The characterization of g-inverses, minimum norm g-inverses and least square g-inverses of
bimatrices are derived as a generalization of the generalized inverses of matrices.
On Least Square, Minimum Norm Generalized Inverses of Bimatrices
1. IOSR Journal of Mathematics (IOSR-JM)
e-ISSN: 2278-5728, p-ISSN: 2319-765X. Volume 11, Issue 1 Ver. VI (Jan - Feb. 2015), PP 19-34
www.iosrjournals.org
DOI: 10.9790/5728-11161934 www.iosrjournals.org 19 | Page
On Least Square, Minimum Norm Generalized Inverses of
Bimatrices
G.Ramesh*, P.Maduranthaki**
*Associate Professor of Mathematics, Govt. Arts College(Autonomous), Kumbakonam.
**Assistant Professor of Mathematics, Arasu Engineering College, Kumbakonam.
Abstract: The characterization of g-inverses, minimum norm g-inverses and least square g-inverses of
bimatrices are derived as a generalization of the generalized inverses of matrices.
Keywords: g-inverse, minimum norm g-inverse, least square g-inverse.
AMS Classification: 15A09, 15A15, 15A57.
I. Introduction
Let n nC be the space of n n complex matrices of order n. A matrix 1 2BA A A is called a bimatrix if
1A and 2A are matrices of same (or) different orders [7,8] and a bimatrix is called unitary if
* *
B B B B BA A A A I [8]. For 1 2, n nA A C , 1 2BA A A and let *
,B BA r A and BA
denote the
conjugate transpose, rank and generalized inverse of the bimatrix BA where BA
is the solution of the equation
B B B BA X A A [1,4,6]. In general, the generalized inverse BA
of BA is not uniquely determined. Since BA
is not unique, the set of generalized inverses of BA is some times denoted as BA
.
In this paper we analyze the characterization of g-inverse, minimum norm g-inverse and least square
g-inverses of bimatrices as a generalization of the g-inverses of matrices [2,3,5] .
II. Generalized Inverses of Bimatrices
In this section some of the properties of generalized inverses of matrices found in [1,2,3,5] are extended
to generalized inverses of bimatrices .
Theorem: 2.1
Let B B BH A A
and B B BF A A
. Then the following relations hold :
(i)
2
B BH H and
2
B BF F .
(ii) ( ) ( ) ( )B B Br H r F r A
(iii) B Br A r A
(iv) B B B Br A A A r A
.
Proof of (i)
Now
2
.B B BH H H
B B B BA A A A
1 2 1 2 1 2 1 2A A A A A A A A
1 2 1 2 1 2 1 2A A A A A A A A
1 1 2 2 1 1 2 2A A A A A A A A
1 1 1 1 2 2 2 2A A A A A A A A
1 1 1 2 2 2 1 2A A A A A A A A
2. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 20 | Page
1 2 1 2 1 2 1 2A A A A A A A A
1 2 1 2 1 2 1 2A A A A A A A A
B B B BA A A A
B BA A
.BH
Hence ,
2
B BH H .
Now
2
.B B BF F F
B B B BA A A A
1 2 1 2 1 2 1 2A A A A A A A A
1 1 2 2 1 1 2 2A A A A A A A A
1 1 1 1 2 2 2 2A A A A A A A A
1 2 1 2 1 2 1 2A A A A A A A A
B B B BA A A A
B BA A
= BF .
Hence,
2
.B BF F
Proof of (ii)
Now B B Br A r A A
B Br A r H (1)
Also Br A B B Br A A A
1 2 1 2 1 2r A A A A A A
1 1 1 2 2 2r A A A A A A
1 1 2 2 1 2r A A A A A A
1 2 1 2 1 2r A A A A A A
B B Br A A A
B Br H A
B Br A r H (2)
From (1) and (2), we get B Br A r H (3)
Now B B Br A r A A
B Br A r F (4)
Also B B B Br A r A A A
1 2 1 2 1 2r A A A A A A
3. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 21 | Page
1 1 1 2 2 2r A A A A A A
1 2 1 1 2 2r A A A A A A
1 2 1 2 1 2r A A A A A A
B B Br A A A
B Br A F
B Br A r F (5)
From (4) and (5), we get B Br A r F (6)
From (3) and (6) we get B B Br A r H r F
.
(7)
Proof of (iii)
Now B B B Br A r A A A
1 2 1 2 1 2r A A A A A A
1 1 1 2 2 2r A A A A A A
1 1 2 2r A A A A
1 2r A A
Br A
Hence, B Br A r A
.
Proof of (iv)
Now B B B B Br A A A r A A
Also we have B B B B B B Br A A A r A A A A
1 2 1 2 1 2 1 2r A A A A A A A A
1 1 1 1 2 2 2 2r A A A A A A A A
1 2 1 1 1 2 2 2r A A A A A A A A
1 2 1 2 1 2 1 2r A A A A A A A A
B B B Br A A A A
B Br A A
Br H
Br A ( by (7) )
Hence, B B B Br A A A r A
.
Theorem: 2.2
Let BA be a bimatrix, then the following relations hold for a generalized inverse BA
of BA .
(i) B BA A
4. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 22 | Page
(ii) B B B B B BA A A A A A
(iii)
*
* * * *
B B B B B B B BA A A A A A A A
Proof of (i)
Let B B B BA A A A
B B B BA A A A
*
1 2 1 2 1 2 1 2A A A A A A A A
** *
1 2 1 1 1 2 2 2A A A A A A A A
* *
1 1 1 2 2 2A A A A A A
** * * *
1 1 1 2 2 2A A A A A A
* ** * * * * *
1 2 1 2 1 2 1 2A A A A A A A A
** * *
B B B BA A A A
Thus * *
B BA A
(8)
On the other hand, * * *
B B B BA A A A
and so
*
*
B BA A
**
B BA A
(9)
From (8) and (9), we get * *
B BA A
.
Proof of (ii)
Let B B B B B B BG A I A A A A
B B B B B BA A A A A A
1 2 1 2 1 2 1 2 1 2 1 2A A A A A A A A A A A A
1 2 1 2 1 2 1 2 1 2 1 2BG A A A A A A A A A A A A
1 2 1 2 1 2 1 2 1 2 1 2BG A A A A A A A A A A A A
1 2 1 2 1 2 1 2 1 2 1 2A A A A A A A A A A A A
1 2 1 1 1 1 1 2 2 2 2 2A A A A A A A A A A A A
1 1 1 1 1 1 2 2 2 2 2 2A A A A A A A A A A A A
1 1 1 1 1 1 2 2 2 2 2 2A A A A A A A A A A A A
5. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 23 | Page
1 1 1 1 1 1 2 2 2 2 2 2A A A A A A A A A A A A
1 1 1 1 1 1 2 2 2 2 2 2A A A A A A A A A A A A
1 2 1 2 1 2 1 2 1 2 1 2A A A A A A A A A A A A
1 2 1 2 1 2 1 2 1 2 1 2BG A A A A A A A A A A A A
1 2 1 2 1 2 1 2 1 2 1 2
1 2 1 2 1 2 1 2 1 2 1 2
B BG G I I A A A A A A A A A A
A A A A A A A A A A A A
1 2 1 2 1 2 1 2 1 2
1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2
B BG G I I A A A A A A A A
A A A A A A A A A A A A A A A A
B B B B B B B B B B B B B B BG G I A A A A A A A A A A A A
B B B B B B B B B B B B B B BG G I A A A A A A A A A A A A
Let B B BB A A
.
B B B B B B B B BG G I B B B B B B
B B B B BI B B B B
0
Hence, 0BG
That is, 0B B B B B BA I A A A A
0B B B B B BA A A A A A
Hence, B B B B B BA A A A A A
.
Proof of (iii)
Let BG denote a generalized inverse of B BA A
.Then BG
is also a generalized inverse of B BA A
and
2
B B
B
G G
S
is a symmetric generalized inverse of B BA A
.
Let B B B B B B B BH A S A A A A A
1 2 1 2 1 2 1 2 1 2 1 2 1 2BH A A S S A A A A A A A A A A
1 2 1 2 1 2 1 2 1 2 1 2 1 2BH A A S S A A A A A A A A A A
1 2 1 2 1 2 1 2 1 2 1 2 1 2BH A A S S A A A A A A A A A A
6. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 24 | Page
1 1 1 2 2 2 1 1 1 1 2 2 2 2A S A A S A A A A A A A A A
1 1 1 1 1 1 1 2 2 2 2 2 2 2A S A A A A A A S A A A A A
* *
1 1 1 1 1 1 1 2 2 2 2 2 2 2A S A A A A A A S A A A A A
1 1 1 1 1 1 1 2 2 2 2 2 2 2A S A A A A A A S A A A A A
* *
1 1 1 1 1 1 1 2 2 2 2 2 2 2AS A A A A A A S A A A A A
* *
1 1 1 2 2 2 1 1 1 1 2 2 2 2AS A A S A A A A A A A A A
* *
1 2 1 2 1 2 1 2 1 2 1 2 1 2A A S S A A A A A A A A A A
* *
1 2 1 2 1 2 1 2 1 2 1 2 1 2BH A A S S A A A A A A A A A A
* *
1 2 1 2 1 2 1 2 1 2 1 2
* *
1 2 1 2 1 2 1 2 1 2 1 2 1 2
B BH H A A S S A A A A A A A A
A A S S A A A A A A A A A A
* *
1 2 1 2 1 2 1 2 1 2
* *
1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2
A A S S A A A A A A
A A A A S S A A A A A A A A A A A A
* * * * * *
B B B B B B B B B B B B B BA S A A A A A S A A A A A A
* * * * * * *
B B B B B B B B B B B B B B B BH H A S A A A A A S A A A A A A
Let *
B B BS A A
* * * * *
B B B B B B B B B B B B B BH H A S A S A A S A A A S A
*
0B BH H .
Hence , 0BH .
That is , * * *
0B B B B B B BA S A A A A A
.
*
* * * *
0B B B B B B B BA A A A A A A A
*
* * * *
B B B B B B B BA A A A A A A A
.
7. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 25 | Page
III. Minimum norm generalized inverses of bimatrices
In this section some of the characteristics of minimum norm g-inverses of matrices found in [2,3] are
extended to minimum norm g-inverses of bimatrices.
Definition: 3.1
A generalized inverse BA
that satisfies both B B B BA A A A
and B B B BA A A A
is called a
minimum norm g-inverse of BA and is denoted by B m
A
.
Theorem: 3.2
Let BA be a bimatrix, then the following three conditions are equivalent:
(i)
*
B B B Bm m
A A A A
and B B B Bm
A A A A
(ii) * *
B B B Bm
A A A A
(iii) * *
B B B B B Bm
A A A A A A
.
Proof of (i) (ii)
From (i), B B B Bm
A A A A
*
*
B B B Bm
A A A A
*
1 2 1 2 1 2m
A A A A A A
*
1 2 1 2 1 2m m
A A A A A A
*
1 1 1 2 2 2m m
A A A A A A
* *
1 1 1 2 2 2m m
A A A A A A
* *
* * * *
1 1 1 2 2 2m m
A A A A A A
* *
* * * *
1 2 1 2 1 2( )m m
A A A A A A
*
* *
B B Bm
A A A
*
*
B B Bm
A A A
*
B B Bm
A A A
Hence, * *
B B B Bm
A A A A
.
Proof of (ii) (iii)
From (ii), * *
B B B Bm
A A A A
Postmultiply by *
B B BA A A
on both sides
* * * *
B B B B B B B B B Bm
A A A A A A A A A A
* * * *
1 2 1 2 1 2 1 2 1 2 1 2m m
A A A A A A A A A A A A
* * * *
1 1 1 2 2 2 1 1 2 2 1 2m m
A A A A A A A A A A A A
8. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 26 | Page
* * * *
1 1 1 2 2 2 1 1 2 2 1 2m m
A A A A A A A A A A A A
* * * *
1 1 1 2 2 2 1 1 2 2 1 2m m
A A A A A A A A A A A A
* * * *
1 1 2 2 1 2 1 2 1 2 1 2m m
A A A A A A A A A A A A
* ** *
1 1 2 2 1 1 2 2 1 2 1 2m m
A A A A A A A A A A A A
* *
1 1 2 2 1 1 2 2 1 1 2 2m m
A A A A A A A A A A A A
*
1 2 1 2 1 1 2 2B Bm m
A A A A A A A A A A
B B B B B Bm
A A A A A A
B B B Bm
A A A A
( by definition 3.1)
( )B m BA A
( by definition 3.1)
Hence, * *
B B B B B Bm
A A A A A A
.
Proof of (iii) (i)
From (ii) of theorem (2.2), * *
B B B B B BA A A A A A
Replace BA by
*
BA we get
* * * * * *
B B B B B BA A A A A A
* * * *
B B B B B BA A A A A A
* * * * * * *
1 2 1 2 1 2 1 2 1 2BA A A A A A A A A A A
* * * * * *
1 2 1 2 1 2 1 2 1 2A A A A A A A A A A
* * * * * *
1 2 1 2 1 2 1 2 1 2A A A A A A A A A A
* * * * * *
1 1 1 1 2 2 2 2 1 2A A A A A A A A A A
* * * * * *
1 1 1 1 2 2 2 2 1 2A A A A A A A A A A
* * * * * *
1 2 1 1 2 2 1 2 1 2A A A A A A A A A A
* * *
B B B B BA A A A A
* *
B B B Bm
A A A A
( by (iii))
* *
B B B Bm
A A A A
* *
1 2 1 2 1 2B m m
A A A A A A A
* *
1 1 1 2 2 2m m
A A A A A A
9. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 27 | Page
*
* *
1 1 1 2 2 2m m
A A A A A A
* *
* *
1 1 1 2 2 2m m
A A A A A A
* *
1 2 1 2 1 2m m
A A A A A A
*
*
B B B B m
A A A A
Premultiply by * *
B B BA A A
on both sides
*
* * * * *
B B B B B B B B B B m
A A A A A A A A A A
*
*
B B B B B Bm m m
A A A A A A
( by (iii))
*
*
B B m
A A
( by(ii))
*
B Bm
A A
Hence,
*
B B B Bm m
A A A A
.
Also, from (iii) of theorem (2.2),
*
* * * *
B B B B B B B BA A A A A A A A
Replace BA by
*
BA we get
* * * ** * * * * * * *
B B B B B B B BA A A A A A A A
*
* * * *
B B B B B B B BA A A A A A A A
*
* * * *
1 2 1 2 1 2 1 2B Bm
A A A A A A A A A A
*
* * * *
1 2 1 1 2 2 1 2A A A A A A A A
*
* * * *
1 2 1 1 2 2 1 2A A A A A A A A
*
* * * *
1 2 1 1 2 2 1 2A A A A A A A A
*
* * * *
1 1 1 1 2 2 2 2A A A A A A A A
* *
* * * *
1 1 1 1 2 2 2 2A A A A A A A A
* ** ** * * *
1 1 1 1 2 2 2 2A A A A A A A A
* ** *
1 1 1 1 2 2 2 2A A A A A A A A
10. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 28 | Page
* ** *
1 2 1 2 1 2 1 2A A A A A A A A
**
B B B BA A A A
*
B B B BA A A A
B B B BA A A A
( by definition 3.1)
B B B Bm
A A A A
( by definition 3.1)
Premultiply by BA on both sides,
B B B B B Bm
A A A A A A
B B B Bm
A A A A
( by definition 3.1)
Hence, B B B Bm
A A A A
.
Theorem: 3.3
Let BA be a bimatrix, then the following relations hold for a minimum norm generalized inverse
B m
A
of BA :
(i) One choice of *
B B m
A A
is
*
.B Bm m
A A
(ii) One choice of 1
B Bm m
A A
where is a non- zero scalar.
(iii) One choice of B B B m
U A V
is B B Bm
V A U
where BU and BV are unitary bimatrices.
(iv)
*
* *
B B B B B B mm
A A A A A A
.
Proof of (i)
Now * * *
1 2 1 2B BA A A A A A
* *
1 1 2 2A A A A
* * *
1 1 2 2B B m m
A A A A A A
* *
1 1 2 2m m
A A A A
* *
1 1 2 2m mm m
A A A A
* *
1 2 1 2m mm m
A A A A
* *
1 2 1 2m m m
A A A A
*
1 2 Bm m m
A A A
*
B Bm m
A A
.
Hence,
*
*
B B B Bm mm
A A A A
.
11. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 29 | Page
Proof of (ii)
Now 1 2B m m
A A A
1 2 m
A A
1 2m m
A A
1 1
1 2m m
A A
1
1 2m m
A A
1
B m
A
Hence , 1
B Bm m
A A
.
Proof of (iii)
1 2 1 2 1 2B B B m m
U A V U U A A V V
1 1 1 2 2 2 m
U AV U AV
1 1 1 2 2 2m m
U AV U AV
1 1 1 2 2 2m m m m m m
V A U V A U
* * * * * * * *
1 1 1 1 1 1 1 2 2 2 2 2 2 2m m
V VV A U U U V V V A U U U
since * *
B B B Bm
A A A A
* * * *
1 1 1 1 1 2 2 2 2 2m m
V I A U I V I A U I
( Since 1 2 1 2, , &U U V V are unitary matrices )
* * * *
1 1 1 2 2 2m m
V A U V A U
* * * *
1 2 1 2 1 2m m
V V A A U U
* *
B B Bm
V A U
.
Hence, * *
B B B B B Bm m
U A V V A U
.
Proof of (iv)
* * * * * *
1 2 1 2 1 2 1 2B B B Bm m
A A A A A A A A A A A A
* * * *
1 1 2 2 1 1 2 2m
A A A A A A A A
* * * *
1 1 2 2 1 1 2 2m m
A A A A A A A A
* * * *
1 1 1 1 2 2 2 2m mm m
A A A A A A A A
* *
* *
1 1 1 1 2 2 2 2m m m m
A A A A A A A A
* *
* *
1 2 1 1 1 2 2 2m m m m
A A A A A A A A
12. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 30 | Page
* *
* *
1 2 1 2 1 2 1 2m m m m
A A A A A A A A
*
*
B B B Bm m
A A A A
*
*
B Bm
A A
( by (ii) of theorem 3.2)
*
B B m
A A
Hence,
*
* *
B B B B B B mm
A A A A A A
.
IV. Least Square Generalized Inverses of Bimatrices
In this section some of the characteristics of least square g-inverses of matrices found in [2,3,5] are
extended to least square g-inverses of bimatrices .
Definition: 4.1
A generalized inverse BA
that satisfies both B B B BA A A A
and
*
B B B BA A A A
is called a least
square generalized inverse bimatrix of BA and is denoted by B l
A
.
Theorem: 4.2
Let BA be a bimatrix, then the following three conditions are equivalent:
(i) B B B Bl
A A A A
and B B B Bl l
A A A A
(ii) * *
B B B Bl
A A A A
(iii) * *
B B B B B Bl
A A A A A A
.
Proof of (i) (ii)
From (i), B B B Bl
A A A A
*
*
B B B Bl
A A A A
*
1 2 1 2 1 2B l l
A A A A A A A
*
1 1 1 2 2 2l l
A A A A A A
1 1 1 2 2 2l l
A A A A A A
* *
* * * *
1 1 1 2 2 2l l
A A A A A A
* *
* * * *
1 2 1 2 1 2l l
A A A A A A
*
* *
B B Bl
A A A
*
*
B B B l
A A A
*
B B B l
A A A
13. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 31 | Page
Hence , * *
B B B Bl
A A A A
.
Proof of (ii) (iii)
From (ii), * *
B B B Bl
A A A A
Premultiply by *
B B BA A A
on both sides
* * * *
B B B B B B B B B B l
A A A A A A A A A A
* * * *
1 2 1 2 1 2 1 2 1 2 1 2l l
A A A A A A A A A A A A
* * * *
1 2 1 2 1 2 1 2 1 2 1 2l l
A A A A A A A A A A A A
* * * *
1 2 1 2 1 2 1 2 1 2 1 2l l
A A A A A A A A A A A A
* * * *
1 1 1 1 1 1 2 2 2 2 2 2l l
A A A A A A A A A A A A
* ** *
1 1 1 1 1 1 2 2 2 2 2 2l l
A A A A A A A A A A A A
( by (i) theorem 2.2)
* * * *
1 2 1 2 1 2 1 2 1 2 1 2l l
A A A A A A A A A A A A
*
B B B B B B l
A A A A A A
B B B B B B l
A A A A A A
B B B B B B l
A A A A A A
( by definition 4.1)
B B B B l
A A A A
( by definition 4.1)
B B l
A A
( by definition 4.1)
Hence, * *
B B B B B Bl
A A A A A A
.
Proof of (iii) (i)
From (iii) of theorem (2.2),
*
* * * *
B B B B B B B BA A A A A A A A
*
* * * *
1 2 1 2 1 2 1 2B B l
A A A A A A A A A A
( by (iii))
*
* * * *
1 2 1 2 1 2 1 2A A A A A A A A
*
* * * *
1 2 1 2 1 2 1 2A A A A A A A A
* * * *
1 1 1 1 2 2 2 2A A A A A A A A
* * * *
1 1 1 1 2 2 2 2A A A A A A A A
14. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 32 | Page
* ** ** * * *
1 1 1 1 2 2 2 2A A A A A A A A
* ** * * ** *
1 1 1 1 2 2 2 2A A A A A A A A
* ** *
1 1 1 1 2 2 2 2A A A A A A A A
* * * *
1 2 1 2 1 2 1 2A A A A A A A A
*
B B B BA A A A
*
B B B BA A A A
B B B BA A A A
( by definition 4.1)
B B B Bl
A A A A
( by definition 4.1)
Postmultiply by BA on both sides
B B B B B Bl
A A A A A A
B B B Bl
A A A A
( by definition 4.1)
Hence , B B B Bl
A A A A
.
Also, from (ii) of theorem (2.4),
* *
B B B B B BA A A A A A
*
* * *
B B B B B BA A A A A A
*
* * * *
1 2 1 2 1 2 1 2 1 2A A A A A A A A A A
*
* * * *
1 2 1 2 1 2 1 2 1 2A A A A A A A A A A
*
* * * *
1 2 1 2 1 2 1 2 1 2A A A A A A A A A A
* *
* * * *
1 1 1 1 1 2 2 2 2 2A A A A A A A A A A
* ** ** * * * * *
1 1 1 1 1 2 2 2 2 2A A A A A A A A A A
* ** * * *
1 1 1 1 1 2 2 2 2 2A A A A A A A A A A
* * * * * *
1 1 1 1 1 2 2 2 2 2A A A A A A A A A A
* * * * * *
1 1 1 1 1 2 2 2 2 2A A A A A A A A A A
* * * * * *
1 2 1 2 1 1 2 2 1 2A A A A A A A A A A
* * * *
1 2 B B B BA A A A A A
15. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 33 | Page
* *
B B B B l
A A A A
( by (iii))
*** *
B B B B l
A A A A
*** * * *
1 2 1 2 1 2 1 2 l
A A A A A A A A
*
* *
1 2 1 2 1 2 1 2l l
A A A A A A A A
*
* *
1 1 1 2 2 2B l l
A A A A A A A
* *
* *
1 1 1 2 2 2l l
A A A A A A
* *
* *
1 1 1 2 2 2l l
A A A A A A
* *
* *
1 2 1 2 1 2l l
A A A A A A
*
*
B B B Bl
A A A A
Postmultiply by * *
B B BA A A
on both sides
*
* * * * *
B B B B B B B B B Bl
A A A A A A A A A A
*
*
B B B B B Bl l l
A A A A A A
( by (iii) )
*
*
B Bl
A A
( by (ii) )
*
B B l
A A
Hence,
*
B B l
A A
= B B l
A A
.
Theorem: 4.3
Let BA be a bimatrix, then the following relations hold for a least square generalized inverse B l
A
of
BA :
(i) 1
B Bl l
A A
where is a non zero scalar.
(ii) One choice of * *
B B B B B Bl l
U A V V A U
where BU and BV are unitary bimatrices.
Proof of (i)
Now 1 2B l l
A A A
1 2 l
A A
1 2l l
A A
1 1
1 2l l
A A
1
1 2l l
A A
1
B l
A
Hence, 1
B Bl l
A A
.
16. On Least Square, Minimum Norm Generalized Inverses of Bimatrices
DOI: 10.9790/5728-11161934 www.iosrjournals.org 34 | Page
Proof of (ii)
Now 1 2 1 2 1 2B B B l l
U A V U U A A V V
1 1 1 2 2 2 l
U AV U AV
1 1 1 2 2 2l l
U AV U AV
1 1 1 2 2 2l l l l l l
V A U V A U
* * * * * * * *
1 1 1 1 1 1 1 2 2 2 2 2 2 2l l
V V V A U U U V V V A U U U
* *
since B B B Bl
A A A A
* * * *
1 1 1 1 1 2 2 2 2 2l l
I V A I U I V A I U
(since 1 2 1 2, ,U U V and V are unitary bimatrices)
* * * *
1 1 1 2 2 2l l
V A U V A U
* * * *
1 2 1 2 1 2l l
V V A A U U
* *
B B Bl
V A U
Hence, * *
B B B B B Bl l
U A V V A U
.
References
[1]. Ben Isreal.A and Greville.T.N.E, Generalized Inverse: Theory and Applications, Wiley-Interscience, New York, 1994.
[2]. Campbell.S.L and Meyer.C.D, Generalized Inverses of Linear Transformations,Society for Industrial and Applied
Mathematics,Philadelphia, 2009.
[3]. Haruo Yanai, Kei Takeuchi, Yashio Takane, Projection Matrices, Geneneralized Inverse Matrices, and Singular Value
Decomposition, Springer, New York, 2011.
[4]. Penrose.R, A generalized Inverse for Matrices, Proc. Cambridge Phil. Soc., Vol.51, 1955, (PP 406-413).
[5]. Radhakrishna Rao.C, Calculus of Generalized Inverses of Matrices Part-I: General Theory, Sankhya, Series A, Vol.29, No. 3,
September 1967, (PP 317-342).
[6]. Radhakrishna Rao.C and Sujit Kumar Mitra, Generalised Inverse of a matrix and its Applications, Wiley, New York, 1971.
[7]. Ramesh.G. and Maduranthaki.P, On Unitary Bimatrices, International Journal of Current Research, Vol.6, Issue 09, September
2014, (PP 8395-8407).
[8]. Ramesh.G. and Maduranthaki.P, On Some properties of Unitary and Normal Bimatrices, International Journal of Recent Scientific
Research, Vol.5, Issue10 , October 2014, ( PP 1936-1940).