The Moore-Penrose inverses of split quaternions
Wensheng Cao, Zhenhu Chang

TL;DR
This paper introduces the Moore-Penrose inverse for split quaternions and applies it to solve specific linear equations, also characterizing similarity and consimilarity of split quaternions.
Contribution
It extends the concept of Moore-Penrose inverse to split quaternions and provides solutions to linear equations involving them, along with conditions for similarity and consimilarity.
Findings
Solutions to equations $axb=d$, $xa=bx$, $xa=bar{x}$ using Moore-Penrose inverse
Necessary and sufficient conditions for similarity of split quaternions
Necessary and sufficient conditions for consimilarity of split quaternions
Abstract
In this paper, we find the roots of lightlike quaternions. By introducing the concept of the Moore-Penrose inverse in split quaternions, we solve the linear equations , and . Also we obtain necessary and sufficient conditions for two split quaternions to be similar or consimilar.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAlgebraic and Geometric Analysis · Mathematics and Applications · Matrix Theory and Algorithms
The Moore-Penrose inverses of split quaternions
Wensheng Cao, Zhenhu Chang
School of Mathematics and Computational Science,
Wuyi University, Jiangmen, Guangdong 529020, P.R. China
e-mail: [email protected]
Abstract In this paper, we find the roots of lightlike quaternions. By introducing the concept of the Moore-Penrose inverse in split quaternions, we solve the linear equations , and . Also we obtain necessary and sufficient conditions for two split quaternions to be similar or consimilar.
Keywords and phrases: Roots of lightlike quaternion, Moore-Penrose inverse, similar, consimilar
MR(2000) Subject Classifications: 15A33; 11R52
1 Introduction
Let and be the field of real numbers and complex numbers, respectively. The split quaternions are elements of a 4-dimensional associative algebra introduced by James Cockle [1] in 1849. Split quaternions can be represented as
[TABLE]
where are basis of satisfying the equalities:
[TABLE]
Let be the conjugate of and A split quaternion is spacelike, timelike or lightlike if or , respectively. We define to be the real part of and to be the imaginary part of . It is easy to verify that Let be the prime of . Then . For , we define
[TABLE]
Obviously and for A split quaternion can be written as
[TABLE]
Denote for , where T denotes the transpose of a matrix. Using the multiplication rule (1), we have the following formulas:
[TABLE]
where
[TABLE]
Unlike the Hamilton quaternion algebra, the split quaternions contain nontrivial zero divisors, nilpotent elements, and idempotents. Accordingly we define the following sets of split quaternions:
[TABLE]
[TABLE]
[TABLE]
In this paper we mainly concentrate on the properties of lightlike split quaternions or some singular matices induced by split quaternions. In Section 2, we will obtain the characteristics of the sets and find the roots of lightlike quaternions. In Section 3, we will obtain some properties of the Moore-Penrose inverse and solve the linear equation . In Section 4, we solve the linear equations and . As applications, we obtain some necessary and sufficient conditions for two split quaternions to be similar or consimilar.
2 The root of a split quaternion in
Proposition 2.1**.**
* and*
[TABLE]
Proof.
It is obvious that . If then , and therefore ∎
Proposition 2.2**.**
**
Proof.
Let . Then and therefore . By proposition 2.1, which implies that . ∎
Proposition 2.3**.**
**
Proof.
Let . By , we have the following equations:
[TABLE]
If then and . If then is [math] or . ∎
Let . Then , and therefore can be written as
[TABLE]
Theorem 2.1**.**
Let and . Then the equation has solution in the following cases:
(1) if ; ;
(2) if .
Proof.
Let be a root of the equation . Then and therefore . So can be written as . It follows from Proposition 2.1 that
[TABLE]
That is
[TABLE]
[TABLE]
It is obvious that there is no solution if . So we only need to consider the case . Thus In this case, the above four equations are solvable if and only if one of the following conditions holds:
- (1)
- (2)
- (3)
The above three cases conclude the proof of Theorem 2.1. ∎
Remark 2.1**.**
Proposition 2.2, Theorem 2.1 and the results in [4] show how to find the roots of any split quaternions.
3 The Moore-Penrose inverse of elements in
We recall that the Moore-Penrose inverse of a complex matrix is the unique complex matrix satisfying the following equations:
[TABLE]
where ∗ is the conjugate of a matrix. We denote the Moore-Penrose inverse of by . The following lemma is well known.
Lemma 3.1**.**
Let and the identity matrix of order . Then the linear equation has a solution if and only if , furthermore the general solution is
[TABLE]
Definition 3.1**.**
The Moore-Penrose inverse of is defined to be
[TABLE]
For , we have the following facts:
[TABLE]
Since for any , We can verify directly the following proposition.
Proposition 3.1**.**
Let . Then
[TABLE]
[TABLE]
[TABLE]
Proposition 3.1 implies that the concept of Moore-Penrose inverse of split quaternions is well defined.
Theorem 3.1**.**
Let and . Then the equation is solvable if and only if
[TABLE]
in which case all solutions are given by
[TABLE]
Proof.
It is obvious that is equivalent to . By Lemma 3.1 is solvable if and only if
[TABLE]
Returning to quaternion form by Proposition 3.1, we have . Using (8), we can rewrite as (10). By Lemma 3.1, the general solution is
[TABLE]
Hence the general solution can be expressed as
[TABLE]
That is
[TABLE]
By (10) we have . ∎
In similar way, we have the following corollaries.
Corollary 3.1**.**
Let . Then the equation has solutions:
[TABLE]
Corollary 3.2**.**
Let . Then the equation is solvable if and only if , in which case all solutions are given by
[TABLE]
Corollary 3.3**.**
Let . Then the equation is solvable if and only if , in which case all solutions are given by
[TABLE]
4 Similarity and consimilarity
Definition 4.1**.**
We say that two split quaternions are similar (resp. consimilar) if and only if there exists a such that (resp. ).
It is obvious that is equivalent to
[TABLE]
Noting the linear equations studied in Section 3, we assume that for in this section.
We can verify the following proposition directly.
Proposition 4.1**.**
Let . Then the eigenvalues of are
[TABLE]
and
[TABLE]
* if only if one of the following conditions holds:*
- (1)
* and ; in this case .*
- (2)
* and ; in this case and .*
Example 4.1**.**
Two examples of the : (1) , , ; (2) , , .
Lemma 4.1**.**
Let and . If then
[TABLE]
Proof.
It follows from that . Let . Since , we have S=R\big{(}\Im(a^{\prime})\big{)}-L\big{(}\Im(b^{\prime})\big{)} and T=R\big{(}\Im(a)\big{)}-L\big{(}\Im(b)\big{)}. Note that
[TABLE]
[TABLE]
[TABLE]
Hence
[TABLE]
Similarly we can verify that , and are symmetric matrices. ∎
Theorem 4.1**.**
Let and . Then the general solution of linear equation is
[TABLE]
Proof.
It follows from Lemma 4.1 that the general solution of linear equation is . Returning to quaternion form by Proposition 3.1, we get the formula (13). ∎
Theorem 4.2**.**
Let , and . Then the general solution of linear equation is
[TABLE]
Proof.
It follows from Proposition 4.1 that \det\big{(}R(a)-L(b)\big{)}=\det\big{(}R(a)-L(\bar{b})\big{)}. Let
[TABLE]
Then which implies that . Note that
[TABLE]
Let be the general solution of the equation . Then \big{(}R(a)-L(b)\big{)}\big{(}R(a)-L(\bar{b})\big{)}\overrightarrow{x_{1}}=\overrightarrow{0}. By Corollary 3.3,
[TABLE]
It follows from that the general solution of is , which is just (14). ∎
Remark 4.1**.**
Kula and Yayli [3, Theorem 5.5] didn’t treat the case ; for example . Also they overlooked the case of Theorem 4.2, therefore [3, Proposition 5.3 (ii)] should be amended. For example, ; in this case, the eigenvalues of are and . The solutions of can be represented by
[TABLE]
We mention that taking leads to in the above formula.
Lemma 4.2**.**
Let . Then there exists a such that
[TABLE]
Proof.
Kula and Yayli [3, Propositions 5.1 and 5.2] treated the case . We only need to consider the case . We at first show that we can conjugate to .
If then and . Note that
[TABLE]
Hence we can find a such that .
If then . In Theorem 4.1, letting and taking in (13), we get That is .
If or , then by (16), we can find a such that . Otherwise, letting and and taking in (13), we get such that . ∎
The fact that a real number is similar to itself, together with Lemma 4.2, implies the following theorem.
Theorem 4.3**.**
Two split quaternions are similar if and only if
[TABLE]
Note that is equivalent to \big{(}R(a)-L(b)F\big{)}\overrightarrow{x}=0. We can verify the following proposition directly.
Proposition 4.2**.**
Let and . Then the eigenvalues of are
[TABLE]
[TABLE]
and
[TABLE]
* if and only if one of the following conditions holds:*
- (1)
* and ; in this case .*
- (2)
* and ; in this case .*
- (3)
* and ; in this case .*
- (4)
* and ; in this case .*
Example 4.2**.**
Four examples of : (1) , , ; (2) , , ; (3) , , ; (4) , , .
Theorem 4.4**.**
Two split quaternions are consimilar if and only if one of the following two conditions holds:
[TABLE]
Proof.
If are consimilar then there exists an such that . Thus . That is , which implies that . It is obvious that is a solution to . If then the following three split quaternions , and are solutions to , among them there is at least one number in . If then it is a basis of the solution space of . The above observation and Proposition 4.2 imply Theorem 4.4. ∎
Acknowledgements. This work is supported by Natural Science Foundation of China (no:11871379 ) and the Innovation Project of Department of Education of Guangdong Province.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] J. Cockle, On systems of algebra involving more than one imaginary, Philosophical Magazine (series 3) 35 (1849) 434-435.
- 2[2] J.Groß,G, Trenkler, S. Troschke, Quaternions: further contributions to a matrix oriented approach, Linear Algebra Appl. 326 (2001) 205-213.
- 3[3] L. Kula, Y. Yayli, Split quaternions and rotations in semi Euclidean space E 2 4 superscript subscript 𝐸 2 4 E_{2}^{4} , J. Korean Math. Soc. 44 (2007) 1313-1327.
- 4[4] M. O ¨ ¨ O \mathrm{\ddot{O}} zdemir, The roots of a split quaternion, Applied Mathematics Letters 22 (2009) 258-263.
