Solutions - Determinants

V
     
 6 5   3 5   3 6 
det(A) = 2  1  2
  

 0 7   4 7   4 0 
= 2(6 7 0) + 1(3 7 5 (4)) 2(0 6 (4))
= 2 42 + 1 41 2 24 = 77.

VI
We can write

(ABC)T = ((AB)C)T

because the matrix multiplication is associative. Now using the property (AB)T = B T AT , then we
have:

(ABC)T = ((AB)C)T = C T (AB)T

Again using this property

(ABC)T = ((AB)C)T = C T (AB)T = C T B T AT

So to prove that (ABC)T = C T B T AT , we use the associativity rule and the property (AB)T = B T AT .
VIII

 
 1 2 3 
       
   5 6   4 6   4 5 
 4 5 6  = 1 2 +3  = 1 (45 48) 2 (36 48) + 3 (32 40) = 3
   

   8 9   8 9   8 8 
 8 8 9 

 

 1 1 2 1 
 
 3 1 4 5 
 = 60

 7 6 1 2 
 
 1 1 3 4 

IX

1. We multiply out each of the bracketed terms in turn, and then re-assemble the calculation.
Firstly, we have

1 1
(A1 B 2 )1 = (B 2 ) (A1 )
1 1
= (BB) (A1 ) = B 1 B 1 A

1
since (A1 ) = A. Next, we get

1
(ABA1 )1 = (A1 ) B 1 A1 = AB 1 A1 ,
2

so

(A1 B 2 )1 (BA1 )(ABA1 )1 (BA)(A1 B)

= (B 1 B 1 A)(BA1 )(AB 1 A1 )(BA)(A1 B)
= (B 1 B 1 A)B(A1 A)(B 1 A1 )B(AA1 )B
= (B 1 B 1 A)B(I)B 1 A1 )B(I)B
= (B 1 B 1 A)BB 1 A1 BB
= (B 1 B 1 A)IA1 BB = B 1 B 1 (AA1 )BB
= B 1 B 1 (I)BB = B 1 (B 1 B)B = B 1 IB
= B 1 B = I.

2. Proof by induction:

we rst verify that what is to demonstrate is true for the rst values of n, so lets consider
n = 2: " #
a11 0
A=
a21 a22

then
det(A) = a11 a22 a21 0 = a11 a22

So for n = 2, it is true that if aij = 0 for i < j then

n
Y
det(A) = a11 a22 ann = aii
i=1

for A = [aij ].

we verify again (just to be fully convinced) for n = 3:

 
 a11 0 0 
       
 a 0   a 0   a a22 
  22  21  21
det(A) =  a21 a22 0  = a11  0  + 0  = a11 a22 a33
 
   a32 a33   a31 a33   a31 a32 
 a31 a32 a33 

So for n = 3, it is true that if aij = 0 for i < j then

n
Y
det(A) = a11 a22 ann = aii
i=1

for A = [aij ].

now we cannot check for every n that it is true (because we would need to check for an
in nite number of values n). So now let assume it is true for a matrix of size n n, and
use this assumption to show it is true for a matrix of size (n + 1) (n + 1):

so we assume  
 a
 11 0 0 0 

 
 a21 a22 0 0 
 . .. ..  = a11 a22 ann
 
 .. . . 
 
 
 an1 ann 
 3

Now lets calculate:

 
 a11 0 0 0   
   a
22 0 0 
 a21 a22 0 0
   
 
.. .. 
.. .. .. . .
   
. . .  = a11 
   
 
   a ann 0 
n,2
 an1 ann 0
   
  
   an+1,2 an+1,n+1 
 a
n+1,1 an+1,n+1 

The other terms of the determinant are multiplied by 0 (and therefore not even
written). So now the determinant of the (n + 1) (n + 1) matrix depends of the
determinant of a n n matrix having all its coef cients above the diagonal being 0.
According to the assumption we made, we know this determinant to be equal to the
product of the elements in this diagonal so:
 
 a11
 0 0 0 

 a21 a22 0 0
 

n+1
.. .. ..
  Y
. . .  = a11 (a22 ann an+1,n+1 ) = aii
 

 
i=1
 an1 ann 0
 

 
 a
n+1,1 an+1,n+1 

So now we now that:

if aij = 0 for i < j then
n
Y
det(A) = a11 a22 ann = aii
i=1

for A = [aij ].
is true for n = 2, 3 and if it is true for n then it is true for n + 1. So by induction, we can
say it is true for any n.