Upload
bryan-penfound
View
230
Download
0
Embed Size (px)
Citation preview
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
1/52
Section Three: Determinants
Textbook: Ch. 2.1, 2.2, 2.3, 2.4, 2.5GOALS OF THIS CHAPTER
- introduce further the matrix determinant
- calculate determinants using the weave method
- calculate determinants using cofactor expansions
- calculate determinants using elementary row operations
- see properties of the matrix determinant
- application to geometry: areas of parallelograms andtriangles
- application to linear systems: solving systems using
Cramers Rule
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
2/52
WHAT IS A MATRIX DETERMINANT?
Sometimes it is nice to know if a matrix isinvertible BEFORE we do all the work toactually find the inverse.
It turns out that there is a special numberassociated to a square matrix that tells us if thatmatrix is invertible!
We call this special number the determinant ofthe matrix A, or write it as det(A).
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
3/52
WHAT IS A MATRIX DETERMINANT?
Recall the 2x2 matrix determinant:
How do we actually get this number?
A =a b
c d
det(A) = ad - bc
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
4/52
WHAT IS A MATRIX DETERMINANT?
On our list of Non-singular Equivalences, wenoted that a matrix was invertible if its RREF
was an identity matrix.
Lets try to reduce this matrix to RREF and see what happens:
A =
a b
c d
Ex. 1 Where the 2x2 Determinant Comes From
R1 R1 / a
(here we are assuming that a is not zero)
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
5/52
1 b/a
c d
WHAT IS A MATRIX DETERMINANT?
A =
Ex. 1 Where the 2x2 Determinant Comes From
R2 R2 c*R1
1 b/a
0 d (bc)/aA =
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
6/52
WHAT IS A MATRIX DETERMINANT?
Ex. 1 Where the 2x2 Determinant Comes From
1 b/a
0 d (bc)/aA =
So, in order for A to be an identity matrix, wewould need the number d (bc)/a to not be zero:
d (bc)/a 0
ad bc 0
det(A)
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
7/52
WHAT IS A MATRIX DETERMINANT?
Ex. 1 Where the 2x2 Determinant Comes From
We have just shown the most important propertyof the determinant:
IF det(A) = 0, THEN THE MATRIX A IS NOTINVERTIBLE.
IF det(A) 0, THEN THE MATRIX A ISINVERTIBLE.
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
8/52
THE WEAVE METHOD
The weave method gives us an easy way toremember the determinant of a 2x2 or 3x3matrix. We cannot use this method for higherorder matrices.
A =a b
c d
det(A) =
Ex. 2 The 2x2 Weave Method
- bc+ ad
+-
Down and to theright, we add.
Down and to the left,we subtract.
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
9/52
THE WEAVE METHOD
A =
a b c
d e f
g h i
det(A) =
Ex. 3 The 3x3 Weave Method
- ceg - afh - bdi+ aei + bfg + cdh
+-
To make it work, you have toadd the first two columns ofthe matrix on the right side.
Down and to the left,we subtract.
a b
d e
g h
Down and to theright, we add.+ +-
-
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
10/52
THE WEAVE METHOD
A =
1 0 3
2 -8 7
5 2 1
det(A) =
Ex. 3 The 3x3 Weave Method
- 3*(-8)*5 - 1*7*2 - 0*2*11*(-8)*1 + 0*7*5 + 3*2*2
+-
1 0
2 -8
5 2
+ +--
det(A) = -8 + 0 + 12 + 120 - 14 - 0
det(A) = 110
REMARK: Since thedeterminant is not
zero, it would bepossible to find theinverse of this matrix.
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
11/52
COFACTOR EXPANSION
How do we deal with determinants of higherorder matrices? Wouldnt it be nice if we
could write these formulas in a simpler way?
To answer these questions, we will use the cofactors of matrixentries. Recall that the cofactor of a matrix entry a(i,j) isdefined to be:
Aij = (-1)i+jMij
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
12/52
a11 a12
a21 a22
Lets revisit the 2x2 determinant:
A =
det(A) =
Ex. 4 The 2x2 Cofactor Expansion
- a12a21+ a11a22
+-Lets try to rewrite
this in terms ofcofactors.
COFACTOR EXPANSION
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
13/52
a11 a12
a21 a22
Lets revisit the 2x2 determinant:
Ex. 4 The 2x2 Cofactor Expansion
A =
det(A) = a11a22 a12a21det(A) = a11M11 a12a21
We can replace a22with the determinantof a minor!
COFACTOR EXPANSION
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
14/52
Lets revisit the 2x2 determinant:
Ex. 4 The 2x2 Cofactor Expansion
a11 a12
a21 a22A =
det(A) = a11a22 a12a21det(A) = a11M11 a12M12
We can do the samething for a21.
COFACTOR EXPANSION
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
15/52
Lets revisit the 2x2 determinant:
Ex. 4 The 2x2 Cofactor Expansion
det(A) = a11A11 + a12A12
a11 a12
a21 a22A =
det(A) = a11a22 a12a21det(A) = a11M11 a12M12
Finally, we get rid ofthe negative signs byusing the cofactor.This is why thecofactor has a (-1) inthe formula!
COFACTOR EXPANSION
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
16/52
Lets revisit the 2x2 determinant:
Ex. 4 The 2x2 Cofactor Expansion
det(A) = a11A11
a11 a12
a21 a22A =
We call this a cofactor expansionalong the first row. Pretend you aredrawing an arrow and whateverentry you hit, multiply that entry byits corresponding cofactor.
+ a12A12
COFACTOR EXPANSION
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
17/52
Next, we will show that a cofactor expansionalong the first row of a 3x3 matrix gives the3x3 weave formula.
Ex. 5 The 3x3 Cofactor Expansion
COFACTOR EXPANSION
A =
a b c
d e f
g h i
det(A) = a*A11 + b*A12 + c*A13
det(A) = a(-1)2M11 + b(-1)3M12 + c(-1)
4M13
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
18/52
a b c
d e f
g h i
Ex. 5 The 3x3 Cofactor Expansion
COFACTOR EXPANSION
A =
det(A) = a*A11 + b*A12 + c*A13
det(A) = a(-1)2M11 + b(-1)3M12 + c(-1)
4M13
det(A) = aM11 bM12 + cM13
det(A) = a(ei - fh) bM12 + cM13
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
19/52
Ex. 5 The 3x3 Cofactor Expansion
COFACTOR EXPANSION
A =
a b c
d e f
g h i
det(A) = a*A11 + b*A12 + c*A13
det(A) = a(ei - fh) b(di - fg) + cM13
det(A) = a(-1)2M11 + b(-1)3M12 + c(-1)
4M13
det(A) = aM11 bM12 + cM13
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
20/52
Ex. 5 The 3x3 Cofactor Expansion
COFACTOR EXPANSION
A =
a b c
d e f
g h i
det(A) = a*A11 + b*A12 + c*A13
det(A) = a(ei - fh) b(di - fg) + c(dh - eg)
det(A) = a(-1)2M11 + b(-1)3M12 + c(-1)
4M13
det(A) = aM11 bM12 + cM13
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
21/52
Ex. 5 The 3x3 Cofactor Expansion
COFACTOR EXPANSION
A =
a b c
d e f
g h i
det(A) = a*A11 + b*A12 + c*A13
det(A) = a(ei - fh) b(di - fg) + c(dh - eg)
det(A) = aei - afh - bdi + bfg + cdh - cegThis is theformula we
saw before!
det(A) = a(-1)2M11 + b(-1)3M12 + c(-1)
4M13
det(A) = aM11 bM12 + cM13
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
22/52
COFACTOR EXPANSION
So this gives us a way to find determinants of4x4 or higher matrices! We expand using
cofactors (since we cant use the weaveformula).
Before I give some examples, I will state a theorem that mightsave you some time.
Thm. 6 If A is a square matrix, then det(A) can be expressedas a cofactor expansion along any row or any columnof thematrix.
In the next few examples, we will see just howawesome this theorem is.
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
23/52
COFACTOR EXPANSION
Ex. 7 Illustration of Thm. 6
- done in class
Ex. 9 Speedy 4x4 Expansion- done in class
Ex. 8 4x4 Cofactor Expansion
- done in class
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
24/52
ELEMENTARY ROW OPERATIONS
Recall the three types of elementary matrices:
TYPE I multiplying a row by a non-zero number
TYPE II adding a multiple of one row to another
TYPE III switching two rows
What we will do next is to calculate the determinantsof these types of elementary matrices, so we can seehow elementary row operations affect determinantcalculations.
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
25/52
1 0 0
0 1 0
0 0 2
TYPE I multiplying a row by a non-zero number
E1 =
ELEMENTARY ROW OPERATIONS
Elementary matrix obtained from operation R3 2R3
det(E1) = 2
So the determinant of aType I elementary matrix isequal to the non-zeronumber you use!
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
26/52
1 0 -3
0 1 0
0 0 1
TYPE II adding a multiple of one row to another
E2 =
ELEMENTARY ROW OPERATIONS
Elementary matrix obtained from operation R1 R1 3R3
det(E2) = 1
So the determinant of aType II elementary matrixis equal to one!
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
27/52
1 0 0
0 0 1
0 1 0
TYPE III switching two rows
E3 =
ELEMENTARY ROW OPERATIONS
Elementary matrix obtained from operation R2 R3
det(E3) = -1
So the determinant of aType III elementary matrixis equal to negative one!
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
28/52
ELEMENTARY ROW OPERATIONS
Fact:
For any square matrix A and any elementary matrix E
det(EA) = det(E) det(A).
This means if we want to use an elementary rowoperation on A to make our determinant calculationeasier, we can balance this action by dividing by thedeterminant of our elementary matrix.
det(EA)
det(E)= det(A)
performing anE.R.O. on a
determinant
balancing bydividing
keeps det(A) in
tact
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
29/52
Ex. 11 Algebraic Row Operation Method- done in class
Ex. 10 5x5 Elementary Row Operation Method
- done in class
ELEMENTARY ROW OPERATIONS
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
30/52
PROPERTIES OF THE DETERMINANT
Thm. 12 Properties of the Determinant- proof done in class
(1) A is singular if and only if det(A) = 0
(2) det(AB) = det(A)det(B)
(3) det(At) = det(A)
(4) If A is upper or lower triangular, thendet(A) is the product of the elements onthe main diagonal.
(5) det(cB) = cndet(B) (c is a scalar number)
(6) If A is invertible det(A-1) = 1 / det(A)
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
31/52
PROVING THE EQUIVALENT STATEMENTS
Now we have enough tools to prove the equivalent
statements. We also add two more items to the list.
(1)The nxn matrix A is invertible.(2) det(A) 0.(3) The system Ax=b has a unique solution for
every b.(4) The only solution of the homogeneous system
Ax=0 is the trivial solution x=0.(5) A is row equivalent to I (ie. there is not a full
row of zeros in the RREF of A).
(6) A can be written as a product of elementarymatrices
Proof done in class
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
32/52
Consider a triangle in the first quadrant:
FINDING AREAS USING DETERMINANTS
(x1,y1)
(x2,y2)
(x3,y3)
To calculate the area, we will calculate thearea of three rectangles and three triangles.
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
33/52
FINDING AREAS USING DETERMINANTS
(x1,y1)
(x2,y2)
(x3,y3)
The area of the first rectangle can becalculated by:
l*w = (x2 - x1)y1
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
34/52
FINDING AREAS USING DETERMINANTS
(x1,y1)
(x2,y2)
(x3,y3)
The area of the first triangle can becalculated by:
b*h/2 = (x2 - x1)(y2 - y1)*1/2
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
35/52
FINDING AREAS USING DETERMINANTS
(x1,y1)
(x2,y2)
(x3,y3)
The total area of this shape is:
l*w + b*h/2 = (x2 - x1)y1 + (x2 - x1)(y2 - y1)*1/2
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
36/52
FINDING AREAS USING DETERMINANTS
(x1,y1)
(x2,y2)
(x3,y3)
The area of the second rectangle can becalculated by:
l*w = (x3 x2)y3
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
37/52
FINDING AREAS USING DETERMINANTS
(x1,y1)
(x2,y2)
(x3,y3)
The area of the second triangle can becalculated by:
b*h/2 = (x3 x2)(y2 - y3)*1/2
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
38/52
FINDING AREAS USING DETERMINANTS
(x1,y1)
(x2,y2)
(x3,y3)
The total area of this shape:
l*w + b*h/2 = (x3 x2)y3 + (x3 x2)(y2 - y3)*1/2
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
39/52
FINDING AREAS USING DETERMINANTS
(x1,y1)
(x2,y2)
(x3,y3)
The area of the final rectangle can becalculated by:
l*w = (x3 x1)y3
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
40/52
FINDING AREAS USING DETERMINANTS
(x1,y1)
(x2,y2)
(x3,y3)
The area of the final triangle can becalculated by:
b*h/2 = (x3 x1)(y1 - y3)*1/2
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
41/52
FINDING AREAS USING DETERMINANTS
(x1,y1)
(x2,y2)
(x3,y3)
The area of the final shape is:
l*w + b*h/2 = (x3 - x1)y3 + (x3 x1)(y1 - y3)*1/2
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
42/52
So the area of the triangle is:
FINDING AREAS USING DETERMINANTS
(x1,y1)
(x2,y2)
(x3,y3)
[(x2 - x1)y1 + (x2 - x1)(y2 - y1)*1/2]
+ [(x3 x2)y3 + (x3 x2)(y2 - y3)*1/2]
- [(x3 - x1)y3 + (x3 x1)(y1 - y3)*1/2]
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
43/52
If we expand, we get:
FINDING AREAS USING DETERMINANTS
1/2*x2y1 1/2*x1y2 + 1/2*x3y2 1/2*x2y3 1/2*x3y1 + 1/2*x1y3
It turns out that this expression above can be written as
x1 y1 1
x2 y2 1
x3 y3 1
det
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
44/52
Sometimes the points will be not in the firstquadrant or the points will be labeled in adifferent order. This has an effect ofchanging the determinant by (-1), so we justtake the absolute value:
FINDING AREAS USING DETERMINANTS
x1 y1 1
x2 y2 1
x3 y3 1
Area of triangle = det
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
45/52
Since a parallelogram is just two triangles puttogether, we can use the previous formula forareas of parallelograms too! All we require arethree vertices of the parallelogram (thefourth one is just extra information).
FINDING AREAS USING DETERMINANTS
x1 y1 1
x2 y2 1
x3 y3 1
Area of parallelogram = det
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
46/52
Ex. 13 Area of Triangle and Parallelogram Using Determinants- done in class
FINDING AREAS USING DETERMINANTS
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
47/52
CRAMERS RULE
We can use determinants to solve equally determinedlinear systems in a much quicker way that usingGauss-Jordan elimination. The next method we see iscalled Cramers Rule.
Thm. 14 Let A be an nxn square matrix that is non-singular. Let bbe an nx1 column matrix. Let Ai be the matrix obtained by replacingthe ith column of A with b. If x is the unique solution of Ax=b, then
xi = det(Ai) , as i ranges from 1 to n.
det(A)
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
48/52
CRAMERS RULE
Proof: Since A is non-singular, we know the inverse ofA exists. We can use it to solve the linear system:
Ax = bx = A-1b
If we rewrite the inverse of A using the adjoint formula, we obtain:
x = (1/det(A) * adj(A))b
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
49/52
CRAMERS RULE
Write the formula using general matrices:
x = (1/det(A) * adj(A))b
A11 A21 An1
A12 A22 An2
A1n A2n Ann
x1
x2
xn
b1
b2
bn
= 1
det(A)
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
50/52
CRAMERS RULE
Lets calculate one of the entries in x:
A11 A21 An1
A12 A22 An2
A1n A2n Ann
x1
x2
xn
b1
b2
bn
=1
det(A)
x1 = 1/det(A) (A11b1 + A21b2+ + An1bn)
x1 = det(A1)det(A)
This is a cofactorexpansion down the firstrow of A1!
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
51/52
CRAMERS RULE
This works for all the other entries in the x-vector, so we have proved that Cramers Ruleworks.
Ex. 15 Cramers Rule for a 3x3 Linear System- done in class
8/3/2019 Determinants - Ch. 2.1, 2.2, 2.3, 2.4, 2.5
52/52
GABRIEL CRAMER (1704-1752)
- Swiss mathematician born in Geneva
- finished his education at eighteen and wasco-chair of a mathematics department attwenty!
- travelled a lot; he met Euler (that guy who has e named after him),Halley (that guy who has a comet named after him), Johann and DanielBernoulli (many things named after the Bernoullis)
- after Johann Bernoulli died, Cramer was asked to edit all ofBernoullis works and publish them
- Cramers Rule appears in Introduction lanalyse des lignes courbesalgbriques
- falling from his carriage while travelling and the over-work from