Topic 3: Determinants

FIN2009 Management Mathematics

Department of Finance
National Taiwan University

Dr. Sandy Lai

1
 A determinant function
The functions that we are familiar with typically have the form:
f(x), with x a real number and the output of the function also a real
number.
However, a function f(X) can take a matrix X as an argument and
still outputs a real number. A matrix’s determinant function is one
such function, which maps a matrix to a real number.
A permutation of a set of integers {1, 2,…, n} is an arrangement
of all the integers in the list without omission or repetitions. A
permutation of {1, 2,…, n} will typically be denoted by (i1, i2, …, in),
where i1 is the first number in the permutation, i2 is the second
number in the permutation, etc.
• Example: List all the permutations of {1, 2,3}
• Solution:
• (1,2,3) (1,3,2) (2,1,3) (3,1,2) (3,2,1)

2
 Permutation and inversion
For the set of integers {1, 2,…, n}, there are have n! permutations.
• E.g., the list {1,2,3,4,5}will have 5!= (5)(4)(3)(2)(1) =120
permutations.
An inversion occurs in the permutation (i1, i2, …, in) whenever a
larger number precedes a smaller number.
• That is, the smaller number is anywhere (need not be
immediately) to the right of the larger number.
A permutation is called even if the number of inversions is even
and odd if the number of inversions is odd.

3
 Example: Number of inversions
Example: Determine the number of inversions in each of the
following permutations.
• (a) (3,1,4,2)
• (b) (1,2,4,3)
• (c) (4,3,2,1)
• (d) (1,2,3,4,5)
• (e) (2,5,4,1,3)
<Solution>
• (a) The permutation (3,1, 4, 2) has a total of 3 inversions.
(3,1,4,2) 2 inversions
(3,1,4,2) 0 inversions
(3,1,4,2) 1 inversion
• (b) There are 0 + 0 +1 = 1 inversions in (1, 2, 4, 3) .
• (c) There are 3 + 2 +1 = 6 inversions in (4, 3, 2,1) .
• (d) There are no inversions in (1, 2, 3, 4, 5) .
• (e) There are 1+ 3 + 2 + 0 = 6 in (2, 5, 4,1, 3) .
4
 Example
Classify as even or odd all the permutations of {1, 2, 3}.

<Solution>

Permutation # Inversions Classification

(1, 2, 3) 0 even
(1, 3, 2) 1 odd
(2,1, 3) 1 odd
(2, 3,1) 2 even
(3,1, 2) 2 even
(3, 2,1) 3 odd

5
 Elementary product of a matrix
A is an n×n matrix. An elementary product from this matrix
will be a product of n entries from A, and none of the entries in the
product can be from the same row or column.
Example: Find all the elementary products for,
• (a) a 2×2 matrix; (b) a 3×3 matrix.
<Solution>
• (a) Each elementary product must have the form: 𝒂𝟏∗ 𝒂𝟐∗ .
Permutation Elementary products
(1, 2) 𝑎 𝑎
(2, 1) 𝑎 𝑎

• (b) Each elementary product must have the form: 𝒂𝟏∗ 𝒂𝟐∗ 𝒂𝟑∗ .
Permutation Elementary products Permutation Elementary products
(1, 2, 3) 𝑎 𝑎 𝑎 (2, 3,1) 𝑎 𝑎 𝑎
(1, 3, 2) 𝑎 𝑎 𝑎 (3,1, 2) 𝑎 𝑎 𝑎
(2,1, 3) 𝑎 𝑎 𝑎 (3, 2,1) 𝑎 𝑎 𝑎
6
 Signed elementary product
For a general an n×n matrix A, it will have n! elementary products
of the form: 𝑎 𝑎 … 𝑎 .
• 𝑖 , 𝑖 ,…, 𝑖 ranges over all the permutations of {1, 2,…, n}.
A signed elementary product from A is the elementary product
𝑎 𝑎 …𝑎 that is multiplied by “+1” if 𝑖 , 𝑖 ,…, 𝑖 is an even
permutation and multiplied by “-1” if otherwise.
Example: Find the signed elementary products for a 3×3 matrix.
<Solution>
Permutation # Inversions Classification Signed elementary
products
(1, 2, 3) 0 even 𝑎 𝑎 𝑎
(1, 3, 2) 1 odd 𝑎 𝑎 𝑎
(2,1, 3) 1 odd 𝑎 𝑎 𝑎
(2, 3,1) 2 even 𝑎 𝑎 𝑎
(3,1, 2) 2 even 𝑎 𝑎 𝑎
(3, 2,1) 3 odd 𝑎 𝑎 𝑎 7
 The determinant of a matrix
Definition 1: If A is square matrix, then the determinant
function is denoted by det. The determinant of A, denoted by
det(A), is the sum of all the signed elementary products of A.
We will sometimes denote determinants as det(A)=|A|. The latter
is most often done with the actual matrix. E.g.,
𝑎 𝑎
det 𝐴 |𝐴| 𝑎 𝑎

The determinant function for a 2×2 matrix:

𝑎 𝑎 𝑎 𝑎
det 𝐴 𝑎 𝑎 𝑎 𝑎 -𝑎 𝑎
𝑎 𝑎

The determinant function for a 3×3 matrix:

𝑎 𝑎 𝑎
𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎
det 𝐴
𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎
𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎
, 𝑎 𝑎 𝑎 𝑎
𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 8
 Example: Determinant
Compute the determinant of each of the following matrices.
3 2
𝑎 𝐴
9 5
3 5 4
𝑏 𝐵 2 1 8
11 1 7

<Solution>

9
 Minor and cofactor
Definition 2: If A is a square matrix, then the minor of aij,
denoted by Mij, is the determinant of the submatrix that results
from removing the ith row and jth column of A.
Definition 3: If A is a square matrix, then the cofactor of aij,
denoted by Cij, is (1)i+j Mij.

Theorem 1: If A is an n× n matrix.
• (a) Choose any row, say row i, then,
det(A)= ai1 Ci1+ ai2 Ci2 +…+ ain Cin
• (b) Choose any column, say column j, then,
det(A)= a1j C1j+ a2j C2j +…+ anj Cnj
• The process of moving across a row or down a column is often
called a cofactor expansion.
<Proof> This follows from the definition of det(A).
10
 Example: Minor and cofactor
For the following matrix compute the cofactors C12 and C24.
4 0 10 4
𝐴 1 2 3 9
5 5 1 6
3 7 1 2
<Solution>
4 0 10 4
1 3 9
1 2 3 9  𝑀 160
5 1 6
5 5 1 6
3 1 2
3 7 1 2

𝐶 1 𝑀 1 160 160

4 0 10 4
4 0 10
1 2 3 9  𝑀 508
5 5 1
5 5 1 6
3 7 1
3 7 1 2

𝐶 1 𝑀 1 508 508

11
 Cofactors
Definition 4: Let A be an n×n matrix and Cij be the cofactor of aij.
The matrix of cofactors from A is,
𝐶 𝐶 … 𝐶
𝐶 𝐶 … 𝐶
⋮ ⋮ ⋱ ⋮
𝐶 𝐶 … 𝐶
The following “sign matrix” summarize the relation between
cofactors and minors.
…
…
…
…
⋮ ⋮ ⋮ ⋮ ⋱

12
 Example: Determinant and cofactors
For the following matrix, compute the determinant by expanding
along the third row.
4 2 1
𝐴 2 6 3
7 5 0

<Solution>
det 𝐴 𝑎 𝐶 𝑎 𝐶 𝑎 𝐶
2 1 4 1 4 2
7 1 5 1 0 1
6 3 2 3 2 6
7 12 5 14 0 20
154

13
 Row operations and determinants
Theorem 2: Let 𝐴 be a square matrix. The following holds
regarding row operations:
• (a) If a multiple of one row of 𝐴 is added to another row to
produce a matrix, 𝐵, then det 𝐵 det 𝐴 .
• (b) If two rows of 𝐴 are interchanged to produce 𝐵, then
det 𝐵 det 𝐴 .
• (c) If one row of 𝐴 is multiplied by 𝑘 to produce 𝐵, then det 𝐵
𝑘 det 𝐴 .

Implication:
• det 𝐸𝐴 det 𝐸 det 𝐴

14
 Proof of Theorem 2
The proof is by induction on the size of 𝐴. The case of a 22 matrix
was verified in Exercises 33–36 of Section 3.1.
Suppose the theorem has been verified for determinants of kk
matrices with k 2. Let n=k+1, and let 𝐴 be nn. The action of 𝐸 on 𝐴
involves either two rows or only one row.
We can expand det 𝐸𝐴 across a row that is unchanged by the action
of 𝐸, say, row i . Let 𝐴 (respectively, 𝐵 ) be the matrix obtained by
deleting row i and column j from 𝐴 (respectively, 𝐸𝐴). Then the rows
of 𝐵 are obtained from the rows of 𝐴 by the same type of
elementary row operation that 𝐸 performs on 𝐴.
Since these submatrices are only kk, the induction assumption
implies that det 𝐵 𝛼 det 𝐴 , where 𝛼 1, 1, 𝑜𝑟 𝑘, depending on
the nature of 𝐸. For det 𝐸𝐴 , the cofactor expansion across row i is
det 𝐸𝐴 𝑎 1 det 𝐵 ⋯ 𝑎 1 det 𝐵
𝛼𝑎 1 det 𝐵 ⋯ 𝛼𝑎 1 det 𝐵 𝛼 det 𝐴
15
 Properties of Determinants
Theorem 3: Let A be an n×n matrix and c be a scalar then. Then,
det 𝑐𝐴 𝑐 det 𝐴 .
<Proof> This follows from Theorem 2.

Theorem 4: Suppose that A, B, and C are all n×n matrices and that
they differ by only one row, say the kth row. Let’s further suppose that
the kth row of C can be found by adding the corresponding entries
from the kth rows of A and B. Then, det 𝐶 det 𝐴 det 𝐵
• The theorem still holds if we replace the word ‘row’ with ‘column.’
<Proof> This follows from Definition 1.

Theorem 5: If A and B are matrices of the same size, then det 𝐴𝐵

det 𝐴 det 𝐵
• Likewise, det 𝐴𝐵𝐶 det 𝐴 det 𝐵 det 𝐶 .

16
 Proof of Theorem 5
If A is not invertible, then neither is AB. We can obtain the proof.
• Suppose AB is invertible, let W be the inverse of AB. Then,
ABW = I. ⟹ A(BW) = I. Since A is square, A is invertible.
• In this case, det 𝐴𝐵 det 𝐴 0. Therefore, det 𝐴𝐵
det 𝐴 det 𝐵 .
If A is invertible, it must be row equivalent to 𝐼 . So, there exist
elementary matrices, 𝐸 , 𝐸 , … , 𝐸 such that 𝐴 𝐸 … 𝐸 𝐸 𝐼
𝐸 …𝐸 𝐸 .
• det 𝐴𝐵 det 𝐸 … 𝐸 𝐸 𝐵 det 𝐸 det 𝐸 …𝐸 𝐸 𝐵 …
det 𝐸 det 𝐸 … det 𝐸 det 𝐸 det 𝐵 ……(Eq.1)
• (Eq.1) det 𝐸 𝐸 det 𝐸 … det 𝐸 det 𝐸 det 𝐵 ⋯
det 𝐸 … 𝐸 𝐸 det 𝐵 det 𝐴 det 𝐵

17
 Example: Determinant
Compute det(A), det(B) and det(A+B) for the following matrices.
10 6 1 2
𝐴 𝐵
3 1 5 6

<Solution>

18
 Example: Determinant
Compute the determinants for the following three matrices.
4 2 1 4 2 1 4 2 1
𝐴 6 1 7 𝐵 2 5 3 𝐶 4 4 10
1 3 9 1 3 9 1 3 9

<Solution>

19
 Properties of determinants
Theorem 6: Suppose that A is an invertible matrix. Then,

• det 𝐴

Theorem 7: Suppose that A is an n×n triangular matrix. Then,

• det 𝐴 𝑎 𝑎 …𝑎

20
 Properties of determinants
Theorem 8: Matrix A is invertible if and only if det 𝐴 0.

Theorem 9: If A is a square matrix, then det 𝐴 det 𝐴 .

Theorem 10: If A is a square matrix with a row or column of all

zeroes, then det 𝐴 0.

Theorem 11: Suppose that A is a square matrix and that two of

its rows are proportional or two of its columns are proportional.
Then det 𝐴 0.

21
 Proof of Theorem 8
Suppose a square matrix A has been reduced to an echelon form U
by row operations (a)-(c):
• (a) Add c times a row to another row (m times).
• (b) Interchange any two rows (r times).
• (c) Multiply a row by a non-zero scalar 𝑘 , another row by 𝑘 ,
another row by 𝑘 , …, another row by 𝑘 .
Based on Theorem 2, we have:
• det 𝐴 1 𝑘 𝑘 …𝑘 det 𝑈 .
Since 𝑈 is in row echelon form, it is triangular. So, det 𝑈 is the
product of the diagonal entries 𝑢 𝑢 … 𝑢 .
• If 𝐴 is invertible, it is row equivalent to an identity matrix. In
this case, the entries 𝑢 , 𝑢 , … , 𝑢 are all leading 1’s. So,
det 𝑈 1. Therefore, det 𝐴 1 𝑘 𝑘 …𝑘 0.
• If 𝐴 is not invertible, at least 𝑢 is zero. Thus, det 𝑈 0.
Therefore, det 𝐴 0. 22
 Proof of Theorem 9
Assume 𝐴 is a nn matrix.
The theorem is obvious for n=1. Suppose the theorem is true for
kk determinants.
Let n=k+1. The cofactor of 𝑎 in A equals the cofactor of 𝑎 in 𝐴
(because the cofactors involve kk determinants). Hence, the
cofactor expansion of det 𝐴 along the first row equals the cofactor
expansion of det 𝐴 down the first column. That is, 𝐴 and 𝐴 have
equal determinants. The theorem is true for n=k+1.
By the Principle of Mathematical Induction, the theorem is true
for all 𝑛 1.

23
 Compute determinants via row reduction
How to compute determinants via row reduction?
• Step 1: Use elementary row operations to reduce the matrix to
row-echelon form.
• Step 2: The row-echelon form of a square matrix is an upper
triangular matrix, for which the determinant can be easily
calculated.
• Step 3: Use Theorem 1 to compute the determinant of the
original matrix.

24
Example: Compute determinants via row reduction
Use row reduction to compute the determinant of the following
matrix.
2 10 2
𝐴 1 0 7
0 3 5
<Solution>

2 10 2 1 0 7
𝑅 ↔𝑅
det 𝐴 1 0 7 2 10 2
0 3 5 0 3 5

1 0 7 1 0 7
𝑅 2𝑅 𝑅
0 10 16 10 0 1
0 3 5 0 3 5

1 0 7 1 0 7
𝑅 3𝑅 𝑅
10 0 1 10 0 1 10 98
0 0 0 0 1 25
 Cramer’s Rule
Theorem 12: Suppose that A is an n×n invertible matrix. Then,
the solution to the system Ax = b is given by,

𝑥 ,𝑥 ,…,𝑥

• where Ai is the matrix found by replacing the ith column of A

with b.

Remarks: If the system is too large, it will probably be quicker to

use one of the other methods that we’ve got for solving systems of
equations.

26
 Proof of Theorem 12 (Cramer’s Rule)
Denote the columns of A by 𝐚 , 𝐚 , … , 𝐚 and the columns of the 𝑛 𝑛
identity matrix I by 𝐞 , 𝐞 , … , 𝐞 .
Let Ai be the matrix found by replacing the ith column of A with b.
Let 𝐼 𝐱 be the matrix found by replacing the ith column of I with x.

𝐴·𝐼 𝐱 𝐴 𝐞 … 𝐱…𝐞 𝐴𝐞 … 𝐴𝐱 … 𝐴𝐞 𝐚 …b…𝐚 𝐴.

det 𝐴 det 𝐼 𝐱 det 𝐴
det 𝐼 𝐱 = product of the main diagonal of 𝐼 𝐱 =𝑥 .

Hence, 𝑥

27
 Example: Cramer’s rule
Use Cramer’s Rule to determine the solution to the following
system of equations.
3𝑥 𝑥 5𝑥 2
4𝑥 𝑥 7𝑥 10
2𝑥 4𝑥 𝑥 3
<Solution>

3 1 5 𝑥 2
4 1 7 𝑥 10
2 4 1 𝑥 3
𝐴 𝐱 𝐛

𝑑𝑒𝑡 𝐴 187 0

2 1 5
det 𝐴 10 1 7 212
3 4 1 28
 3 2 5
det 𝐴 4 10 7 273
2 3 1
3 1 2
det 𝐴 4 1 10 107
2 4 3

𝑥 𝑥 𝑥

𝑥 𝑥 𝑥

29
 Adjoint (or Adjugate)
Definition 5: The adjoint of A is the transpose of the matrix of
cofactors and is denoted by adj(A).
• The adjoint of A is also called the adjugate of A.
For a matrix A, the minor of aij is Mij, and the cofactor of aij is Cij.
Cij = (1)i+j Mij.
The matrix of cofactors from A is
𝐶 𝐶 … 𝐶
𝐶 𝐶 … 𝐶
⋮ ⋮ ⋱ ⋮
𝐶 𝐶 … 𝐶
The adjoint of A is
𝐶 𝐶 … 𝐶
𝐶 𝐶 … 𝐶
⋮ ⋮ ⋱ ⋮
𝐶 𝐶 … 𝐶

30
 Example: adjoint
Compute the adjoint of the following matrix.
4 2 1
𝐴 2 6 3
7 5 0
<Solution>
2 1
𝐶 1 1 5 5
5 0
4 2
𝐶 1 1 34 34
7 5

𝐶 15 𝐶 21 𝐶 52 𝐶 7
𝐶 12 𝐶 14 𝐶 20

15 21 52
• The matrix of cofactors from A = 5 7 34
15 5 12
12 14 20
𝑎𝑑𝑗 𝐴 21 7 14
52 34 20
31
 Inverse and adjoint
Theorem 13: If A is an invertible matrix, then
1
𝐴 𝑎𝑑𝑗 𝐴
det 𝐴

𝐶 𝐶 … 𝐶
𝐶 𝐶 … 𝐶
• 𝑎𝑑𝑗 𝐴
⋮ ⋮ ⋱ ⋮
𝐶 𝐶 … 𝐶

32
 Proof of Theorem 13
Cramer’s rule says: If Ax = b, then 𝑥 , where Ai is the
matrix found by replacing the ith column of A with b.
We know that 𝐴𝐴 𝐼. So, the jth column of 𝐴 is a vector y that
satisfies 𝐴𝐲 𝐞 , where 𝐞 is the jth column of the identify matrix.
Let 𝐴 𝐞 be the matrix found by replacing the ith column of A
with 𝐞 .
𝐞
By Cramer’s rule, 𝑦 .
Recall 𝐴 denotes the submatrix of A formed by deleting row j and
column i. A cofactor expansion down column i of 𝐴 𝐞 show that
det 𝐴 𝐞 1 det 𝐴 𝐶

That is, the (i,j) entry of 𝐴 is .

33
 Example: Inverse and adjoint
Use the adjoint matrix to compute the inverse of the following
matrix.
4 2 1
𝐴 2 6 3
7 5 0
<Solution>
det 𝐴 154

15 5 12
𝑎𝑑𝑗 𝐴 21 7 14
52 34 20

15 5 12
𝐴 21 7 14
52 34 20

34
 Inverse Matrices (Summary)
Theorem 14: If A is an n× n matrix then the following statements
are equivalent.
• (a) A is invertible.
• (b) The only solution to the system Ax = 0 is the trivial
solution.
• (c) A is row equivalent to In .
• (d) A is expressible as a product of elementary matrices.
• (e) Ax =b has exactly one solution for every n×1 matrix b.
• (f) Ax =b is consistent for every n×1 matrix b.
• (g) det 𝐴 0.

35