Chapter 5 Determinants (Soham Bhosale).ipynb

Kernel: SageMath 10.6

Chapter 5: Determinants

By: Soham Bhosale

The determinant of a matrix is a single numerical value which is used when calculating the inverse or when solving systems of linear equations. Let's look at some basic examples first.

A = Matrix(ZZ, [[-3,1,2],[5,5,-8],[4,2,-5]])
A

[-3  1  2]
[ 5  5 -8]
[ 4  2 -5]

A.det()

0

B = Matrix(ZZ, [[4, -1, 1, 0],[1, 7, 3, 5],[0,-3, -2, 1], [2, 4, 8, -1]])
B

[ 4 -1  1  0]
[ 1  7  3  5]
[ 0 -3 -2  1]
[ 2  4  8 -1]

B.det()

-413

The determinants of 2x2 matrices can be generated using a special formula

var('a11', 'a12', 'a21', 'a22')
A=matrix(2, 2, [a11, a12, a21, a22])
print(A)
print("det of 2 x 2 ", det(A))

[a11 a12]
[a21 a22]
det of 2 x 2  -a12*a21 + a11*a22

Let's apply to an example matrix

A= random_matrix(ZZ, 2, 2)
print(A)
print("Sage determined determinant of A" , det(A))
d = A[0,0]*A[1,1]-A[0,1]*A[1,0]
print("Determinant calculated by formula " , d)

[  2  -1]
[ -2 -11]
Sage determined determinant of A -24
Determinant calculated by formula  -24

Now we are going to look at some interesting properties of Determinants.

The first property of determinants we are going to look at is regarding multiplying determinants. Multiplying matrices and computing arithmetic operations requires a lot of computation. However, the relationship between the determinant of the product of matrices and the product of their individual determinants, it more simple.

$det(A*B) = det(A) * det(B)$

And

$det(B*A) = det(B) * det(A)$

C=random_matrix(ZZ,4,4)
C

[ -1   3   0 -29]
[  1  -1  -1   7]
[  0   0  -1  -2]
[  1  -3   1  -1]

D=random_matrix(ZZ,4,4)
D

[-2 -1  0 -8]
[ 1 -4 27  1]
[ 2  0 -1 -4]
[ 4 -1  3 -3]

# Multiplying the two matrices
C*D

[-111   18   -6   98]
[  23   -4   -5  -26]
[ -10    2   -5   10]
[  -7   12  -85  -12]

# Determinant of the product of the two matrices
det(C*D)

29376

# Product of the determinants
det(C) * det(D)

29376

This property holds commutatively as well!

D*C

[  -7   19   -7   59]
[  -4    4  -22 -112]
[  -6   18   -3  -52]
[  -8   22   -5 -126]

det(D*C)

29376

det(D) * det(C)

29376

After seeing this neat property of multiplying determinants, it may be tempting to assume that adding determinants work the same way. Namely, that the determinant of the sum of two matrices is the same as the sum of their individual determinants. In math notation, $det(A+B) = det(A)+det(B)$. However, this property is not true. Here's an example of when it is not true.

print("det(C+D) = ", det(C + D))
print("det(C) + det(D) = ", det(C) + det(D))

det(C+D) =  -5430
det(C) + det(D) =  -523

However, there are some instances where this property does hold true. Here is an example:

Determinants of matrices can be calculated using something called the cofactor expansion method.

Here is the formula:

$det(A) = a_{11}C_{11} + a_{12}C_{12} + ... a_{1n}C_{1n}$

Where $C_{11}...C_{1n}$ are the cofactors of $a_{11}...a_{1n}$.

We can represent this using an example to see how sage calculates it

var('a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i')
A=matrix(3, 3, [a, b, c, d, e, f, g, h, i])
A

[a b c]
[d e f]
[g h i]

A.det()

-(f*h - e*i)*a + (c*h - b*i)*d - (c*e - b*f)*g

Elementary Matrices

Instead of using cofactor expansion to compute the determinant. It is more efficient to convert the matrix to echleon form using row operations and then multiply the terms along the diagnonal

First, let's see how elementary matrices change a matrix (row-operations)

Remember the three elementary operations are

1. Interchanging two rows

2. Multiplying a row by a constant c

3. Adding a mulitple of one row to another

# Matrix representing adding a multiple of row 3 to row 2
E = elementary_matrix(3, row1 = 1, row2 = 2, scale = 4)
E

[1 0 0]
[0 1 4]
[0 0 1]

#Effect of multiplying elementary matrix E by A
E*A

[      a       b       c]
[d + 4*g e + 4*h f + 4*i]
[      g       h       i]

# Order of multiplication matters in linear algebra or else we get a completely different effect
A*E

[      a       b 4*b + c]
[      d       e 4*e + f]
[      g       h 4*h + i]

Affect of Elementary Matrices on the Determinant

Here is the affect of elementary matrices on the Determinant. This table from the textbook does a great job summarizing it.

#Adding a muliple of one row to the other does not change the determinant
B=random_matrix(ZZ, 3, 3)
print(B)
print("Determinant Before", det(B))
E*B
print("Determinant After Multiplying by elementary matrix", det(E*B))

[176  18  -1]
[ -1  -1  -2]
[  1  -1  -2]
Determinant Before -74
Determinant After Multiplying by elementary matrix -74

Elementary Matrix Switching Rows

The determinant of the matrix produced by interchanging two rows of the original matrix is -det(B)

F=elementary_matrix(3, row1=0, row2=2)
print(F)

[0 0 1]
[0 1 0]
[1 0 0]

print(B)
print("Determinant Before", det(B))
print('Matrix after row operation')
print(F*B)
print("Determinant After Multiplying by elementary matrix", det(F*B))

[176  18  -1]
[ -1  -1  -2]
[  1  -1  -2]
Determinant Before -74
Matrix after row operation
[  1  -1  -2]
[ -1  -1  -2]
[176  18  -1]
Determinant After Multiplying by elementary matrix 74

Multiplying a row by a constant multiplies the determinant by that constnant c

G=elementary_matrix(3, row1=2, scale=5)
print(G)

[1 0 0]
[0 1 0]
[0 0 5]

print(B)
print("Determinant Before", det(B))
print('Matrix after row operation')
print(G*B)
print("Determinant After Multiplying by elementary matrix", det(G*B))

[176  18  -1]
[ -1  -1  -2]
[  1  -1  -2]
Determinant Before -74
Matrix after row operation
[176  18  -1]
[ -1  -1  -2]
[  5  -5 -10]
Determinant After Multiplying by elementary matrix -370

Now here is a function I wrote that represents the recursive nature of finding determinants using cofactor expansion

def myDet(M):
    if M.ncols() != M.nrows():
        return("not a square matrix so no determinant")
    else: 
        col=M.ncols()
        if col==1: 
            return M[0,0]
        else:
            D = 0
            N=M.matrix_from_columns(range(1,col))
            for i in range(col):
                D=D + (-1)^i * M[i,0] * myDet(N.matrix_from_rows(list(range(i)) + list(range(i+1, col))))
            return D

A=random_matrix(ZZ, 5,5)

det(A)

147

myDet(A)

147

Some more properties!

A = Matrix(ZZ, [[3,1,0],[-1,2,1],[0,-1,2]])
print(A)

[ 3  1  0]
[-1  2  1]
[ 0 -1  2]

Determinant of the inverse of a matrix A is $\dfrac{1}{det(A)}$

B= A.inverse()
print("A inverse")
print(B)
print("Determinant of matrix A", det(A))
print("Determinant of A inverse", det(B))
print("It can be seen that the determinant of the inverse is 1 over the determinant of the original matrix")

A inverse
[ 5/17 -2/17  1/17]
[ 2/17  6/17 -3/17]
[ 1/17  3/17  7/17]
Determinant of matrix A 17
Determinant of A inverse 1/17
It can be seen that the determinant of the inverse is 1 over the determinant of the original matrix

$A^{T}$has the same Determinant As A

print(A)
B= A.transpose()
print("Transposed Matrix")
print(B)
print("Original Determinant", det(A))
print("Determinant of Transposed Matrix", det(B))

[ 3  1  0]
[-1  2  1]
[ 0 -1  2]
Transposed Matrix
[ 3 -1  0]
[ 1  2 -1]
[ 0  1  2]
Original Determinant 17
Determinant of Transposed Matrix 17

Inverses from Determinants

In the textbook, we learn that If A is an invertible matrix, then $A^{-1} = \dfrac{1}{det(A)} * adj(A)$

A=random_matrix(ZZ, 5,5)
print(A)
C=A.adjugate()
print("Adjugate matrix for A")
print(C)

[-8  0  1 -1 56]
[ 2  0 -1  0  0]
[ 1 -2 -2  1  1]
[-3 -1  1  1  7]
[ 1  5  2  0 -3]
Adjugate matrix for A
[  33 -732  312 -279   69]
[ -36  474 -294  258 -168]
[  66 -954  624 -558  138]
[  32 -138 -192 -286 -134]
[  -5  -90   30  -35    5]

D= 1/(det(A)) * C
D

[-11/170  122/85  -52/85  93/170 -23/170]
[   6/85  -79/85   49/85  -43/85   28/85]
[ -11/85  159/85 -104/85   93/85  -23/85]
[-16/255   23/85   32/85 143/255  67/255]
[  1/102    3/17   -1/17   7/102  -1/102]

# We see that the formula holds true
E = A.inverse()
E

[-11/170  122/85  -52/85  93/170 -23/170]
[   6/85  -79/85   49/85  -43/85   28/85]
[ -11/85  159/85 -104/85   93/85  -23/85]
[-16/255   23/85   32/85 143/255  67/255]
[  1/102    3/17   -1/17   7/102  -1/102]

E== D

True

Cramer's Rule

Cramer's rule is a formula using determinants that can be used to find the solution to Ax = b where A is an n x n matrix and b is a vector.

If A = [$a_1, a_2,...a_n$] and b is in $R^n$, then let $A_i$ denote the matrix A after replace $a_i$ with b

def mysolve(A, b):
    
    if A.nrows()!=A.ncols():
        return "first matrix is not a square"
    else:
        if A.det() == 0:
            return "first matrix has determinant zero so Cramer's rule does not apply"
        else:
            if b.nrows()!=A.ncols() or b.ncols()!=1:
                return "second matrix is not a column matrix of the right side"
            else:
                n=A.ncols()
                x=matrix(QQ, n, 1)
                for j in range(n):
                    Aj=copy(A)
                    Aj[:,j]=b
                    x[j] = Aj.det()/A.det()
                return x

A=random_matrix(ZZ, 3, 3)
print(A)
print("Determinant of A", A.det())
b=matrix(QQ, 3, 1, [5, 3, 7])
print(b)
print("Using the Cramer Rule Calculator We get x=")
print(mysolve(A,b))

[ 2  0  1]
[ 5 -1 10]
[ 0  1  0]
Determinant of A -15
[5]
[3]
[7]
Using the Cramer Rule Calculator We get x=
[ 8/3]
[   7]
[-1/3]

print("Using in built in sage functionality we get: ") 
print(A.solve_right(b))

Using in built in sage functionality we get: 
[ 8/3]
[   7]
[-1/3]

Determinants Can Also Represent Area

Let T:

Example Problems

A functionality that computers provide is the ability to compute fast. Thus we can use this to our advantage to find examples of certain examples or counterxamples of matrices. Here are some practice problems from the book representing this

# Exercise 4 (c): Fill in entries for A and B to find a counterexample.
A = Matrix(ZZ, [[_, _, _], [_, _, _], [_, _, _]])
B = Matrix(ZZ, [[_, _, _], [_, _, _], [_, _, _]])

A.determinant() - B.determinant() == (A-B).determinant()

True

#5.2 Question 41
for a in range(-10, 10):
    for b in range(-10, 10):
        for c in range(-10, 10):
            for d in range(-10, 10):
                A = Matrix(ZZ, [[a,b],[c,d]])
                # Skip the all zeroes matrix.
                if a == 0 and b == 0 and c == 0 and d == 0:
                    continue
                # In order for this to hold, we need det(A) == 0.
                if 3*det(A) == det(3*A) and det(A) != 0:
                    print("Example found!\n", A)

def tri_diagonal(n):
    m = matrix(n)
    for j in range(min(2,n)):
        m[0,j]=1
    for i in range(1,n-1):
        for j in range(i-1,i+2):
            m[i,j]=1
    for j in range(1,1+min(2,n)):
        m[n-1,n-j]=1   
    return(m)            

tri_diagonal(4)

[1 1 0 0]
[1 1 1 0]
[0 1 1 1]
[0 0 1 1]

[det(tri_diagonal(n)) for n in range(8)]

[1, 1, 0, -1, -1, 0, 1, 1]

det(tri_diagonal(4))

-1