Professional Documents
Culture Documents
Surigao State College of Technology Surigao City
Surigao State College of Technology Surigao City
Surigao City 1
Chapter 1
INTRODUCTION
in this chapter.
identification [1,5].
equivalent to the NIEP when n=1 and n=2, as the solution applies
when we ask that the eigenvalues be real. The cases n=3, 4 were
solved by Loewy and London [11]. The RNIEP is open for any n ≥ 5,
[2, 3, 8, 10, 11, 13, 14, 15]. This paper investigates a question
permutative matrices.
Surigao State College of Technology
Surigao City 3
of permutative matrices.
Methodology
Basic/Preliminary Concepts
vertical columns:
a11 a12 ⋯ a1 n
a
A= 21
⋮
[a22 ⋯ a2 n
⋮ aij ⋮
a m 1 am 2 ⋯ amn
.
]
The ith row of A is
[ ai 1 ai 2 ⋯ a ¿ ] (1 ≤ i≤ m ) ;
a1 j
[]
a2 j
⋮
amj
( 1 ≤ j≤ n ) .
diagonal matrix if [ aij ]=0 for i≠ j . For a square matrix [ aij ], the
diagonal entries are the entries a 11 ,a 22 , … , ann, which form the main
diagonal. Thus, for a diagonal matrix, the terms off the main
integer, by A P= A ∙ A ∙∙ ∙∙ ∙ A .
1 0 0 2 0 0 1 0 0
[ ] [ ] [ ]
A= 0 2 0 , B= 0 2 0 ,∧I 3= 0 1 0 .
0 0 3 0 0 2 0 0 1
if AT = A .
1 2 3
[ ]
Example 1.5 The matrix A= 2 3 5 is symmetric.
3 5 6
3
Example 1.7 Let A= 2 3 ,andB=
[ ]
2 2
−1
[ ]
2 .
1 −1
Since AB=BA=I n, we
Definition 1.8 [10] If we start out with an n × n matrix A=[ aij ] and
then cross out some, but not all, of its rows or columns, we
1 2 −4
[
A= 0 1 1
3 0 3 ]
If we consider the second row and third column, we get the
submatrix
A23= [ 13 20 ]
Surigao State College of Technology
Surigao City 7
If we have the first row and second column of submatrix A23, then
A12=[ 3 ] .
exactly one entry in each row and column is equal to 1 and all
sends the first entry to the second position, sends the second
entry to the first position, and leaves the third entry in the
third position:
0 1 0 1 2
[ ][ ] [ ]
1 0 0 2=1
0 0 1 3 3
P2 , P3 , … , P n such that
xT
[ ]
(P x)
B= 2
⋮
( Pn x )
T
T
.
PT AP= X Y ,
0 Z[ ] where X and Z are both square.
T
P AP= [10 11] . On the other hand, if we let this matrix A=[ 01 10],
0 1
then the irreducible matrix is P AP=
T
[ ] 1 0
.
appropriate sign, with exactly one entry from each row and
polynomial.
a realizable list of σ .
writing A ≥ 0.
[ OA OB ]
where the O ' s represent the zero matrices. The O on the top
right is an m× n matrix, while the O on the bottom left is n × p.
4 3 2 1
Example 1.22 If A=
3 −1
and B= [
5 −1 ]
, then [ ]
4 3 0 0
[ A O =3
O B] 0
0
[ −1 0 0
0 2 1
0 5 −1
.
]
Definition 1.23 [7] The eigenvalues of an n × n matrix A are the
σ =( λ :det ( λI − A )=0 ) .
n
tr ( A )=∑ aii .
i=1
follows:
ρ ( A )=max ¿ λi∨¿ . ¿
λ i ∈σ
A= [−23 10]
Then the characteristic polynomial
n
tr ( A )=∑ aii =a11 +a 22=3+0=3
i=1
vertex to itself.
vertex u and edge x are incident with each other, as are v and x .
Figure 1.
sk ( σ) of σ defined by
n (1.
k k
sk ( σ ) =∑ λ =tr( A ), k =1,2, …
1
i=1 1)
k
when tr means trace. As A is a nonnegative matrix for k =1,2 , … it
follows that
(1.
sk ( σ ) ≥ 0 , k=1,2 , …
2)
Surigao State College of Technology
Surigao City 12
Preliminary Results
paper.
5
Theorem 1.1 [14] Let σ = ( λ 1 , λ2 , λ 3 , λ 4 , λ 5 ) and let sk =∑ λki . Suppose
i=1
1.s1=0 ,
2. s3 ≥0 ,
3. λ 2+ λ 5 ≤ 0.
2.4 s 4 ≥ s 22 , and
Chapter 2
Loewy and London (1978) solved the RNIEP for n ∈{3 , 4 }. The
Laffey and Loewy in 1996 that the RNIEP and the SNIEP are
equivalent for n ≤ 4 but that they are not for n ≥ 5. Hence for
In 1996, Johnson et al. [8] proved that RNIEP and the SNIEP
they used symbolic dynamics to study the NIEP. The main results
At the end of [13], the author raises the following problem: “Can
Chapter 3
RESULTS
Main Results
matrices.
Let
(3.
8 51 8 51
a= + √ and b= − √
25 50 25 50 1)
and let
(3.
4 21
(
σ = 1 , a , b ,− ,−
5 25 )
2)
For simplicity, and since σ is fixed in this section, we write sk
5
s1=∑ λ11=¿ 1+
i =1
( 258 + √5051 )+( 258 − √5051 )+( −45 )+( −21
25 )
=0 ¿
5 2 2
8 √ 51 8 √ 51 −4 2 −21 2 3239
s2=∑ λ22=¿ 12 +
i =1
( +
25 50 ) (
+ −
25 50
+
5
+ )( )( )
25
=
1250
¿
Surigao State College of Technology
Surigao City 16
5 3 3
8 √ 51 8 √ 51 −4 3 −21 3
s3=∑ λ33=¿ 13 +
i=1
( +
25 50 )(+ −
25 50
+
5
+ ) ( )( )
25
=0 ¿
5 4 4
8 √ 51 8 √51 −4 4 −21 4 6107321
i=1
4
s4 =∑ λ =¿ 1 +
4
4
+
25 50 ( + −
25 50 ) (
+
5
+
25
= ) ( ) ( )
3125000
¿
5 5 5
8 √ 51 8 √51 −4 5 −21 5 861
s5=∑ λ55=¿ 15 +
i=1
( +
25 50 )(+ −
25 50
+
5
+
25) ( )( )
=
3125
¿
6107321 3239 2 2
Observe that s1=0, sk ≥ 0 for k =2 ,3 , 4 , 5, 4 s 4 =4
3125000
≥
1250( =s 2, )( )
and
matrix. This follows from Theorem 1.1 since s1=0 and s3=0, while
x1 x2 x3 x4 x5
[. .
A= . .
. .
. .
.
.
.
.
.
.
.
.
.
. ,
.
.
]
where x i , i=1,2,3,4,5, are nonnegative real numbers. Also, here and
0 1 0 0 0
0
A= 0
1
0
[ ] 0
0
0
0
1
0
0
0
0
0
0
1
0
1 .
0
0
By Definition 1.30,
5
s1 ( σ )=∑ λ11=tr ( A)=0
i =1
and
5
s3 ( σ )=∑ λ31=tr ( A3 )=0
i=1
1
3
4 5
and column has the value of 1 and the rest are 0 . Then it follows
It follows that x 1=0. If no other x iis equal to zero, then all the
x 3 >0.
0 1 0 0 0 0 x3 0 0 0
0
A=x3 0
1
0
[ ][
0
0
0
0
1
0
0
0
0
0
0
1
0
0
0
0 0 x3 0 0
1 = 0 0 0 0 x3
x3 0 0 0 0
0 0 0 x3 0
]
which is also impossible. Hence, at least one of x 4 , x 5 must be
x1 x2 x3 x4 x5
.[
A= . 0 . .
. 0 +¿+¿ 0 ¿ ]
. + ¿ 0 ¿ +¿+¿+¿ 0 ¿ ,
is impossible,
x1 x2 x3 x4 x5
A=
[ +¿ 0 + ¿ . ¿ . ¿
0 ¿ +¿+ ¿ 0 ¿ +¿+¿ 0 ¿ +¿ 0 ¿+ ¿+¿+¿ 0 ¿
]
which is impossible, where i and j are the vertices of G( A).
x1 x2 x3 x4 x 5
A=
[ +¿ 0 ¿ . . ¿
+¿+¿ 0 ¿ +¿+ ¿ 0 ¿ +¿ 0 ¿ +¿+ ¿+¿ 0 ¿
]
which is also impossible. Hence, this case is not possible.
x1 x2 x3 x4 x5
A=
[.
+¿+¿ 0
0 .
¿ +¿+¿ 0
.
0 ¿
. +¿ 0 ¿
]
and this leads to a contradiction as follows: At least one
x1 x2 x3 x4 x5
A=
[ . 0 . ¿ ¿
+¿ 0 ¿ +¿+ ¿ 0 ¿ 0 ¿ +¿+ ¿+¿ 0 ¿ +¿ 0 ¿ ,
]
while if a 25 is positive, then a 54 and a 42 must be positive
x1 x2 x3 x4 x5
A=
[ . 0 . . ¿
+¿ 0 ¿ +¿+ ¿ 0 ¿ 0 ¿ +¿+ ¿+¿ 0 ¿ +¿ 0 ¿
]
Hence, this case is impossible.
x1 x2 x3 x4 x5
A=
[ .
+¿+¿ 0
0 .
0 +¿+¿+ ¿ 0 0
.
¿
. +¿ 0 ¿
]
and we get a contradiction as in (II), namely either a 24> 0,
This shows that when x 5=0 then a 21>0. Suppose now that x 5=0. Then
A looks like
x1 x2 x3 x4 x5
A=
[ +¿ 0 . . ¿ ¿
0 ¿ 0 ¿ . ¿ . ¿ . ¿ 0 ¿ .¿ 0 ¿ . ¿ . ¿ +¿ . ¿ . ¿ 0 ¿
]
leading us to consider two cases.
a 21, a 25>0.
x1 x2 x3 x4 x5
A=
[ +¿ 0 0 0 + ¿+¿ 0 ¿
0 ¿ +¿+ ¿ 0 ¿ 0 ¿ 0 ¿ + ¿0 ¿ +¿ . ¿ . ¿ 0 ¿
]
It follows that each of the second, third and fourth rows
and the other is x 4. This means that two of these rows are
Suppose that a 31=0 then this implies when a 21>0 . Then A looks
like
Surigao State College of Technology
Surigao City 23
x1 x2 x3 x 4 x 5
A=
[ +¿ 0 0 0 +¿0 ¿ ]
0 ¿ . ¿ . ¿ . ¿ 0 ¿ . ¿ 0 ¿ . ¿ . ¿ +¿ . ¿ . ¿ 0 ¿ ,
x1 x2 x3 x 4 x 5
A=
[ +¿ 0 0 0 +¿0 ¿
0 ¿ +¿+¿ 0 ¿ 0 ¿ +¿ 0 ¿ +¿ 0 ¿ +¿ 0 ¿+ ¿ 0 ¿ ,
]
This leads to a contradiction, as we get the 3-cycle
3 ↦5 ↦4 ↦ 3.
contradiction.
x1 x2 x 3 x4 x5
A=
[ 0 0 ¿ +¿ . ¿
.¿ . ¿ . ¿ . ¿ . ¿ 0 ¿ . ¿ 0 ¿ 0 ¿ +¿+¿ 0 ¿
]
Surigao State College of Technology
Surigao City 24
x1 x2 x3 x4 x5
A=
[ 0 0 ¿ +¿ . ¿
.¿ 0 ¿ . ¿ . ¿ . ¿ 0 ¿ 0 ¿ 0 ¿ 0 ¿ +¿+ ¿ 0 ¿
]
Suppose that a 24=0. Then a 25>0 . We claim that a 31+a 32>0 .
so A looks like
x 1 x 2 x3 x4 x5
A=
[ 0 0 +¿ 0 + ¿ 0 ¿ ]
.¿ 0 ¿ 0 ¿ . ¿ 0 ¿ 0 ¿ . ¿ 0 ¿ . ¿ . ¿ . ¿ 0 ¿ ,
x 1 x 2 x3 x4 x5
A=
[ 0 0 +¿ 0 + ¿+¿ . ¿ ]
.¿ . ¿+ ¿ . ¿ 0 ¿ 0 ¿ . ¿ . ¿ . ¿ . ¿ . ¿ 0 ¿ ,
implying a 34=0.
Surigao State College of Technology
Surigao City 25
A=[ 00 0 + ¿+¿ 0
0 ¿ ]
+¿+¿+ ¿ 0 ¿ 0 ¿ 0 ¿ + ¿ 0 ¿ 0 ¿ 0 ¿ + ¿ 0 ¿ +¿ 0 ¿ +¿ 0 ¿ ,
1
3
4 5
2
Figure 2 Undirected graph corresponding to A
Surigao State College of Technology
Surigao City 26
q ( t )=t 5− ( 3239 3
t +
21919
t−
861
2500 ) ( 62500 ) 15625
.
−3239
Then, the coefficient oft 3 in q ( t )= . We call the 16
2500
−3239
in q i ( t ) must be . This will be used to determine x . We
2500
(i +1)-th row of Ai .
v5 =( 1 ,2 , 3 , 4 ) , v 6 =( 1, 2 , 3 ,1 ) , v 7= (1 , 2 ,5 , 4 ) , v 8=( 1 , 2 ,5 , 1 ) ,
0 x 0 0 1−x
x
A1= 0
0
1−x
[
0
1−x
0
0
1−x 0
0
x
0
x
x
0
0
0 1−x
0
]
and
0 x 0 0 1−x
1−x
A12= 0
0
x
[0
1−x
0
0
x
0
1−x
0
0
x
0
1−x
0
0
x
0
]
Computing the characteristic polynomials of the matrices
satisfy
−3239
−( 3 x 2−3 x+ 2 )= ,
2500
1 38
which yields the two solutions x 1, x 2, given by x 1 ,= − √ ,
2 50
1 38
x 2 ,= + √ . Substituting either one of those numbers into A1 we
2 50
t 5−(3239/2500)t 3 +(2192069/6250000)t−344569/6250000 ,
−( x 2−x +1 )=¿−3239/2500,
equation
−5 x+ 5 x 2=¿ −3239/2500
these are the only distinct cases we have shown that σ cannot be
Chapter 4
Summary
of such matrices.
Recommendation
unsolved.
REFERENCES
Surigao State College of Technology
Surigao City 31
Chapman Hall/CRC.
8. Johnson, C., Laffey, T.J. & Loewy, R. (1996). The real and the
With Applications 9 th
Applications, 302/303:295–302.
Algebra, 6, 83–90.
Appendix A
Surigao State College of Technology
Surigao City 34
CURRICULUM VITAE
PERSONAL DATA
EDUCATIONAL ATTAINMENT