Beruflich Dokumente
Kultur Dokumente
OR IGINA L R ESEA R CH
Open Access
Abstract
In this article, we are to find the root of a square matrix A. Specially, if matrix A has multiple eigenvalues, we present a
manual solution so as to find the root of it. In other words, we focus on solving the equation X 2 = A and find the
solutions.
Keywords: Square root of matrix; Matrix equation; Eigenvalue
MSC: 15A24
Introduction
In recent years, several articles are written about the root
of a matrix, and we can refer to [1-6] for instance. In 2007,
Kh. Ikramov tried to solve the matrix equation XX = I
and XX = I.
At first, he considered the problem as
a b
and then generalized it to general case
XX =
c d
XX = In and XX = In [7]. If matrix A has different eigenvalues 1 , 2 , . . . , n , then the root of A, i.e. solution of the
equation X 2 = A is achieved as the following:
X = VEV 1 , E = diag( 1 , 2 , , n ),
(1)
2
Solution of
X =
a b
. In this section, we are to solve the
c d
equation (2) in four cases as the following:
Let A =
3a+d
b
X=
a+d
2
a+d
2
2 a+d
2
a+3d
4
a+d
2
*Correspondence: a-nazari@araku.ac.ir
Department of Mathematics, Arak University, P. O. Box 38156-8-8349, Arak, Iran
1 =
a+d
,
2
2 =
a+d
+ t,
2
(t 0).
2013
2013 Nazari et al.; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction
in any medium, provided the original work is properly cited.
Page 2 of 6
a+d
2
0
a+d
2
+t
2b
2b
d a d a + 2t
V 1 =
da+2t 1
4
2t
1
ad
2t
4bt
.
In this case,
X = lim VEV 1 = lim
t0
t0
2b
a+b
2 (ba+2t)+(ad)
2b a+d
2 +t
4bt
a+b
(da) a+d
2 (da+2t)
2 +t(ad)
4bt
3a+d
b
4 a+d 2 a+d
2
= c2
a+3d
2
a+d
2
(ab)
a+d
2b a+d
2 +2b
2 +t
2t
a+b
2
+ (da+2t)
a+d
2
+t
a+d
2
X=
c
a , X = 2c
a
a
2 a
are some solutions of (2).
a 0
x y
If A =
and X =
is the root of it, then
c d
z w
setting X 2 = A is the solution above.
Corollary 1. If A has a zero multiple eigenvalue and
b = c = 0, then in addition to the two solutions
in (4), an infinite number of solutions exist as the following
for (2):
a c
a c
, X = 2
,X =
,
X=
a
a a
a
in which , , = 0 C.
Lemma 2. (Case 4) If A has two zero eigenvalues, then
one of the following cases occurs:
a) If b = c = 0, then X = 0 is a solution of the equation.
b) If b = 0 and c = 0, then the equation has no solution.
c) If c = 0 and b = 0, then the equation has no solution.
d) If bc = 0, then the equation has no solution.
x y
, we can verify all four cases prez w
sented before. Case (a) is trivial. For case (b), if b = 0 and
c = 0, then from X 2 = A, we must have:
2
w + xy = 0
x2 + yz = 0
y = 0 or (x + w) = 0.
(x + w)y = 0
z(x + w) = c
Assuming X =
(3)
x2 + yz = a,
(4)
xy + yw =
a2
,
c
(5)
xz + wz = c,
(6)
(7)
z(x + w) = c x + w = 0 and z = 0.
According to both (5) and (6), we achieve
c2
y,
a2
c2
y) = c.
a2
x2
y2 c2
a2
(8)
y
z
a2
c2
(9)
Page 3 of 6
c2
a2
y2 = a
(10)
2cy
a
;
= 2a x w =
(x w)
cy
a
(11)
a2
;
cy
(12)
Proof. 1 = a, 2 = d, 3 = d + t (t 0) E =
a 0
0
0
d 0 .
d+t
0 0
AX1 = 1 X1 X1 = 0
d
AX2 = 2 X2 X2 =
betc e T
AX3 = 3 X3 X3 = t(adt) t 1
2c
az
c
a2
y
x= y
;
a
cy
a
2cy
1 b
V = 0 a d
0 0
2
c
a2
a2
c 2 y2
y
= a 2 = 0 a = 0,
a
2cy
az
4c yz
X = lim VEV 1 =
0
t0
0
a b c
A = 0 d e ,
0 0 d
m=
be( d a) + (d a)(
be
2 d
(a d)2
m=
b
a+ d
2 d
be
2 d
(a d)2
+ c d c a)
1 4 9
Example 1. The second root of A = 0 4 1 equals:
0 0 4
4 26
1 3 9
X = 0 2 14
.
0 0 2
d b c
Theorem 3. The second root of A = 0 d e , in which
0 0 a
a = d and d = 0 equals X such that
b
d 2d
X = 0 d
0 0
+ c d c a)
be( d a) + (d a)(
a b m
a+
d
e ,
X= 0
d
2 d
0
0
d
in which
be+tc
td+t 2 ta
e
t
(ad)2
e
a+ d
1 = d, 2 = d + t (t 0),
d 0
0
3 = a E = 0
d+t 0
a
0
0
Page 4 of 6
1 bt be+cacd
2
(ad)
e
Proof. AXi = i Xi , i = 1, 2, 3 V = 0 1
ad
0 0
1
(ad)2
2 d
.
e
X = lim VEV 1 =
0
d
t0
a+
d
a
0 0
9 6 17
Example 2. Second root of A = 0 9 14 is equal
0 0 25
3 1 75
32
with X = 0 3 74 .
0 0 5
a b c
Theorem 4. Second root of A = 0 d e in which a = d
0 0 a
a b
n
a+
d
and n is as n =
(beca+cd)(a+d)
+eb
2 a
(ad)2
d+c a(ad)
1 = a, 2 = d, 3 = a + t(t 0)
a 0
0
E= 0
d 0 , AXi = i Xi , i = 1, 2, 3
a+t
0 0
1 b
Proof. V = 0 d a
0 0
be+ct+cacd
(ad)2
e
t+ad
X = lim VEV 1 = 0
(beca+cd)(a+d)
+eb
2 a
(ad)2
X= 0
0
e
a+
d
d+c a(ad)
Example 3.
b
a+
d
t0
n=
and
9 7 5
Second root of A = 0 16 4 equals:
0 0 9
31
1 42
4 47 .
0 3
a b c
Theorem 5. Second root of A = 0 a d , that a = 0 is
0 0 a
4cabd
b
a 2
a 8ad a
X= 0
a 2 a .
0 0
a
1 = a, 2 = a + t, (t 0), 3 = a + 2t,
a 0
0
,
a+t 0
(t 0) E = 0
a + 2t
0
0
bd+2ct
1 bd+tc
t2
4t 2
d
,
Proof. AXi = i Xi , i = 1, 2, 3 V = 0 dt
2t
0 1
1
4cabd
b
a 2
a 8a a
1
d
X = lim VEV = 0
a 2a .
t0
0 0
a
4 9 3
Example 4. Calculate the second root of A = 0 4 11 ,
0 0 4
9 147
2 4 64
.
X = 0 2 11
4
0 0 2
0 0 0
Theorem 6. If A = a 0 0 , d = 0 then
b c d
(a) If a = 0, then A has no lower triangular second root.
(b) If a = 0, then A has infinite second roots as
0
0 0
0 0
X = t
, where t is a free parameter.
b dtc c
d
d
d
0 0 1, 100
Example 5. Find the second root of 0 0 1, 200
0 0 1, 600
such
that sum of allentries is
100.
0 10 20
0 t 1103t
4
X = 0 0 30 X = 0 0 30 .
0 0 40
0 0 40
Theorem 7. Find the second root of a lower triangular
matrix which has three eigenvalues (all are 0);
Page 5 of 6
0 0 0
considering A = a 0 0 , with b = 0, we have the
b c 0
following:
0 0 1, 390
0 0 0 such that sum of all entries is 1,391.
0 0 0
0
X = 0
0
1,390
t
0
0
0
0
t X = 0
0
0
0
X = 0
0
1 0
0 1, 390 or
0 0
0 1, 390
0 1 .
0 0
Theorem
0
0
bc d a+ ad
2 a
(ad)2
a
0
0
d
0
0
bc d a+ ad
2 a
(ad)2
c
a+ d
b
a+ d
.
0
1
V = 0
0
bc
t(ad+t)
1
c
ad+t
b
0 ,
da
da 0
0
1 bc
1
c
,
adjA =
=
d a ad+t
t
detA
da
b
0
1
a x12
VEV 1 = 0 a + t
0 x23
b
a+
d
d ,
1
(13)
where
x12
a 0 b
Theorem 9. The second root of matrix A = 0 a 0 ,
0 0 d
when a = b, and a = 0, equals
V 1
Example 7.
b
a+ d
a 0
0
a + t 0 .
E= 0
0
d
x32
c a + t(d a) c(d a) d
.
=
(d a)(a d + t)
X = lim VEV 1
t0
a
=
0
0
)
bc( d a+ 2ad
a
(ad)2
c
a+ d
b
a+ d
.
0
0 1 0
0 0 1
matrices P12 = 1 0 0 , P13 = 0 1 0 and
0 0 1
1 0 0
i+1 1i i+1
1i
1 0 0
2
2 0
2 0 2
i+1
, 0 1 0
P23 = 0 0 1 equals 1i
2
2 0
1i
i+1
0 1 0
0 0 1
2 0 2
1
0
1i
, respectively, where i = 1.
and 0 i+1
2
2
i+1
0 1i
2
2
Page 6 of 6
1 0
0
1 1 1
E = 0 1 + t 0 , and the matrix V = 1 1 1 .Then,
1 0 0
0
0
i
1+t+i
1+ti
1
1
+
t
2