Chapter 2: DETERMINANTS

AM
PH
Duong T. PHAM

T.
g.
APPLIED LINEAR ALGEBRA
on
Du

Duong T. PHAM September 30, 2020 1 / 28

Outline

1 Motivation

AM
2 Determinant

PH
3 Basic properties

T.
4 Laplace’s formula g.
on
5 Inverse matrix
Du

6 Cramer’s Rule

Duong T. PHAM September 30, 2020 2 / 28

Motivation

Ex.1 Consider Equ. ax = 0 (1)

AM
If a 6= 0, Equ.(1) has a UNIQUE solution x = 0

PH
If a = 0, Equ.(1) has INFINITELY MANY solutions (all x ∈ R are

T.
solutions)
g.
⇒ number a decides the solvability of Equ.1
on
Du

a is called the determinant of [a]

Duong T. PHAM September 30, 2020 3 / 28

Motivation

Ex.2:  
ax + by = 0 ax + by = 0

AM

 

( 
 

ax + by = 0 cx + dy = 0 cx + dy = 0
⇐⇒ ⇐⇒

PH
cx + dy = 0 acx + bcy = 0


acx + bcy = 0



acx + ady = 0 
(ad−bc)y = 0

T.
If ad−bc 6= 0, has a unique solution x = 0, y = 0
g.
on
If ad−bc = 0, has infinitely many solutions
Du

 
a b
ad−bc is called determinant of
c d

Duong T. PHAM September 30, 2020 4 / 28

Matrices of order 1, 2, 3

Definition

AM
 
A = a11 ; det A := a11
 

PH
a a12
A = 11 ; det A := a11 a22 − a12 a21
a21 a22

T.
 
a11 a12 a13
A = a21 a22 a23 ;
a31 a32 a33
g.
on
det A := a11 a22 a33 − a11 a23 a32 − a12 a21 a33
Du

+a12 a23 a31 + a13 a21 a32 − a13 a22 a31

Duong T. PHAM September 30, 2020 5 / 28

Matrices of order n

 
a11 a12 ... a1n

AM
a21 a22 ... a2n 
Given a square matrix A=
. . . . . .

... . . .

PH
an1 an2 ... ann
Determinant of A,

T.
 
a11 a12 ...
g. a1n 

a a ... a2n  X
det A =  21 22
on
:= s(f )a1f (1) a2f (2) . . . anf (n)
. . . . . . ... . . . 
Du

an1 an2 f ∈Sn

... ann 

Duong T. PHAM September 30, 2020 6 / 28

Examples

AM
 
 
a11  = ?
X 1
s(f )a1f (1) = s a = a11
1 11

PH
f ∈S1
 
a11 a12  P
a21 a22  = ? f ∈S2 s(f )a1f (1) a2f (2)
 

T.
   
1 2 1 2
=s g. a a +s a a
1 2 11 22 2 1 12 21
on
= a11 a22 − a12 a21
Du

Duong T. PHAM September 30, 2020 7 / 28

Examples

 
a11 a12 a13 
  X
a21 a22 a23  = s(f )a1f (1) a2f (2) a3f (3)

AM
 
a31 a32 a33  f ∈S3

PH
   
1 2 3 1 2 3
=s a a a +s a a a
1 2 3 11 22 33 1 3 2 11 23 32

T.
   
1 2 3 1 2 3
+s a12 a21 a33 + s a a a
2 g.
1 3 2 3 1 12 23 31
on
   
1 2 3 1 2 3
+s a13 a21 a32 + s a a a
2 1 13 22 31
Du

3 1 2 3
=
a11 a22 a33 − a11 a23 a32 − a12 a21 a33 + a12 a23 a31 + a13 a21 a32 − a13 a22 a31

Duong T. PHAM September 30, 2020 8 / 28

Determinant of transpose matrix

Proposition.
A ∈ Mn (R); there holds det A = det At

AM
Proof. Assume that A = [aij ]. Then At = [aijt ], where aijt = aji

PH
P
det A = f ∈Sn s(f )a1f (1) . . . anf (n) (1)
det At = f ∈Sn s(f )a1f
t t

T.
P P
(1) . . . anf (n) = f ∈Sn s(f )af (1)1 . . . af (n)n (2)
Any f ∈ Sn ; f −1 ∈ Sn , g.
s(f −1 )af −1 (1)1 . . . af −1 (n)n = s(f )af −1 (1)1 . . . af −1 (n)n
on
= s(f )af −1 (1),f (f −1 (1)) . . . af −1 (n),f (f −1 (n))
Du

= s(f )a1f (1) . . . anf (n)

Sums (1) and (2) have n! terms. Each term in (1) appears in
(2) =⇒ (1) = (2) 

Duong T. PHAM September 30, 2020 9 / 28

Determinant

Proposition.
     
 a11 ... a1n  a11 . . . a1n  a11 . . . a1n 

AM
   

 ... ... ... 

. . .
 . . . . . .  . . .
 . . . . . . 
λbi1 +µci1 . . . λbin +µcin  = λ  bi1
 . . . bin  + µ  ci1
 . . . cin 

PH


 ... ... ... 

. . .
 . . . . . .  . . .
 . . . . . . 
 an1 ... ann  an1 . . . ann  an1 . . . ann 

T.
Proof.
X g.
LHS = s(f )a1f (1) . . . (λbif (i) + µcif (i) ) . . . anf (n)
on
f ∈Sn
Du

=X X
λ s(f )a1f (1) . . . bif (i) . . . anf (n) + µ s(f )a1f (1) . . . cif (i) . . . anf (n)
f ∈Sn f ∈Sn

= RHS 

Duong T. PHAM September 30, 2020 10 / 28

Determinant

Corollaries:
     
 a11 . . . a 1n  a11 . . . a1n  a11 . . . a1n 
    
 ... ... . . .   . . . . . . . . .   . . . . . . . . . 

AM

bi1 +ci1 . . . bin +cin  =  bi1 . . . bin  +  ci1 . . . cin 
    
 ... ... . . .   . . . . . . . . .   . . . . . . . . . 

PH

 an1 ... ann  an1 . . . ann  an1 . . . ann 

T.
   
 a11 . . . a1n  a11 . . . a1n 
   
 ... ... ...  . . . . . . . . .
  g.
λai1 . . . λain  = λ  ai1 . . . ain 

on
   
 ... ... ...  . . . . . . . . .
   
Du

 an1 . . . ann  an1 . . . ann 

   
λa11 . . . λa1n  a11 . . . a1n 
   
 . . . . . . . . .  = λn  . . . . . . . . . 
   
λan1 . . . λann  an1 . . . ann 

Duong T. PHAM September 30, 2020 11 / 28

Rows exchange and Determinant

Proposition.
   
. . . ... . . . . . . ... . . .

AM
 
ai1 ... ain  aj1 ... ajn 
 
. . . ... . . . = − . . . ... . . .

PH

aj1 ... ajn  ai1 ... ain 
 
. . . ... . . . . . . ... . . .

T.
Proof. Denote B := matrix in RHS
LHS =
X g.
s(f )a1f (1) . . . aif (i) . . . ajf (j) . . . anf (n)
on
f ∈Sn
Du

P
RHS = s(f )b1f (1) . . . bif (i) . . . bjf (j) . . . bnf (n)
P f ∈Sn
= s(f )a1f (1) . . . ajf (i) . . . aif (j) . . . anf (n)
Pf ∈Sn
= f ∈Sn s(f (i j))s(i j) a1f (1) . . . aif (j) . . . ajf (i) . . . anf (n)
P
=− f ∈Sn s(f )a1f (1) . . . anf (n) = −LHS 

Duong T. PHAM September 30, 2020 12 / 28

Properties

Proposition.
det A = 0 if A satisfies one of the following:

AM
has one zero row or one zero column

PH
has 2 equal rows or 2 equal columns
has 2 proportional rows or 2 proportional columns

T.
Proof.

a11 . . . a1n 
  g.
a11 . . . a1n 

on
   
. . . . . . . . . . . . . . . . . .
   
Du

 0 ... 0  = 0 0 ... 0  = 0
   
. . . . . . . . . . . . . . . . . .
   
an1 . . . ann  an1 . . . ann 
Other statements: Exercises

Duong T. PHAM September 30, 2020 13 / 28

Properties

Corollary.
A ∈ Mn (R); Determinant of A does NOT change when multiplying one

AM
row with a number and adding it into another row, i.e.

PH
   
. . . . . . . . .  . . . ... . . . 
  
ai1 . . . ain  ai1 + λaj1 . . . ain + λajn 

T.
   
. . . . . . . . . =  . . . ... . . . 
  
aj1 . . . ajn   aj1 ... ajn
 g.  
. . . . . . . . .  . . . ... ... 


on
Du

Corollary.
A ∈ Mn (R); Determinant of A does NOT change when adding into any
row with a linear combination of other rows.

Duong T. PHAM September 30, 2020 14 / 28

Product of matrices

Proposition.
A, B ∈ Mn (R); there holds det(AB) = det A det B

AM
 
1 + x1 y1 1 + x1 y2 . . . 1 + x1 yn

PH
1 + x2 y1 1 + x2 y2 . . . 1 + x2 yn 
Ex. Find det A where A =   ...

... ... 
1 + xn y1 1 + xn y2 . . . 1 + xn yn

T.
  
1 x1 0 . . . 0 g. 1 1 ... 1
 1 x2 0 . . . 0   y1 y2 . . . yn 
on
  
A=  1 x3 0 . . . 0   0 0 ... 0 

Du


. . . . . . . . . . . . . . . . . . . . .
1 xn 0 . . . 0 0 0 ... 0
(
0 when n > 2
det A =
(x2 − x1 )(y2 − y1 ) when n = 2

Duong T. PHAM September 30, 2020 15 / 28

Minor and Cofactor

Definition.
A ∈ Mn (R);

AM
The minor Mi,j is the determinant of the (n − 1) × (n − 1)-matrix
that results from A by removing the ith row and the jth column

PH
The expression (−1)i+j Mi,j is known as cofactor

T.
 
−2 2 −3
Ex. A = −1 1 3 
2 0 −1
g.
on
   
−1 3  −2 −3
Du

M1,2 = 
  = 1 − 6 = −5; M2,2 =    = 8;
2 −1 2 −1
 
−2 −3
M3,2 =
  = −3
−1 3 

Duong T. PHAM September 30, 2020 16 / 28

Laplace’s formula

Theorem.

AM
 
a11 a12 . . . a1j ... a1n
a21 a22 . . . a2j ... a2n 

PH
 
. . . . . . . . . ... ... . . .
A= ; Aij := (−1)i+j Mij cofactors of A
 ai1 ai2 . . . aij ... ain 

T.
 
. . . . . . . . . ... ... . . .
an1 an2 . . . anj g.
... ann
on
det A = ai1 Ai1 + ai2 Ai2 + . . . + ain Ain
Du

det A = a1j A1j + a2j A2j + . . . + anj Anj

Duong T. PHAM September 30, 2020 17 / 28

Example

AM
 
−2 2 −3    
  1+2
−1 3  2+2
−2 −3
Ex. −1 1 3  = 2(−1)
 
 2 −1 + 1(−1)
   
 2 −1

PH
 2 0 −1
 
3+2
−2 −3
+0(−1)

T.
 
−1 3 
g.
= −2 × (−5) + 1 × 8
on
= 18
Du

Duong T. PHAM September 30, 2020 18 / 28

Determinants of triangular matrices

Corollary.

AM
 
a11 a12 . . . a1n 

PH

 0 a22 . . . a2n 

. . . = a11 a22 . . . ann
 ... . . . 

T.
 0 0 ... ann 
 
a11
 0 . . . 0  g.
on
a21 a22 . . . 0 

. . . = a11 a22 . . . ann
... . . . 
Du


an1 an2 . . . ann 

Duong T. PHAM September 30, 2020 19 / 28

Triangularization of matrices
   
 1 0 1 0 0 r2 →r2 +r1 1 0 1 0 0
  
−1 2 7 8 9 r3 →r3 −3r1 0 2 8 8 9
  
Ex.  3 0 −1 0 0 = 0 0 − 4 0
 0

AM
 4 1 7 1 0 r4 →r4 −4r1 0 1 3 1 0
  
 8 7 −1 2 3 r5 →r5 −8r1 0 7 − 9 2 3

PH
   
1 0 1 0 0 1 0 1 0 0
  
r2 ↔r4 0 1 3 1 0 r4 →r4 −2r2 0 1 3 1 0

T.
 
= − 0 0 −4 0 0 = − 0 0 −4 0 0
0 2 8 8 9 r5 →r5 −7r2 0 0
g. 2 6 9
  
0 7 −9 2 3 0 0 − 30 −5
on
3
Du

   
1 0 1 0 0 1 0 1 0 0
1
 
r3 → 4 r3 0 1
 3 1 0 r4 →r4 +2r2 0
 1 3 1 0
= − 4 0 0 −1
 0 0  = − 4 0 0 −1 0 0
0 0
 2 6 9 r5 →r5 −30r2 0
 0 0 6 9
0 0 −30 −5 3 0 0 0 −5 3
Duong T. PHAM September 30, 2020 20 / 28
Triangularization of matrices

AM
   
1 0 1 0 0 1 0 1 0 0 

PH
5
 
0
 1 3 1 0 r5 →r5 + 6 r4 0
 1 3 1 0 
−4 0 0 −1 0 0 = − 4 0 0 −1 0 0  = 252

T.
0 0 0 6 9 0 0 0 6 9 
 
21 
0 0 0 −5 3  g. 0 0 0 0 2
on
Du

Duong T. PHAM September 30, 2020 21 / 28

Inverse of a matrix
Recall: A matrix A ∈ Mn (R) is invertible if ∃B ∈ Mn (R) such that
AB = BA = In and B = A−1

AM
Theorem.
Given A ∈ Mn (R); A is invertible if and only if det A 6= 0. Then

PH
 
A11 A21 . . . An1
A12 A22 . . . An2 

T.
A−1 = det1 A 

 Here, Aij are cofactors of A
 ... ... ... 
A1n g.
A2n . . . Ann
on
| {z }
:=P
Du

Proof. (=⇒) A be invertible =⇒ ∃B ∈ Mn (R) satisfying AB = BA = In .

then
det(AB) = det A det B = det In = 1 =⇒ det A 6= 0.
(⇐=) det A 6= 0. We will prove AP = PA = det A · In
Duong T. PHAM September 30, 2020 22 / 28
Inverse of a matrix

(AP)ii = ai1 Ai1 + ai2 Ai2 + . . . + ain Ain = det A ( Laplace’s formula )

AM
(
bk` = ak` ∀k 6= j, ` = 1, . . . , n
i 6= j; Define B ∈ Mn (R) whose
bj` = ai` ∀` = 1, . . . , n

PH
det B = 0 since rows ith and jth equal.

T.
Using Laplace’s formula
g.
0 = det B = bj1 Bj1 + . . . + bjn Bjn = ai1 Aj1 + . . . + ain Ajn = (AP)ij
on
Hence, AP = det A · In . Similarly, PA = det A · In .
Du

1
Therefore, A−1 = P 
det A

Duong T. PHAM September 30, 2020 23 / 28

Find inverse of a matrix
 
1 0 −1
Method 1: Ex. A = −1 1 2  ; det A = 1
1 1 1

AM
   
1+1
1 2 1+2
−1 2
A11 =(−1) 1 1 = − 1; A12 =(−1)  1 1 =3;
   

PH
   
1+3
−1 1 2+1
0 −1
A13 =(−1)  1 1 = − 2; A21 =(−1) 1 1  = − 1;
   

T.
   
2+2
1 −1 g. 2+3
1 0
A22 =(−1) 1 1  =2; A23 =(−1) 1 1 = − 1;
   
on
   
0 −1  1 −1
Du

3+1 3+2
A31 =(−1) 1 2  =1; A32 =(−1) −1 2  = −1;
   
 
  −1 −1 1
 1 0  = 1; ⇒ A−1 = 1  3

A33 =(−1)3+3  1 2 − 1
−1 1
−2 −1 1
Duong T. PHAM September 30, 2020 24 / 28
Find inverse of a matrix

Method 2: Using row elementary operations

     
0 1 1 1 1 0 0 0 3 3 3 3 1 1 1 1

AM
1 0 1 1 0 1 0 0  −→ 1 0 1 10 1 0 0
  
A=1 1 0 1 0

0 1 0 1 1 0 10 0 1 0

PH
1 1 1 0 0 0 0 1 1 1 1 0 0 0 0 1

1 31 1 1 1 
 1 1 1 1   
1 1 1 1 3 3 3 1 1 1

T.
3 3 3 3
1 0 1 10 1 0 0 0 −1 0 0 − 1 2
− 1
− 1
   →  13 3
1 2
3 3
1 1 0 10 0 1 g.
0  0 0 −1 0 − 3 − 3 3 − 13 
0 −1 − 13 − 13 − 31 2
on
1 1 1 0 0 0 0 1 0 0 3
Du

0 − 23 1 1 1 
   2 1 1 1 
1 0 0 3 3 3 −3 3 3 3
0 −1 0  1 2 1 1 1 2 1 1 
− 31
0 − − −

3 3 3  →I  3 3 3 3 
4 1

0 0 −1 0 −
3 − 13 2
3 − 31  3
1
3 − 23 1 
3
0 0 0 −1 − 13 − 13 − 31 2
3
1
3
1
3
1
3 − 2
3

Duong T. PHAM September 30, 2020 25 / 28

Find inverse of a matrix

Method 3: Given A ∈ Mn (R); Find A−1 .

       
x1 y1 x1 y1
 ..   ..   ..  −1  .. 

AM
A  .  =  .  =⇒  .  = A  . 
xn yn xn yn

PH

      x + x + x = y1
0 1 1 1 x1 y1 


2 3 4
1 0 1 1 x2  y2  x1 + x3 + x4 = y2

T.
Ex.  1 1 0 1 x3  = y3  ⇔ x + x + x
   
1 2 4 = y3
1 1 1 0 x4 g. y4



x + x + x
1 2 3 = y4
on
⇔

x1 = − 23 y1 + 31 y 2 + 31 y3 + 13 y4
Du

 2 1 1 1 


 −3 3 3 3
x = 1 y − 2 y 2 + 1 y + 1 y 1
− 23 1 1

2 3 1 3 3 3 4 ⇒A−1 = 
3
 
3 3 3 
1 1
 x 3 = 1
y1 + 1
y 2 − 2
y3 + 1
y 4

3 3 − 23 1
3

 3 3 3 3 1 1 1
− 23

x = − 2 y + 1 y 2 + 1 y − 2 y

3 3 3
4 3 1 3 3 3 3 4

Duong T. PHAM September 30, 2020 26 / 28

Cramer’s Rule

    
a11 a12 ... a1n x1 b1
a21 a22 ... a2n  x2  b2 

AM

 ... ...
  =   and det A 6= 0
...   ...   ... 

PH
an1 an2 ... ann xn bn
| {z }
A

T.
The system has a unique solution
g.
xi =
det Ai
, i = 1, 2, . . . , n,
on
det A
Du

where Ai is the matrix formed by replacing the ith column of A by

the column vector b.

Duong T. PHAM September 30, 2020 27 / 28

Cramer’s Rule

x1 − x2 + x3 = 1
 
 1 −1 1
Ex. x1 + x2 − x3 = 1 ; Denote A = 1 1 −1.
1 1 1

x1 + x2 + x3 = 3


AM
We have det A = 4 6= 0 =⇒ The system has a unique solution

PH
   
1 −1 1  1 1 1 
   

T.
1 1 −1 1 1 −1
   
3 1 1 4 1 3 1  4
x1 = = = 1;
g. x2 = = = 1;
det A 4 det A 4
on
Du

 
1 −1 1
 
1 1 1
 
1 1 3 4
x3 = = = 1.
det A 4

Duong T. PHAM September 30, 2020 28 / 28