You are on page 1of 12

≻≻≻≻—≻≻≻≻—Lecture Seven —≺≺≺≺—≺≺≺≺

Chapter Three: Properties of Determinant


 
Recall. Let A = aij be an n × n matrix. We define the determinant of A by det(A)=
|A| = Σσ∈Sn ǫ(σ)a1σ(1)a2σ(2) · · · anσ(n), where Sn is the collection of all permutations of
 
a11 a12
Jn = {1, 2, . . . , n}. For example, for A2×2 = , det(A)= a11a22 − a12a21 and
  a21 a22
a11 a12 a13
for A3×3 = a21 a22 a23, det(A)= a11a22a33 − a12a21a33 − a13a22a31 − a11a23a32+
 
a31 a32 a33
a12a23a31 + a13a21a32.

Linear Algebra lecture 7−1


Theorem 3.1.1. The determinants of a matrix and its transpose are equal, that is,
det(A)=det(AT ).

◦ T
   
Proof. 1 Let An×n = aij and A = Bn×n = bij .
Then bij = aji for i, j = 1, 2, . . . , n.
Denote Sn the collection of all permutations of Jn = {1, 2, . . . , n}.
Note the set Tn = {σ −1 : σ ∈ Sn} is the same as Sn
since size of Tn equals the size of Sn and Tn ⊆ Sn.
2◦ By definition of determinant, ǫ(σ −1) = ǫ(σ) ∀ σ ∈ Sn and Sn = Tn,
det(B) = Σσ∈Sn ǫ(σ)b1σ(1)b2σ(2) · · · bnσ(n)
= Σσ∈Sn ǫ(σ)aσ(1)1aσ(2)2 · · · aσ(n)n
= Σσ∈Sn ǫ(σ)aσ(1)σ−1(σ(1))aσ(2)σ−1(σ(2)) · · · aσ(n)σ−1(σ(n))
ǫ(σ −1 )=ǫ(σ)
= Σσ−1∈Tn ǫ(σ −1)a1σ−1(1)a2σ−1(2) · · · anσ−1(n)
Tn =Sn
= Σσ∈Sn ǫ(σ)a1σ(1)a2σ(2) · · · anσ(n) = det (A).
Linear Algebra
This theorem is proved. lecture 7−2
Theorem 3.1.5. If B is obtained from A by multiplying a row (column) of A by a
real number c, then det(B)= c det(A).

Note 3.1.ET3. If A is an n × n matrix and c is a real number, then det(cA)= cn det(A).

Linear Algebra lecture 7−3



   
Proof. 1 Suppose Bn×n = bij results from An×n = aij
( of A by c ∈ R.
by multiplying rth row
c arj if i = r
Then we have bij = .
aij if i 6= r
So det(B) = Σσ∈Sn ǫ(σ)b1σ(1) · · · brσ(r) · · · bnσ(n)
= Σσ∈Sn ǫ(σ)a1σ(1) · · · (carσ(r)) · · · anσ(n)
= c Σσ∈Sn ǫ(σ)a1σ(1) · · · arσ(r) · · · anσ(n)
= c det(A).

2◦ Now suppose that B is obtained from A


by multiplying rth column of A by c.
Then B T is obtained from AT by by multiplying rth row of AT by c.
So, along with Theorem 3.1.1, we have
det(B)=det(B T )= c det(AT )= c det(A).

Linear Algebra lecture 7−4


Theorem 3.1.2. If matrix B results from matrix A by interchanging two rows (columns)
of A, then det(B)= −det(A)


   
proof. 1 Suppose Bn×n = bij results from An×n = aij
by interchanging rthand sth rows of A, 1 ≤ r < s ≤ n.
 asj if i = r

Then we have bij = arj if i = s .

aij if i 6= r, s

Linear Algebra lecture 7−5



s
 if i = r
proof. 2◦ Let τ : Jn = {1, 2, . . . , n} −→ Jn defined by τ (i) = r if i = s .

i otherwise

Note τ is a permutation and has the inverse itself.
Moreover, ǫ(σ ◦ τ ) = −ǫ(σ) for all σ ∈ Sn since τ is a transposition.
Consider the set Tn ≡ {σ ◦ τ : σ ∈ Sn} ⊆ Sn.
Since σ1 ◦ τ = σ2 ◦ τ if and only if σ1 = σ2 beacuse τ (τ (i)) = i for i = 1, . . . , n.
It follows |Tn| = |Sn| = n! and hence Tn = Sn.
3◦ det(B) = Σσ∈Sn ǫ(σ)b1σ(1) · · · brσ(r) · · · bsσ(s) · · · bnσ(n) .
= Σσ∈Sn ǫ(σ)a1σ(1) · · · asσ(r) · · · arσ(s) · · · anσ(n)
= Σσ∈Sn ǫ(σ)a1σ(1) · · · arσ(s) · · · asσ(r) · · · anσ(n)
= Σσ∈Sn − ǫ(σ ◦ τ )a1σ(τ (1)) · · · arσ(τ (r)) · · · asσ(τ (s)) · · · anσ(τ (n))
= Σσ∈Tn − ǫ(σ)a1σ(1) · · · arσ(r) · · · asσ(s) · · · anσ(n)
= −Σσ∈Sn ǫ(σ)a1σ(1) · · · arσ(r) · · · asσ(s) · · · anσ(n) = −det(A)
4◦ Now if B is obtained from A by interchanging two columns of A,
then B T is obtained from AT by interchanging two rows of AT .
So, along with Theorem 3.1.1, det(B)=det(B T )= −det(AT )= −det(A).

Linear Algebra lecture 7−6


Corollary 3.1.3. If two rows (columns) of A are equal, then det(A)= 0.

Proof. 1◦ Suppose that rows r and s of A are equal.


Interchanging rows r and s of A to obtain the matrix B,
one gets B = A. Hence det(B)=det(A).
On the other hand, Theorem 3.1.2 implies that det(B)= −det(A).
We have det(A)= −det(A), so det(A)= 0.

Linear Algebra lecture 7−7


    
Fact. Consider three n × n matrix An×n = aij , Bn×n = bij and Cn×n = cij so
that Rowr (C) =Rowr (A)+Rowr (B) and Rowi(C) =Rowi(A) =Rowi(B) for i 6= r. Then
det(C)=det(A)+det(B).

Proof. 1◦ Since Rowr (C) =Rowr (A)+Rowr (B),


we have crj = arj + brj for 1 ≤ j ≤ n.
Since Rowi(C) =Rowi(A) =Rowi(B) for i 6= r,
we obtain cij = aij = bij for i 6= r and j = 1, . . . , n. Thus,
det(C) = Σσ∈Sn ǫ(σ)c1σ(1) · · · crσ(r) · · · cnσ(n)
= Σσ∈Sn ǫ(σ)c1σ(1) · · · (arσ(r) + brσ(r)) · · · cnσ(n)
= Σσ∈Sn ǫ(σ)[(c1σ(1) · · · arσ(r) · · · cnσ(n)) + (c1σ(1) · · · brσ(r) · · · cnσ(n))]
= Σσ∈Sn ǫ(σ)a1σ(1) · · · arσ(r) · · · anσ(n) + Σσ∈Sn ǫ(σ)b1σ(1) · · · brσ(r) · · · bnσ(n)
= det(A) + det(B).
We prove this fact.

Linear Algebra lecture 7−8


Theorem 3.1.6. If B is obtained from A by adding to each element of the sth row
(column) of A a constant c times the corresponding element of the rth row (column),
r 6= s, of A, then det(B)=det(A).

Proof. 1◦ We prove the theorem for rows. Since B is obtained from A


by adding the sth row of A c times of the rth row of A,
We have Rows(B) =Rows(A) + c Rowr (A)
and Rowi(B) =Rowi(A) for i 6= s.
2◦ We let C be obtained from A by replacing the sth row of A by Rowr (A).
Note C has equal r and s rows and hence det(C)= 0 by Corollary 3.1.3 .
Next let D be obtained from C by multiplying the sth row by c.
Then Theorem 3.1.5 says det(D)= c det(C)= c 0 = 0.
3◦ Now back to B, we have Rows(B) =Rows(A)+Rows(D)
and Rowi(B) =Rowi(A) =Rowi(D) for i 6= s.
The previous fact gives det(B)=det(A)+det(D).
It follows det(B)=det(A) immediately.

Linear Algebra lecture 7−9


Theorem 3.1.7. If a matrix is upper (lower) triangular, then its determinant is the
product of the elements on the main diagonal.


 
Proof. 1 Let An×n = aij be an upper triangular. Then aij = 0 for i > j.
Thus, a term a1σ(1)a2σ(2) · · · anσ(n) in the expression for det(A)
is zero if 1 > σ(1), 2 > σ(2), . . . , n > σ(n).
Note the only permutation not satisfying the above n inequality
is the identity permutation;
namely, σ(n) = n, σ(n − 1) = n − 1, . . . , σ(1) = 1.
Thus det(A)= a11a22 · · · ann.

2 If An×n = aij is lower triangular, then AT is upper triangular.
 

It follows 1◦ and Theorem 3.1.1 immediately that


det(A)=det(AT )= a11a22 · · · ann.
Linear Algebra lecture 7−10
Corollary 3.1.1. The determinant of a diagonal matrix is the product of the entries on
its main diagonal.

Proof. This corollary simply follows Theorem 3.1.7


and the fact that a diagonal matrix is also upper triangular.

Example. The determinant of an identity matrix is 1, i.e., det(In)= 1.

Remark. Let E be an elementary matrix.


(a) If E is the elementary matrix of rows interchange, then det(E)= −1.
(b) If E is the elementary matrix of multiplying a row by c, then det(E)= c.
(c) If E is the elementary matrix of adding c times of a row to another row, then det(E)= 1.

Proof. Skip.

Linear Algebra lecture 7−11


♣ Determinant computation (via reduction to triangular (row echelon) form): we can com-
pute det(A) by reducing A to a triangular matrix via appropriate elementary row (col-
umn) operations.
 
4 3 2
Example 3.1.17. Let A = 3 −2 5. Please compute det(A).
 
2 4 6
4 3 2 ← −
"
4 3 2
# ! " # !
det(A) = det 3 −2 5 = 2 det 3 −2 5
2 4 6 × 12 1 2 3 ← −
" # ! " # !
1 2 3 ×(−3) ×(−4) 1 2 3
= −2 det 3 −2 5 ← −+ = −2 det 0 −8 −4
4 3 2 ←−−−−− + 0 −5 −10 ×(− 51 )
" # ! " # !
1 2 3 1 2 3
= 10 det 0 −8 −4 ← − = −10 det 0 1 2 ×8
0 1 2 ← − 0 −8 −4 ← −+
"
1 2 3
# !
= −10 det 0 1 2 = (−10)(1)(1)(12) = −120.
0 0 12

Linear Algebra lecture 7−12

You might also like