Keywords
Generalized inverses, Secondary generalized inverses, Secondary transpose, EP matrices
This article is included in the Manipal Academy of Higher Education gateway.
Generalized inverses, Secondary generalized inverses, Secondary transpose, EP matrices
Version 2 of the article incorporates the valuable suggestions from the reviewers. Changes include language correction, a preliminary section highlighting basic definitions, examples elaborating the definitions, etc. Additionally, a few references have been added, updating the bibliography.
See the author's detailed response to the review by P Sam Johnson
See the author's detailed response to the review by Mehsin Jabel Atteya
See the author's detailed response to the review by Pankaj Kumar Manjhi
The theory of symmetric matrices as well as range symmetric matrices are well known in literature. A matrix is said to be EP (or range symmetric), whenever the range space of the matrix is equal to the range space of its conjugate transpose. In other words, matrix is EP whenever its null space is same as that of the null space of its conjugate transpose. Ballantine1 has studied about the product of two EP matrices of specific rank to be again an EP matrix. In2 new characterizations of EP matrices are given. Also, weighted EP matrix is defined and characterized. Meenakshi3 extended the concept of range symmetric matrices over Minkowski space. In 2014, the same author defined range symmetric matrices in indefinite inner product space.4 In Ref. 5, the concept of EP matrices to bounded operator with closed range is defined on a Hilbert space. For more characterizations of EP and hypo EP operators one can refer.6,7
For an matrix , the secondary transpose is related to transpose of the matrix by the relation . Here, the matrix has non zero unitary entries only on the secondary diagonal. For a matrix with complex entries, the secondary transpose will be renamed as secondary conjugate transpose and is given by .
The concept of secondary conjugate transpose is gaining importance in recent years. Shenoy8 has defined Outer Theta inverse by combining the outer inverse and secondary transpose of a matrix . Drazin-Theta matix 9 is a new class of generalized inverse introduced for a square matrix of index m. One can refer10 for the extension of these inverses over rectangular matrices. R. Vijayakumar11 introduced the concept of secondary generalized inverse with the help of secondary transpose of a matrix. This concept is similar to Moore Penrose inverse. But unlike Moore penrose inverse, existence of s-g inverse is not assured in general. In Ref. 12, a necessary and sufficient conditions of existence of s-g invese is given. In the same article, a few characterizations and a determinantal formula for s-g inverse also has been discussed. In 2009, Krishnamoorthy and Vijayakumar13 has defined the concept of S - normal matrices with the help of secondary transopse for a class of complex square matrices. Jayashree14 has defined secondary k - range symmetric fuzzy matrices. Its relation with S - range symmetric fuzzy matrices, k - range fuzzy symmetric matrices and EP matrices are defined.
In this article, we define secondary range symmetric matrices. Several equivalent conditions for a matrix to be secondary range symmetric, is obtained here. Also, the existence of secondary generalized inverse of a secondary range symmetric matrix is discussed.
Below are some useful deifnitions and results related to secondary conjugate transpose.
The set of all matrices with complex entries are denoted by ( for matrices with real entries. Also, represents the null space of the matrix A. The column space and rank of A are denoted by respectively.
(Ref. 15) Let . Then the secondary transpose (or secondary conjugate transpose in the case of complex matrices) of denoted by and is defined as , where , .
(Ref. 16) Let . Then the conjugate secondary transpose of denoted by and is defined as where .
The secondary transpose of a matrix is defined by reflecting the entries through sendary diagonal.
(Ref. 16) A matrix is said to be secondary normal (S - normal) if .
(Ref. 12) is said to be secondary generalized inverse of if
Note that the matrix A is said to be secondary symmetric or if and only if .
(Ref. 12) Given an matrix . The following statements are equivalent.
(Ref. 17) A matrix is said to be EP (or range symmetric) if .
Meenakshi3 has defined EP in Minkowski space and has given equivalent conditions for a matrix to be range symmetric.
Let the components of a complex vector from 0 to n-1, be Let us denote the Minkowski metric tensor by and it is defined as . Now, the Minkowski metric matrix is . The Minkowski inner product on is defined by where is the Hilbert inner product. A space with Minkowski inner product is defined as Minkowski space. The idea of Minkowski space arised when Xing18 tried to study the optical devices described by the Mueller matrix which may not have a singluar value decomposition. The problem was solved by Renardy19 by defining Minkowski space and obtained the singular value decomposition of Mueller matrix over the Minkowski space.
(Ref. 3) A matrix is said to be range symmetric in Minkowski space if and only if .
Here represents the Minkowski adjoint given by where is the Minkowski metric tensor.
In this section, we define secondary range symmetric matrices which is analogous to that of range symmetric matrices. Some equivalent conditions for a matrix to be range symmetric is also given here.
is secondary right (left) normalized g inverse of where A ∈ R(n×n) if , and is - symmetric. ( is - symmetric).
Example 1:
Let Note that the matrix is both secondary right normalized g-inverse and secondary left normalized inverse of A. The conditions can be easily verified.
Also . Therefore X is secondary right normalized g-inverse of A.
Here Here X is also a left normalized g-inverse.
Consider . The S - transpose of is defined as where , .
Example 2:
Consider a matrix The S – transpose of A is given by .
A matrix secondary range symmetric if and only if
Example 3:
Let Here But . Therefore the matrix is not range symmetric, but it is secondary range symmetric.
Let . Then the following conditions are equivalent.
Hence holds true.
This proves the equivalence of and .
Hence and are equivalent.
Hence .
Thus equivalence of and are proved.
Thus equivalence of and are proved.
Thus the equivalence of and is proved.
From the following example, it is clear that EP matrices in Minkowski space defined by Meenakshi3 and secondary range symmetric matrices are two different concepts.
Consider a matrix where . Note that, here the secondary transpose AS of the matrix A coincides with A. Clearly A is secondary range symmetric (i.e., Here, is secondary normal as well as . However, the matrix is not range symmetric in Minkowski’s space since . It is clear that .
In this example, the matrix A is secondary range symmetric, but not range symmetric in Minkowski space.
In the following example, B is range symmetric in Minkowski space. But it is not secondary range symmetric.
Let and . Clearly is not secondary range symmetric.
Observe that is range symmetric in Minkowski space. Since, so that .
These examples shows that secondary range symmetric matrices and range symmetric matrices in Minkowski inverse are two different matrices even though the proof techniques adopted here are similar.
Note that .
A necessary condition for a matrix to be a S - EP (secondary range symmetric) is proved here.
Let . If is secondary normal and , then is secondary range symmetric.
Since is secondary normal, . Hence which implies . Thus is secondary range symmetric.
A relation connecting range symmetric and secondary range symmetric matrices is given below:
Let . Then any two of the following conditions imply the third one.
Since is EP, . By condition (7) of Theorem 0.2, as is secondary range symmetric. Hence is secondary EP .
(1), (3) ⇒ (2)
Since A is EP, Also, from (3) we have which gives Hence, by Theorem 0.2, A is secondary range symmetric.
Since is S - EP, by condition (2) of Theorem 0.2, is range symmetric. Hence . Also, By (3), from which it follows that . Hence is range symmetric. Thus (1) holds.
For any square complex matrix , there exists unique - symmetric matrices such that where and . In the following theorem, an equivalent condition for a matrix to be secondary range symmetric is obtained interms of , the -symmetric part of .
For , is secondary range symmetric if and only if where is the S - symmetric part of .
If is secondary range symmetric, then . For , and . Hence . Thus, , then and hence . Therefore . Thus . Since, both and are -symmetric, they are secondary range symmetric.
andNow, and . Therefore, . Thus is secondary range symmetric.
We shall discuss the existence of secondary generalized inverse inverse of a secondary range symmetric matrix. First, we shall prove certain lemmas, to simplify the proof of the main result.
For an matrix , if exists, then .
For an matrix , if exists, then is the projection on and is the projection on .
if and only if . By definition 2, being S - symmetric, idempotent is the projection on . Similarly, if and only if and is S - symmetric and idempotent. Hence is the projection on .
For an matrix , the following are equivalent:
. Since and is secondary symmetric, by using Theorem 0.2 we have, and . Thus . Hence by Thoerem 0.1, it follows that exists, By Lemma 1 and Theorem 0.2, . Hence is range symmetric. Thus (2) holds.
. Since exists, by Lemma 1, , by equivalence of condition (1) and (5) of Theorem 0.2, is secondary range symmetric which implies that . Hence . By Lemma 2, it follows that , hence . By definition 2, , is -symmetric, idempotent and ; hence and , which implies . Thus holds.
. Since, is S - symmetric and idempotent, , by lemma 1.1, exists and implies is the projection on . For all reflexive g-inverses of , . Since is S - symmetric and idempotent, is -symmetric. Hence by definition 7, exists and which implies . By hypothesis . Therefore . Thus both and are S - symmetric. By definition 2, exists and . By taking secondary transpose on , we get . and . Therefore . By theorem 0.2, is secondary ramge symmetric. , . Thus . Thus (1) holds. Hence the theorem.
Let A be secondary range symmetric matrix. Then exists if and only if .
Since is secondary range symmetric and , the existence of follows from equivalence of (1) and (2) of Thereom 0.6. Conversly, if is secondary range symmetric, and exists, then by equivalence of (2) and (3) of Theorem 0.1 and by Theorem 0.2, . Hence .
In this article we defined and characterized the concept of secondary range symmetric matrices. The Moore Penrose inverse exists for any matrix. But, in the case of secondary generalized inverse this is not true. Here, we obtained a necessary condition for a secondary range symmetric matrix to have an s-g inverse. In fact this condition holds true for the existence of secondary generalized inverse for any matrix.
As an extension of this work, the sum of range symmetric matrices are discussed in Ref. 20. One can think of defining weighted secondary EP matrices and its characterizations. Also, extending secondary range symmetric matrix to indefinite inner product spaces will open up a new area of research.
Views | Downloads | |
---|---|---|
F1000Research | - | - |
PubMed Central
Data from PMC are received and updated monthly.
|
- | - |
Competing Interests: No competing interests were disclosed.
Is the work clearly and accurately presented and does it cite the current literature?
Yes
Is the study design appropriate and is the work technically sound?
Yes
Are sufficient details of methods and analysis provided to allow replication by others?
Yes
If applicable, is the statistical analysis and its interpretation appropriate?
Not applicable
Are all the source data underlying the results available to ensure full reproducibility?
Yes
Are the conclusions drawn adequately supported by the results?
Yes
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: The paper tried to supply new results with avoid any mistake.
Is the work clearly and accurately presented and does it cite the current literature?
Yes
Is the study design appropriate and is the work technically sound?
Partly
Are sufficient details of methods and analysis provided to allow replication by others?
Partly
If applicable, is the statistical analysis and its interpretation appropriate?
Yes
Are all the source data underlying the results available to ensure full reproducibility?
Partly
Are the conclusions drawn adequately supported by the results?
Partly
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: Combinatorial matrices, Discrete Mathematics, computer science
Is the work clearly and accurately presented and does it cite the current literature?
No
Is the study design appropriate and is the work technically sound?
Yes
Are sufficient details of methods and analysis provided to allow replication by others?
Yes
If applicable, is the statistical analysis and its interpretation appropriate?
Not applicable
Are all the source data underlying the results available to ensure full reproducibility?
Yes
Are the conclusions drawn adequately supported by the results?
Yes
References
1. Johnson P, Vinoth A: Product and factorization of hypo-EP operators. Special Matrices. 2018; 6 (1): 376-382 Publisher Full TextCompeting Interests: No competing interests were disclosed.
Reviewer Expertise: Functional Analysis
Alongside their report, reviewers assign a status to the article:
Invited Reviewers | |||
---|---|---|---|
1 | 2 | 3 | |
Version 2 (revision) 11 Sep 24 |
read | ||
Version 1 19 Feb 24 |
read | read | read |
Provide sufficient details of any financial or non-financial competing interests to enable users to assess whether your comments might lead a reasonable person to question your impartiality. Consider the following examples, but note that this is not an exhaustive list:
Sign up for content alerts and receive a weekly or monthly email with all newly published articles
Already registered? Sign in
The email address should be the one you originally registered with F1000.
You registered with F1000 via Google, so we cannot reset your password.
To sign in, please click here.
If you still need help with your Google account password, please click here.
You registered with F1000 via Facebook, so we cannot reset your password.
To sign in, please click here.
If you still need help with your Facebook account password, please click here.
If your email address is registered with us, we will email you instructions to reset your password.
If you think you should have received this email but it has not arrived, please check your spam filters and/or contact for further assistance.
Comments on this article Comments (0)