C = Yˆ . This means that X C = Yˆ is a solution of AC Y = B C , so X

is a solution of AX = B .
From the statement above we get following results. When the v-quaternionic
linear equations AX = B have a solution, we can find a solution by a solution
of complex representation equation AC Y = B C from the formula (3.1).
The following two special cases about quaternions and split quaternions come
from Theorem 3.1 respectively with v = −1 and v = 1 .
Corollary 3.2 For A ∈ H m×n , B ∈ H m× p . Then
1) The quaternionic linear equations AX = B have a solution if and only if
rank ( A ) = rank ( A, B ) , i.e. AX = B has a solution if and only if AC Y = B C
rank ( A ) rank
has a solution, and if= = ( A, B ) n , then quaternionic linear equa-
tions AX = B have a unique solution.
2) If Y is a solution to AC Y = B C , then the following quaternion matrix is a
solution to AX = B ,
 I 
=X
1
(
[ I n , I n j] Y + Qn−1YQp  − Ip j , ) (3.8)
4  p 

DOI: 10.4236/jamp.2019.78118 1722 Journal of Applied Mathematics and Physics

 G. Wang et al.

 0 − It 
in which Qt =  .
 It 0 
Corollary 3.3 For A ∈ H ms ×n , B ∈ H ms × p . Then
1) The split quaternionic linear equations AX = B have a solution if and
only if rank ( A ) = rank ( A, B ) , i.e. AX = B has a solution if and only if
rank ( A ) rank
AC Y = B C has a solution, and if= = ( A, B ) n , then split quater-
nionic linear equations AX = B have a unique solution.
2) If Y is a solution to AC Y = B C , then the following split quaternion matrix
is a solution to AX = B ,

I 
=X
1
( )
[ I n , I n j] Y + Qn−1YQp  I p j , (3.9)
4  p 

0 It 
in which Qt =  .
 It 0 
In the similarly way, we have the following result.
Theorem 3.4 For A ∈ H vm×n , C ∈ H vp×q , B ∈ H vm×q . Then
1) V-quaternionic matrix equation AXC = B has a solution if and only if
C 
rank ( A ) = rank ( A, B ) and rank ( C ) = rank   , i.e. AXC = B has a solution
B
rank ( A ) rank
if and only if AC YC C = B C has a solution, and if= = ( A, B ) n ,
C 
=
rank ( C ) rank
=   p , then v-quaternionic matrix equation AXC = B has a
B
unique solution.
2) If Y is a solution to AC YC C = B C , then the following v-quaternion matrix
is a solution to AXC = B ,

 Ip 
=X
1
4
( )
[ I n , I n j] Y + Qn−1YQp  1  ,
Ip j
(3.10)
 v 

0 vI t 
in which Qt =  .
 It 0 
The proof process is similar to the Theorem 3.1.
Remark 1 The above theorems and corollaries not only give the necessary and
sufficient conditions for quaternion and split quaternion matrix equations
= , AXC B to have a solution, but also a unification of representation
AX B=
for a solution.
Example 3.1
Let

 i 1 + j  i 
=A =  and B   .
 −1 + j − k   −1

Find all solutions of the v-quaternionic linear equations AX = B .

By the complex representation of the v-quaternion matrix, we know

DOI: 10.4236/jamp.2019.78118 1723 Journal of Applied Mathematics and Physics

G. Wang et al.

i 1 0 v  i 0
 −1 0 v −vi   −1 0 
=AC =  , BC  ,
 0 1 −i 1   0 −i 
   
 1 i −1 0   0 −1
7 ± 45
and if v ≠
2
=
, then rank AC ( )
=
rank AC , B C (
4 , i.e. )
rank ( A ) rank
= = ( A, B ) 2 , then the v-quaternionic linear equations AX = B
have a unique solution.
For the matrix equation AC Y = B C , the unique solution is easily found to be
 −5v + 1 v (1 − 2v ) 
 2 
 v − 7v + 1 v 2 − 7v + 1 
 −3v v ( −1 − v ) 
 2 i i
Y =  v − 7v + 1 v 2 − 7v + 1  .
 1 − 2v −5v + 1 
 2 
 v − 7v + 1 v − 7v + 1 
2

 1+ v 3v 
 v 2 − 7v + 1 i v − 7v + 1 
2
i

By (3.1), we easily find the unique solution X of v-quaternionic linear equa-

tions AX = B , and
1
=X
1
4
[2 2]
I , I j Y(+ Qn
−1
YQ p
1 
 j )
 v 
 −5v + 1 1 − 2v −3v 1+ v
T

=
 v 2 − 7v + 1 + v 2 − 7v + 1 j v 2 − 7v + 1 i − v 2 − 7v + 1 k  .
 

The following two examples are special cases of the above conclusion.
Case 1: For quaternionic linear equations AX = B with v = −1 . It is easy to
know AC and B C by (2.6a),
 i 1 0 −1 i 0
 −1 0 −1 i   −1 0 
=AC =  , BC  
 0 1 −i 1   0 −i 
   
 1 i −1 0   0 −1
=
and ( )
rank AC =
rank (
AC , B C )
4 . Clearly, the linear equations AC Y = B C
have a unique solution. The unique solution is easily found to be
2 1
 3 −3 
 
1 i 0 
3 
Y = .
 1 2 
3 3 
 1 
 0 − i
 3 
By (3.8), we easily find the unique solution X of quaternionic linear equations

DOI: 10.4236/jamp.2019.78118 1724 Journal of Applied Mathematics and Physics

 G. Wang et al.

AX = B , and
 1  2 1 1 
T
1
X = (
[ I 2 , I 2 j] Y + Qn−1YQp   =)  + j i  .
4  − j  3 3 3 
Case 2: For split quaternionic linear equations AX = B with v = 1 . It is easy
to know AC and B C by (2.6b),
 i 1 0 1 i 0
 −1 0 1 −i   −1 0 
=AC =  , BC  
 0 1 −i 1   0 −i 
   
 1 i −1 0   0 −1
=
and rank AC( ) =
rank (
AC , B C )
4 . Clearly, the linear equations AC Y = B C
have a unique solution. The unique solution is easily found to be
 4 1 
 5 5 
 
 3i 2 
i
 5 5 
Y = .
 1 4 
 5 5 
 2 3 
− i − i
 5 5 
By (3.9), we easily find the unique solution X of split quaternionic linear equa-
tions AX = B , and
1  4 1
T
3 2 
X =
1
( )
[ I 2 , I 2 j] Y + Qn−1YQp   = + j i + k
5 5 
.
4  j  5 5

4. Algebraic Method of Real Representation

If A ∈ H vm×n , B ∈ H vm× p , then by the definition of real representation, AX = B if
and only if AR X R = B R . That is AX = B has a solution X if and only if
ARY = B R has a solution Y = X R .
Theorem 4.1 For A ∈ H vm×n , B ∈ H vm× p . Then
1) V-quaternionic linear equations AX = B have a solution if and only if
rank ( A ) = rank ( A, B ) , i.e. AX = B has a solution if and only if ARY = B R
has a solution, and if= rank ( A ) rank
= ( A, B ) n , then v-quaternionic linear eq-
uations AX = B have a unique solution.
2) If Y is a solution to ARY = B R , then the following v-quaternion matrix is a
solution to AX = B ,
 Ip 
 −I i 
 p 
 1
[ I n , I n i, I n j, I n k ]  Y − PnYPp + RnYR p + SnYS p   I p j  .
1 1 1
=X (4.1)
16  v v  v 
1 
 v I p k 

Proof: If Y is a solution of ARY = B R , by (2.5),

DOI: 10.4236/jamp.2019.78118 1725 Journal of Applied Mathematics and Physics

G. Wang et al.

B R ⇔ Pm−1 AR Pn Pn−1YPp =
ARY = ( )
Pm−1 B R Pp ⇔ AR Pn−1YPp =
BR , (4.2) ( )
i.e. Pn−1YPp is a solution of ARY = B R . Similarly, Rn−1YR p , Sn−1YS p are also so-
lution of ARY = B R .
1
(
Yˆ = Y + Pn−1YPp + Rn−1YR p + Sn−1YS p
4
)
(4.3)
1 1 1 
=  Y − PnYPp + RnYR p + SnYS p 
4 v v 
is also a solution of ARY = B R . Let
 z11 z12 z13 z14 
z z22 z23 z24 
=Y  21 ∈ R 4 n×4 p , zts ∈=
R n× p , s, t 1, 2,3, 4. (4.4)
 z31 z32 z33 z34 
 
 z41 z42 z43 z44 

It is easy to get, by direct calculation,

 zˆ1 − zˆ2 vzˆ3 vzˆ4 
 zˆ zˆ1 vzˆ4 −vzˆ3 
Yˆ =  2 ∈ R 4 n× 4 p , (4.5)
 zˆ3 zˆ4 zˆ1 − zˆ2 
 
 zˆ4 − zˆ3 zˆ2 zˆ1 

in which
1 1
zˆ=
1 ( z11 + z22 + z33 + z44 ) , zˆ2= ( z21 − z12 + z43 − z34 ) , (4.6a)
4 4
1 1 1  1 1 1 
zˆ=
3  z31 − z42 + z13 − z24  , zˆ=
4  z41 + z32 + z23 + z14  . (4.6b)
4 v v  4  v v 
By (4.5), we construct a v-quaternion matrix.
 Ip 
 −I i 
 p 
1  
X = zˆ1 + zˆ2 i + zˆ3 j + zˆ4 k = [ I n , I n i, I n j, I n k ] Y  1 I p j  .
ˆ (4.7)
4
v 
1 
I k
 v p 

Clearly X R = Yˆ . This means that X R = Yˆ is a solution of ARY = B R , so X

is a solution of AX = B .
From the statement above we get following results.
The following two special cases about quaternions and split quaternions come
from Theorem 4.1 respectively with v = −1 and v = 1 .
Corollary 4.2 For A ∈ H m×n , B ∈ H m× p . Then
1) The quaternionic linear equations AX = B have a solution if and only if
rank ( A ) = rank ( A, B ) , i.e. AX = B has a solution if and only if ARY = B R
rank ( A ) rank
has a solution, and if= = ( A, B ) n , then quaternionic linear equa-
tions AX = B have a unique solution.
2) If Y is a solution to ARY = B R , then the following quaternion matrix is a
solution to AX = B ,

DOI: 10.4236/jamp.2019.78118 1726 Journal of Applied Mathematics and Physics

 G. Wang et al.

 Ip 
 −I i 
[ I n , I n i, I n j, I n k ] (Y − PnYPp − RnYR p − SnYS p )  − I p j  .
1
=X (4.8)
16 p
 
−
 p k 
I

Corollary 4.3 For A ∈ H ms ×n , B ∈ H ms × p . Then

1) The split quaternionic linear equations AX = B have a solution if and
only if rank ( A ) = rank ( A, B ) , i.e. AX = B has a solution if and only if
rank ( A ) rank
ARY = B R has a solution, and if= = ( A, B ) n , then split quater-
nionic linear equations AX = B have a unique solution.
2) If Y is a solution to ARY = B R , then the following split quaternion matrix
is a solution to AX = B ,

 Ip 
− I i
[ I n , I n i, I n j, I n k ] (Y − PnYPp + RnYR p + SnYS p )  I pj  .
1
=X (4.9)
16 p
 
 p k 
I

In the similarly way, we have the following result.

Theorem 4.4 For A ∈ H vm×n , C ∈ H vp×q , B ∈ H vm×q . Then
1) V-quaternionic matrix equation AXC = B has a solution if and only if
C 
rank ( A ) = rank ( A, B ) and rank ( C ) = rank   , i.e. AXC = B has a solution
B
rank ( A ) rank
if and only if ARYC R = B R has a solution, and if= = ( A, B ) n ,
C 
=
rank ( C ) rank
=   p , then v-quaternionic matrix equation AXC = B has a
B
unique solution.
2) If Y is a solution to ARYC R = B R , then the following v-quaternion matrix
is a solution to AXC = B ,

 Ip 
 
 −I pi 
[ I n , I n i, I n j, I n k ]  Y − PnYPp + RnYR p + SnYS p   1 I p j  .
1 1 1
=X (4.10)
16  v v  v 
1 
I k
 v
p 


The proof process is similar to the Theorem 4.1.

Remark 2 The above theorems and corollaries not only give the necessary and
sufficient conditions for quaternion and split quaternion matrix equations
= , AXC B to have a solution, but also a unification of representation
AX B=
for a solution.
Example 4.1
For two v-quaternion matrices A and B in Example 3.1, find solutions of the
v-quaternionic linear equations AX = B .
By the real representation of the v-quaternion matrix, we know

DOI: 10.4236/jamp.2019.78118 1727 Journal of Applied Mathematics and Physics

G. Wang et al.

 0 1 −1 0 0 v 0 0  0 −1 0 0 
 −1 0 0 0 v 0 0 −v   −1 0 0 0 
   
 1 0 0 1 0 0 0 −v  1 0 0 0
   
0 0 −1 0 0 −v −v 0  R  0 −1 0 0  ,
=AR = ,B
 0 1 0 0 0 1 −1 0   0 0 0 −1
   
 1 0 0 −1 −1 0 0 0   0 0 −1 0 
 0 0 0 −1 1 0 0 1  0 0 1 0
   
 0 −1 −1 0 0 0 −1 0   0 0 0 −1

7 ± 45
and if v ≠
2
=
, then rank AR ( ) =
rank (
AR , B R 8 , i.e.)
rank ( A ) rank
= = ( A, B ) 2 , then the v-quaternionic linear equations AX = B
have a unique solution.
For the matrix equation ARY = B R , the unique solution is easily found to be
 −5v + 1 0 (1 − 2v ) v 0 

 0 3v 0 ( −1 − v ) v 
 0 −5v + 1 0 ( 2v − 1) v 
 
Y= 2
1  −3v 0 ( −1 − v ) v 0 .

v − 7v + 1 1 − 2v 0 −5v + 1 0 
 
 0 −1 − v 0 3v 
 0 2v − 1 0 −5v + 1 
 
 −1 − v 0 −3v 0 

By (4.1), we easily find the unique solution X of v-quaternionic linear equa-

tions AX = B , and
 1 
 −i 
 
[ I 2 , I 2i, I 2 j, I 2 k ]  Y − PnYPp + RnYR p + SnYS p   1 j 
1 1 1
=X
16  v v  v 
1 
 v k 

 −5v + 1 1 − 2v −3v 1+ v
T

=
 v 2 − 7v + 1 + v 2 − 7v + 1 j v 2 − 7v + 1 i − v 2 − 7v + 1 k  .
 
The following two examples are special cases of the above conclusion.
Case 1: For quaternionic linear equations AX = B with v = −1 . It is easy to
know AR and B R by (2.6a),
 0 1 −1 0 0 −1 0 0   0 −1 0 0 
 −1 0 0 0 −1 0 0 1   −1 0 0 0 
   
 1 0 0 1 0 0 0 1 1 0 0 0
   
0 0 −1 0 0 1 1 0  R  0 −1 0 0  ,
=AR = ,B
 0 1 0 0 0 1 −1 0   0 0 0 −1
   
 1 0 0 −1 −1 0 0 0   0 0 −1 0 
 0 0 0 −1 1 0 0 1  0 0 1 0
   
 0 −1 −1 0 0 0 −1 0   0 0 0 −1

DOI: 10.4236/jamp.2019.78118 1728 Journal of Applied Mathematics and Physics

 G. Wang et al.

and=
rank AR( )
=
rank (
AR , B C )
8 . Clearly, the linear equations ARY = B R
have a unique solution. The unique solution is easily found to be

6 0 −3 0 
0 −3 0 0 
 
0 6 0 3 
 
1 3 0 0 0
Y=  .
9 3 0 6 0 
 
0 0 0 −3
0 −3 0 6 
 
0 0 3 0 

By (4.8), we easily find the unique solution X of quaternionic linear equations

AX = B , and
1
 −i 
[ I 2 , I 2i, I 2 j, I 2 k ] (Y − PnYPp − RnYR p − SnYS p )  
1
=X
16 −j
 
 −k 
T
2 1 1 
=  + j i .
3 3 3 

Case 2: For split quaternionic linear equations AX = B with v = 1 . It is easy

to know AR and B R by (2.6b),

 0 1 −1 0 0 1 0 0   0 −1 0 0 
   
 −1 0 0 0 1 0 0 −1  −1 0 0 0 
 1 0 0 1 0 0 0 −1 1 0 0 0
   
 0 0 −1 0 0 −1 −1 0  R  0 −1 0 0 
=A R
= ,B
 0 1 0 0 0 1 −1 0   0 0 0 −1
   
 1 0 0 −1 −1 0 0 0   0 0 −1 0 
 0 0 0 −1 1 0 0 1  0 0 1 0
   
 0 −1 −1 0 0 0 −1 0   0 0 0 −1

and=
rank AR( )
=
rank (
AR , B R )
8 . Clearly, the linear equations ARY = B R
have a unique solution. The unique solution is easily found to be

 −4 0 −1 0 
0 3 0 −2 

 0 −4 0 1
 
1 −3 0 −2 0 
Y= −  .
5  −1 0 −4 0 
 
 0 −2 0 3
0 1 0 −4 
 
 −2 0 −3 0 

By (4.9), we easily find the unique solution X of split quaternionic linear equa-
tions AX = B , and

DOI: 10.4236/jamp.2019.78118 1729 Journal of Applied Mathematics and Physics

G. Wang et al.

1
 −i 
[ I 2 , I 2i, I 2 j, I 2 k ] (Y − PnYPp + RnYR p + SnYS p )  
1
=X
16 j
 
 
k
T
4 1 3 2 
=
 5 + 5 j 5 i + 5 k .
 

5. Conclusion
The goal of this paper is to solve the quaternion and split quaternion linear equ-
ations in a unified manner. First, we give the definition of the v-quaternion and
two new matrix representations of v-quaternion matrix. Then we derive two al-
gebraic methods for solving the linear equations of v-quaternion. It is notewor-
thy that this paper not only gives algebraic techniques for solving the linear equ-
ations over v-quaternion algebras, but also a unification of algebraic techniques
for linear equations in quaternionic and split quaternionic theory.

Acknowledgements
We thank the Editor and the referee for their comments. Research of T. Jiang is
funded by the National Natural Science Foundation of China (11771188) and
Shandong Natural Science Foundation (ZR201709250116). This support is
greatly appreciated.

Conflicts of Interest
The authors declare no conflicts of interest regarding the publication of this pa-
per.

References
[1] Hamilton, W.R. (1843) On a New Species of Imaginary Quantities Connected with
a Theory of Quaternions. Proceedings of the Royal Irish Academy, 2, 424-434.
[2] Özdemir, M. and Ergin, A.A. (2006) Rotations with Unit Timelike Quaternions in
Minkowski 3-Space. Journal of Geometry and Physics, 56, 322-336.
https://doi.org/10.1016/j.geomphys.2005.02.004
[3] Schmidt, J. and Nieman, H. (2001) Using Quaternions for Parametrizing 3-D Rota-
tions in Unconstrained Nonlinear Optimization. Vision Modeling, and Visualiza-
tion, Stuttgart, 21-23 November 2001, 399-406.
[4] Cockle, J. (1849) On Systems of Algebra Involving More than One Imaginary and
on Equations of the Fifth Degree. Philosophical Magazine, 35, 434-437.
https://doi.org/10.1080/14786444908646384
[5] Kula, L. and Yayli, Y. (2007) Split Quaternions and Rotations in Semi Euclidean
Space E 42 . Journal of the Korean Mathematical Society, 44, 1313-1327.
https://doi.org/10.4134/JKMS.2007.44.6.1313
[6] Özkaldi, S. and Gündoğan, H. (2011) Dual Split Quaternions and Screw Motion in
3-Dimensional Lorentzian Space. Advances in Applied Clifford Algebras, 21,
193-202. https://doi.org/10.1007/s00006-010-0236-6

DOI: 10.4236/jamp.2019.78118 1730 Journal of Applied Mathematics and Physics

 G. Wang et al.

[7] Özdemir, M., Erdoğdu, M. and Simsek, H. (2014) On the Eigenvalues and Eigen-
vectors of a Lorentzian Rotation Matrix by Using Split Quaternions. Advances in
Applied Clifford Algebras, 24, 179-192. https://doi.org/10.1007/s00006-013-0424-2
[8] Adler, S.L. (1995) Quaternionic Quantum Mechanics and Quantum Fields. Oxford
University Press, New York.
[9] Özdemir, M. and Ergin, A.A. (2005) Some Geometric Applications of Timelike Qu-
aternions. International Conference of the Jangjeon Mathematical Society, Vol. 16,
108-115.
[10] Ramis, C. and Yayli, Y. (2013) Dual Split Quaternions and Chasles’ Theorem in
3-Dimensional Minkowski Space E13 . Advances in Applied Clifford Algebras, 23,
951-964. https://doi.org/10.1007/s00006-013-0405-5
[11] Aslan, S. and Yayli, Y. (2016) Split Quaternions and Canal Surfaces in Minkowski
3-Space. International Journal of Geometry, 5, 51-61.
[12] Brody, D.C. and Graefe, E.M. (2011) On Complexified Mechanics and Coquater-
nions. Journal of Physics A: Mathematical and Theoretical, 44, Article ID: 072001.
https://doi.org/10.1088/1751-8113/44/7/072001
[13] Janovská, D. and Opfer, G. (2013) Linear Equations and the Kronecker Product in
Coquaternions. Mitteilungen der Mathematische Gesellschaft in Hamburg, 33,
181-196.
[14] Jiang, T. (2004) An Algorithm for Quaternionic Linear Equations in Quaternionic
Quantum Theory. Journal of Mathematical Physics, 45, 4218-4222.
https://doi.org/10.1063/1.1794368
[15] Zhang, Z., Jiang, Z. and Jiang, T. (2015) Algebraic Methods for Least Squares Prob-
lem in Split Quaternionic Mechanics. Applied Mathematics and Computation, 269,
618-625. https://doi.org/10.1016/j.amc.2015.07.072

DOI: 10.4236/jamp.2019.78118 1731 Journal of Applied Mathematics and Physics