Download MATH 304 Linear Algebra Lecture 7: Evaluation of determinants

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
MATH 304
Linear Algebra
Lecture 7:
Evaluation of determinants.
The Vandermonde determinant.
Cramer’s rule.
Determinants
Determinant is a scalar assigned to each square matrix.
Notation. The determinant of a matrix
A = (aij )1≤i,j≤n is denoted det A or
a a ... a 1n 11 12
a a ... a 2n 21 22
.
..
.
.
.
.
. ... .
.
an1 an2 . . . ann Principal property: det A 6= 0 if and only if a
system of linear equations with the coefficient
matrix A has a unique solution. Equivalently,
det A 6= 0 if and only if the matrix A is invertible.
Explicit definition in low dimensions
a b
= ad − bc,
Definition. det (a) = a, c d
a11 a12 a13 a21 a22 a23 = a11 a22 a33 + a12 a23 a31 + a13 a21 a32 −
a31 a32 a33 −a13 a22 a31 − a12 a21 a33 − a11 a23 a32 .
+:
−:

*

∗
∗

∗

∗
*
∗
*
∗
∗
*
∗
 
∗
∗
 
∗ ,  ∗
*
*
 
*
∗
 
∗ ,  *
∗
∗
*
∗
∗
*
∗
∗
 
∗
∗
 
* ,  *
∗
∗
 
∗
*
 
∗ ,  ∗
*
∗
∗
∗
*
∗
∗
*

*

∗ .
∗

∗

* .
∗
Properties of determinants
Determinants and elementary row operations:
• if a row of a matrix is multiplied by a scalar r ,
the determinant is also multiplied by r ;
• if we add a row of a matrix multiplied by a scalar
to another row, the determinant remains the same;
• if we interchange two rows of a matrix, the
determinant changes its sign.
Properties of determinants
Tests for singularity:
• if a matrix A has a zero row then det A = 0;
• if a matrix A has two identical rows then
det A = 0;
• if a matrix has two proportional rows then
det A = 0.
Properties of determinants
Special matrices:
• det I = 1;
• the determinant of a diagonal matrix is equal to
the product of its diagonal entries;
• the determinant of an upper triangular matrix is
equal to the product of its diagonal entries.
Properties of determinants
Determinant of the transpose:
• If A is a square matrix then det AT = det A.
Columns vs. rows:
• if one column of a matrix is multiplied by a
scalar, the determinant is multiplied by the same
scalar;
• adding a scalar multiple of one column to
another does not change the determinant;
• interchanging two columns of a matrix changes
the sign of its determinant;
• if a matrix A has a zero column or two
proportional columns then det A = 0.
Row and column expansions
Given an n×n matrix A = (aij ), let Mij denote the
(n − 1)×(n − 1) submatrix obtained by deleting the
ith row and the jth column of A.
Theorem For any 1 ≤ k, m ≤ n we have that
n
X
det A =
(−1)k+j akj det Mkj ,
j=1
(expansion by kth row )
det A =
n
X
(−1)i+m aim det Mim .
i=1
(expansion by mth column)
Signs for row/column expansions

+
−

+


−
..
.
−
+
−
+
..
.
+
−
+
−
..
.
−
+
−
+
..
.

···
· · ·

· · ·


· · ·
...


1 2 3
Example. A = 4 5 6.
7 8 9
Expansion by


1 ∗ ∗
 ∗ 5 6
∗ 8 9
5
det A = 1 8
the 1st row:

 
∗ 2 ∗
∗ ∗
4 ∗ 6 4 5
7 8
7 ∗ 9
4 6
6 + 34
−
2
7 9
7
9

3
∗
∗
5 8
= (5 · 9 − 6 · 8) − 2(4 · 9 − 6 · 7) + 3(4 · 8 − 5 · 7) = 0.


1 2 3
Example. A = 4 5 6.
7 8 9
Expansion by the 2nd column:

 
 

1 ∗ 3
1 ∗ 3
∗ 2 ∗
 4 ∗ 6  ∗ 5 ∗  4 ∗ 6 
∗ 8 ∗
7 ∗ 9
7 ∗ 9
4 6
1 3
1 3
+ 5
det A = −2 7 9 − 84 6
7 9
= −2(4 · 9 − 6 · 7) + 5(1 · 9 − 3 · 7) − 8(1 · 6 − 3 · 4) = 0.

1 2 3
Example. A = 4 5 6.
7 8 9

Subtract the
the 3rd row:
1 2
4 5
7 8
1st row from the 2nd row and from
3 1 2 3 1 2 3 6 = 3 3 3 = 3 3 3 = 0
9 7 8 9 6 6 6
since the last matrix has two proportional rows.


1 2 3
Another example. B = 4 5 6 .
7 8 13
Let’s do some row reduction.
Add −4 times the 1st row to the 2nd row:
1 2 3 1 2 3
4 5 6 = 0 −3 −6 7 8 13 7 8 13 Add −7 times the 1st row to the 3rd row:
1 2 3 1 2 3
0 −3 −6 = 0 −3 −6 7 8 13 0 −6 −8 Expand the determinant by the 1st column:
1 2 3
−3 −6 0 −3 −6 = 1 −6 −8 0 −6 −8 Thus
1 2
−3 −6 = (−3) det B = −6 −8 −6 −8 1 2
= (−3)(−2)(−2) = −12.
= (−3)(−2) 3 4

2 −2 0 3
−5 3 2 1

Example. C = 
 1 −1 0 −3, det C =?
2 0 0 −1

Expand the determinant by the 3rd column:
2 −2 0 3 2 −2 3 −5 3 2 1 = −2 1 −1 −3 1 −1 0 −3 2 0 −1 2 0 0 −1 Add −2 times the 2nd row to the 1st row:
0 0 9
2 −2 3 det C = −2 1 −1 −3 = −2 1 −1 −3 2 0 −1 2 0 −1 Expand the determinant by the 1st row:
0 0 9
1 −1 det C = −2 1 −1 −3 = −2 · 9 2
0
2 0 −1 Thus
1 −1 = −18 · 2 = −36.
det C = −18 2 0
Problem. For what values of a will the following
system have a unique solution?

 x + 2y + z = 1
−x + 4y + 2z = 2

2x − 2y + az = 3
The system has a unique solution if and only if the
coefficient matrix is invertible.


1 2 1
A = −1 4 2, det A =?
2 −2 a
Add −2 times the
1 2
−1 4
2 −2
3rd column to the 2nd column:
1 1
0
1 2 = −1
0
2 a 2 −2 − 2a a Expand the determinant
1
0
0
det A = −1
2 −2 − 2a
by the 2nd column:
1 1 1
2 = −(−2 − 2a) −1
2
a
Hence det A = −(−2 − 2a) · 3 = 6(1 + a).
Thus A is invertible if and only if a 6= −1.
More properties of determinants
Determinants and matrix multiplication:
• if A and B are n×n matrices then
det(AB) = det A · det B;
• if A and B are n×n matrices then
det(AB) = det(BA);
• if A is an invertible matrix then
det(A−1 ) = (det A)−1 .
Determinants and scalar multiplication:
• if A is an n×n matrix and r ∈ R then
det(rA) = r n det A.
Examples




−1 2 1
1 0 0
X =  0 2 −2, Y = −1 3 0.
0 0 −3
2 −2 1
det X = (−1) · 2 · (−3) = 6, det Y = det Y T = 3,
det(XY ) = 6 · 3 = 18, det(YX ) = 3 · 6 = 18,
det(Y −1 ) = 1/3, det(XY −1 ) = 6/3 = 2,
det(XYX −1 ) = det Y = 3, det(X −1 Y −1 XY ) = 1,
det(2X ) = 23 det X = 23 · 6 = 48,
det(−3X T XY −4 ) = (−3)3 · 6 · 6 · 3−4 = −12.
The Vandermonde determinant
Definition. The Vandermonde determinant is
the determinant of the following matrix


1 x1 x12 · · · x1n−1


1 x2 x22 · · · x2n−1 


n−1 
2
V =
1
x
x
·
·
·
x
3
3
3

,
..
.
 ... ...
.
. . .. 
.


1 xn xn2 · · · xnn−1
where x1 , x2 , . . . , xn ∈ R. Equivalently,
V = (aij )1≤i,j≤n , where aij = xij−1 .
Examples.
1 x1 = x2 − x1 .
• 1 x2 1 x1 x12 1 x1
0
2
2
• 1 x2 x2 = 1 x2 x2 − x1 x2 1 x x2 1 x x2 − x x 3
3
1 3
3
3
1
0
0
x2 − x1 x22 − x1 x2 2
= 1 x2 − x1 x2 − x1 x2 = 2
x
−
x
x
−
x
x
3
1
1
3
3
1 x − x x2 − x x 3
1
1
3
3
1
1 x2 x2
= (x2 − x1 ) = (x2 − x1 )(x3 − x1 ) x3 − x1 x32 − x1 x3 1 x3 = (x2 − x1 )(x3 − x1 )(x3 − x2 ).
Theorem
1 x
1
1 x2
1 x3
... ...
1 xn
x12
x22
x32
···
···
···
...
..
.
xn2 · · ·
Y
=
(xj − xi ).
..
.
1≤i<j≤n
xnn−1 x1n−1
x2n−1
x3n−1
Corollary The Vandermonde determinant is not
equal to 0 if and only if the numbers x1 , x2 , . . . , xn
are distinct.
Let x1 , x2 , . . . , xn be distinct real numbers.
Theorem For any b1 , b2 , . . . , bn ∈ R there exists a
unique polynomial p(x) = a0 +a1 x+ · · · +an−1 x n−1
of degree less than n such that p(xi ) = bi ,
1 ≤ i ≤ n.

a0 + a1 x1 + a2 x12 + · · · + an−1 x1n−1 = b1



a0 + a1 x2 + a2 x22 + · · · + an−1 x2n−1 = b2
············



a0 + a1 xn + a2 xn2 + · · · + an−1 xnn−1 = bn
a0 , a1 , . . . , an−1 are unknowns. The coefficient
matrix is the Vandermonde matrix.
A general system of n linear equations in n variables:

a11 x1 + a12 x2 + · · · + a1n xn = b1



a21 x1 + a22 x2 + · · · + a2n xn = b2
⇐⇒ Ax = b,
·········



an1 x1 + an2 x2 + · · · + ann xn = bn
where

a11 a12
a a
 21 22
A =  ..
..
 .
.
an1 an2
 
 

b1
x1
. . . a1n





. . . a2n 
 b2 
x2 
, x =  .. , b =  .. 
. . . ... 
.
.

bn
xn
. . . ann
Cramer’s rule

a11 x1 + a12 x2 + · · · + a1n xn = b1



a21 x1 + a22 x2 + · · · + a2n xn = b2
⇐⇒ Ax = b
·
·
·
·
·
·
·
·
·



an1 x1 + an2 x2 + · · · + ann xn = bn
Theorem Assume that the matrix A is invertible.
Then the only solution of the system is given by
det Ai
xi =
, i = 1, 2, . . . , n,
det A
where the matrix Ai is obtained by substituting the
vector b for the ith column of A.
Example.

 x + 2y + 3z = 0
4x + 5y + 6z = 0

7x + 8y + 13z = 1
Augmented matrix of the system:


1 2 3 0
(A | b) = 4 5 6 0
7 8 13 1
As obtained earlier in this lecture, det A = −12.
Since det A 6= 0, there exists a unique solution of
the system.
Example.

 x + 2y + 3z = 0
4x + 5y + 6z = 0

7x + 8y + 13z = 1


1 2 3 0
4 5 6 0
7 8 13 1
By Cramer’s rule,
0 2 3 0 5 6 2 3
1 8 13 5 6 −3
1
=
x=
=
=
1 2 3 −12
−12 4
4 5 6 7 8 13 Example.

 x + 2y + 3z = 0
4x + 5y + 6z = 0

7x + 8y + 13z = 1
By Cramer’s
1
4
7
y=
1
4
7
rule,


1 2 3 0
4 5 6 0
7 8 13 1
0 3 1 3
0 6 − 4 6
1 13 6
1
=
=
=−
−12
−12
2
2 3 5 6 8 13 Example.

 x + 2y + 3z = 0
4x + 5y + 6z = 0

7x + 8y + 13z = 1


1 2 3 0
4 5 6 0
7 8 13 1
By Cramer’s rule,
1 2 0
4 5 0
1 2
7 8 1
4 5
−3
1
=
z=
=
=
1 2 3 −12
−12 4
4 5 6 7 8 13 System of linear equations:

 x + 2y + 3z = 0
4x + 5y + 6z = 0

7x + 8y + 13z = 1
Solution: (x, y , z) =
1 1
1
.
,− ,
4
2 4
Determinants and the inverse matrix
Given an n×n matrix A = (aij ), let Mij denote the
(n − 1)×(n − 1) submatrix obtained by deleting the ith row
and the jth column of A.
e = (αij )
The cofactor matrix of A is an n×n matrix A
i+j
defined by αij = (−1) det Mij .
e T A = AA
e T = (det A)I .
Theorem A
Corollary If det A 6= 0 then the matrix A is invertible and
eT .
A−1 = (det A)−1 A
e T = (det A)I means that
AA
Pn
k+j
akj det Mkj = det A for all k,
j=1 (−1)
Pn
k+j
amj det Mkj = 0 for m =
6 k.
j=1 (−1)
Related documents