9 8 7
6 5 4
3 2 1
X
or Y =
1
1
1
]
1
i h g
f e d
c b a
Both X and Y are comprised of 3 rows and 3 columns each.
In general, an m n matrix has m rows and n columns (note the ordering: rows then
columns). An m n matrix may be represented as:
A =
1
1
1
1
]
1
mn 2 m 1 m
n 2 22 21
n 1 12 11
a a a
a a a
a a a
or more compactly A = (a
ij
)
m n
The order of the matrix is m n, read as m by n.
A square matrix is one in which the number of rows equals the number of columns i.e. m = n.
Given the ordering of matrices, it is possible to identify a particular element by its row and
column position. Thus (a
12
) represents the element in the first row and second column of the
matrix.
In general, matrices are derived from systems of linear simultaneous equations. Thus:
3x + 2y = 10
4x + 7y = 17
may be decomposed in matrix form as:
A z = b
1
We would like to acknowledge the research assistance of M. Maharaj and E. Bahaw
where A =
,
_
7 4
2 3
is a matrix of coefficients and both b and z are column vectors of constants and unknowns
respectively, with
b =
,
_
17
10
and z =
,
_
y
x
5.2 Matrix Addition
The sum of two matrices A and B are obtained by adding corresponding elements of A and
B. Note that for matrix addition to be possible, the order of A must be the same as B. Thus:
1
1
1
]
1
mn m m
n
a a a
a a a
2 1
1 12 11
+
1
1
1
]
1
mn m m
n
b b b
b b b
2 1
1 12 11
=
1
1
1
]
1
+ +
+ +
) ( ) (
) ( ) (
1 1
1 1 11 11
mn mn m m
n n
b a b a
b a b a
The resulting matrix A + B is of the same order as A and B. In general, for A = [a
ij
] and B =
[b
ij
] both A and B of order m n, their sum or difference may be defined as:
A t B = [a
ij
t b
ij
]
mxn
Note that the addition (and subtraction) of matrices is:
(i) Commutative
(A + B) = B + A
(ii) Associative
(A + B) + C = A + (B + C)
Note also that:
(i)
1
]
1
d c
b a
+
1
]
1
j i h
g f e
cannot be evaluated as both matrices dont have the same
dimension.
(ii) A + 0 = [a
ij
] + [0] = [a
ij
] = A
(iii) If A = [a
ij
]
m n
, then A = [a
ij
]
m n
(iv) If B = [a
ij
]
m n
, then A + B = [a
ij
]
m n
+ [a
ij
]
m n
= [0]
m n
Example 5.1
New chains of One stop stores have opened in T&T with 4 outlets, San Fernando, P.O.S.,
Arima and Pt. Fortin. Three of its main items and their stock levels in the various districts
are:
Coca Cola (case) Potato Chips (00 pks.) Bread (loaves)
San Fernando 9 3 6
P.O.S. 10 2 8
Arima 11 7 12
Pt. Fortin 12 1 19
The retail outlet recorded sales as shown below:
Coca Cola (case) Potato Chips (00pks.) Bread (loaves)
San Fernando 2 2 4
P.O.S. 8 1 5
Arima 7 4 8
Pt. Fortin 5 0 17
What is the new level of stocks?
(a) per district
(b) per commodity
Solution
,
_
19 1 12
12 7 11
8 2 10
6 3 9

,
_
17 0 5
8 4 7
5 1 8
4 2 2
=
,
_
2 1 7
4 3 4
3 1 2
2 1 7
=
,
_
Fortin . Pt
Arima
. S . O . P
Fernando San
Example 5.2
A popular hardware in Malabar advertises the following price and delivery cost per unit of
wood, cement and galvanize.
Wood Cement Galvanize
Purchase 12 8 4
Delivery 2 2 1
Assume that government regulation raises the purchase price by 10% per unit and the
delivery cost by 40%. Write out expressions for:
a) The change in unit prices and delivery costs
b) The new price and delivery cost levels
Solution
(i) C =
1
]
1
1 2 2
4 8 12
C =
1
]
1
) 1 ( 4 . 0 ) 2 ( 4 . 0 ) 2 ( 4 . 0
) 4 ( 1 . 0 ) 8 ( 1 . 0 ) 12 ( 1 . 0
C =
1
]
1
4 . 0 8 . 0 8 . 0
4 . 0 8 . 0 2 . 1
C* = C + C
Note that C* = C + C = C + C
=
1
]
1
1 2 2
4 8 12
+
1
]
1
4 . 0 8 . 0 8 . 0
4 . 0 8 . 0 2 . 1
=
1
]
1
4 . 1 8 . 2 8 . 2
4 . 4 8 . 8 2 . 13
Example 5.3
The table below shows the trends in the life expectancy of vegetarians and non vegetarians in
a Caribbean island. From the data in the table, form a matrix which demonstrates how many
years more vegetarians are expected to live as compared to nonvegetarians.
Year Vegetarians Meat Eaters
1960 60 45
1970 70 50
1980 75 52
1990 71 55
2000 80 56
Solution
,
_
80
71
75
70
60

,
_
56
55
52
50
45
=
,
_
24
16
23
25
15
5.3 Matrix Multiplication
Two matrices A and B are conformable for multiplication if the number of columns in A is
equal to the number of rows in B. Let A be of order m n and B of order n p, then AB = C
where C is of order m p. More technically, matrix multiplication refers to a repeat process
of vector multiplication. Thus consider the first row of A and the first column of B, where A
and B are m n and n p matrices respectively.
[a
11
a
12
a
1n
]
1
1
1
1
]
1
1
11
n
b
b
= a
11
b
11
+ a
12
b
21
+ + a
1n
b
n1
= a
1
b
1
which indicates that corresponding elements are multiplied and the results are summed.
In general A
(m n)
B
(n p)
= C
(m p)
or
1
1
1
1
]
1
mn 2 m 1 m
n 2 22 21
n 1 12 11
a a a
a a a
a a a
1
1
1
1
1
]
1
np 2 n 1 n
p 2 22 21
p 1 12 11
b b b
b b b
b b b
=
1
1
1
1
]
1
m
2
1
a
a
a
1
1
1
]
1
  
b b b
  
p 2 1
The element in the ijth position in the matrix AB is the product of the i
th
row of A and j
th
column of B. To form the product AB, we premultiply B by A, and to form the product BA
we post multiply B by A.
Example 5.4
,
_
4 3
2 1
,
_
8 7
6 5
=
1
1
1
]
1
+ +
+ +
) 8 4 ( ) 6 3 ( ) 7 4 ( ) 5 3 (
) 8 2 ( ) 6 1 ( ) 7 2 ( ) 5 1 (
=
,
_
50 43
22 19
Example 5.5
A =
1
]
1
22 21
12 11
a a
a a
and B =
1
]
1
23 22 21
13 12 11
b b b
b b b
AB=
1
]
1
+ + + +
+ + + +
23 22 13 21 22 22 12 21 21 22 11 21
23 12 13 11 22 12 12 11 21 12 11 11
b a b a b a b a b a b a
b a b a b a b a b a b a
=
1
]
1
23 22 21
13 12 11
c c c
c c c
5.3.1Scalar and Vector Multiplication
Example 5.6
Let the value of the commodities for our 'One Stop' retail outlet stores be:
Coca Cola (case) Potato Chips (00pks.) Bread (loaves)
San Fernando 360 60 30
P.O.S. 400 40 40
Arima 440 140 60
Pt. Fortin 480 20 45
If the firm is celebrating its 1
st
anniversary in T&T and cuts all prices by 10%, what is the
value of its stock
(a) per district?
(b) per commodity?
Solution
0.9
480
440
400
360
20
140
40
60
,
_
45
60
40
30
=
432
396
360
324
18
126
36
54
,
_
5 . 40
54
36
27
=
,
_
Fortin . Pt
Arima
. S . O . P
Fernando San
Example 5.7
Wills icecream chain sells 200 strawberry, 120 chocolate and 240 orangepine icecreams per
hour in the period 611 on a Friday evening during the dry season. The price of a strawberry
icecream is $0.45, a chocolate icecream is $0.60 and an orangepine is $0.50. The cost to
Wills of a strawberry icecream is $0.38, a chocolate is $0.42 and an orangepine is $0.32.
Derive the firms total profits using:
(i) total concepts
(ii) marginal concepts
Solution
TR = P Q
1
1
1
]
1
240
120
200
5 Q
=
1
1
1
]
1
1200
600
1000
1
1
1
]
1
50 . 0
60 . 0
45 . 0
P
1
1
1
]
1
1
1
1
]
1
1200
600
1000
50 . 0
60 . 0
45 . 0
PQ
but PQ as is has no mathematical significance. Transposing Q and forming PQ' gives:
TR = PQ' =
[ ] 1200 600 1000
50 . 0
60 . 0
45 . 0
1
1
1
]
1
= $1 016.00
= TR  TC = $1 410.00  $1 016.00 = $394.00
(i) From a per unit perspective
Profit per unit = price per unit cost per unit
1
1
1
]
1
1
1
1
]
1
1
1
1
]
1
18 . 0
18 . 0
07 . 0
32 . 0
42 . 0
38 . 0
50 . 0
60 . 0
45 . 0
Total profit = profit per unit quantity
[ ] 1200 600 1000
18 . 0
18 . 0
07 . 0
1
1
1
]
1
= $394.00
5.3.2 Multiplication by a Scalar
If we multiply A
(m n)
by a vector, v
(n 1)
, then
1
1
1
1
]
1
m
2
1
a
a
a
1
1
1
]
1

v

=
1
1
1
1
]
1
v a
v a
v a
m
2
1
= C
where C is of order m 1, i.e. it is a column vector.
Note that a row vector is simply a matrix with one row. Similarly, a column vector is really a
matrix with one column. If we have a matrix, A
(m x n)
and we multiply it by a scalar k and form
a matrix B, then
b
ij
= k a
ij
If we use vectors Brv
i
= k(Arv
i
), where Brv
i
is the new row vector formed when the existing
row vectors of A is multiplied by k.
It can also be shown that:
kA = Ak
kA + IA = (k + I)A
k(A + B) = kA + kB
This implies that linear equations with scalar coefficients follow the laws of scalar algebra.
Example 5.8
Assume that the purchase price and delivery cost of three important building materials are as
follows:
Wood Cement Gravel
Purchase Price 120 20 600
Delivery Cost 15 5 50
We may represent these unit costs as
1
]
1
50 5 15
600 20 120
AC
With the introduction of a 20% VAT, new Average Total Cost per unit would increase to
AC* =
1
]
1
60 6 18
720 24 144
ATC 2 . 1
5.3.3 Distributive Laws and Associative Laws of Multiplication of Matrices
5.3.3.1 Distributive Law
With scalar multiplication, the distribution law states that a (b + c) = ab + ac. This also holds
for matrix multiplication, i.e. A (B + C) = AB + AC.
To prove that this is so, we can show that the ij
th
element of A (B + C) = ij
th
element of
(AB+BC).
Proof: Let A = [a
ij
]
m x n
, B = [b
ij
]
n x p
and C = [c
ij
]
n x p
The i
th
row of A is given by
a
i1
, a
i2
, a
in
The elements of the jth column of B + C are
b
1j
+ c
1j
, b
2j
+ c
2j
, , b
nj
+ c
nj
This implies that the element in the ith row and jth column of A (B + C) is:
a
i1
(b
1j
+ c
1j
) + a
i2
(b
2j
+ c
2j
) + + a
in
(b
nj
+ c
nj
) =
+ +
n
1 k
n
1 k
n
1 k
ky ik kj ik kj kj ik
c a b a ) c b ( a
which represents the sum of the elements in the ith row and jth column of AB and AC.
5.3.3.2 Associative Law
With scalar multiplication, the associative law states that a(bc) = (ab)c. This also holds for
matrix multiplication i.e. A(BC) = (AB)C. we will illustrate this with the proof below.
Proof: If A = [a
ij
] is of order mn, if B = [b
ij
] is of order np, and if C = [c
ij
] is of order pq.
The element of the ith row of A are a
i1
, a
i2
, a
in
and the elements of the jth column of
BC are
p
1 h
b
1h
c
hj
,
p
1 h
b
2h
c
hj
, ,
p
1 h
b
nh
c
hj
Hence the element standing in the ith row and jth column of A(BC) is
a
i1
p
1 h
b
1h
c
hj
+ a
i2
p
1 h
b
2h
c
hj
+ + a
in
p
1 h
b
nh
c
hj
=
p
1 k
a
ik
(
p
1 h
b
kh
c
hj
) =
p
1 h
(
p
1 k
a
ik
b
kh
) c
hj
= (
p
1 k
a
ik
b
k1
) c
ij
+ (
p
1 k
a
ik
b
k2
) c
2j
+ + (
p
1 k
a
ik
b
kp
) c
pj
This is the element standing in the ith row and jth column of (AB)C. Hence, A(BC) = (AB)C.
However, note that AB BA in general.
Note that also AB = 0 does not necessarily mean that A=0 or B=0 as with scalar
multiplication. Consider the simple matrix multiplication below:
,
_
0 0
y x
,
_
x x
y y
=
,
_
0 0
0 0
5.4 Equality of Matrices
Two matrices A and B are equal if the corresponding elements in A and B are equal. Thus:
1
1
1
1
]
1
mn 2 m 1 m
n 2 22 21
n 1 12 11
a a a
a a a
a a a
=
1
1
1
1
]
1
mn 2 m 1 m
n 2 22 21
n 1 12 11
b b b
b b b
b b b
if every a
ij
= b
ij
Note both matrices have the same order.
In vector form
a
i
= b
i
or Arv
i
= Brv
i
5.5 Transpose of a Matrix
The transpose of a matrix refers to a reordering of its rows and columns to give a matrix of
columns and rows. Thus A
m x n
becomes (A
T
)
n x m
under the transpose operation. Thus for:
A =
1
1
1
]
1
mn 1 m
n 2 21
n 1 11
a a
a a
a a
A' = A
T
=
1
1
1
]
1
mn n 2 n 1
1 m 21 11
a a a
a a a
The transpose of the sum of two matrices A and B, i.e. (A+B)', is A'+B' (see proof 1 below).
The transpose of the product of AB is B'A'. In general, the transpose of the product of any
number of matrices is equal to the product of their transpose in reverse order (see proof 2
below).
Note further that:
(A)
T
= A
T
(A
T
)
T
= A
Proof 1: (A + B)' = A' + B'
Let A = [a
ij
] and B = [b
ij
]. We need only check that the element in the i
th
row and j
th
column of
A', B', and (A + B)' are respectively a
ij
, b
ij
and a
ji
+ b
ji
.
Proof 2: (AB)' = B'A'
Let A = [a
ij
] be of order m n, B = [b
ij
] be of order np; then C = AB = [c
ij
] is of order m p.
The element standing in the i
th
row and j
th
column of AB is c
ij
=
n
1 k
a
ik
. b
kj
and this is also
the element standing in the j
th
row and ith column of (AB)'.
The element of the j
th
row of B' are b
1j
, b
2j
b
nj
and the element of the i
th
column of A' are a
i1
,
a
i2
a
in
. Then, the element in the j
th
row and i
th
column of B'A' is
p
1 k
b
kj
. a
ik
=
p
1 k
a
ik
. b
kj
= c
ij
Thus, (AB)' = B'A'
Proof 3: (A
T
)
1
= (A
1
)
T
To show this post multiply the expression by A
T
(A
T
)
1
A
T
= (A
1
)
T
A
T
L.H.S. = I
R.H.S. = (A
1
)
T
A
T
= (A A
1
)
T
= I
T
= I
LHS = RHS
Proof 4: (A)
T
= (A)
T
for A
nxm
and a scalar
The ij
th
entry of A is a
ij
and the corresponding entry of (A)
T
= a
ji
Since (a
ij
)
T
= a
ji
then [a
ij
]
T
= [a
ji
]
= a
ji
which demonstrates that (A)
T
= (A
T
)
5.6 Symmetric Matrices
A symmetric matrix is such that A
T
=A. Consider the matrix A below:
A= A
T
=
1
1
1
]
1
8 7 3
7 4 2
3 2 1
=A
Example 5.9
Let A be a square matrix and A
T
its transpose. Show that A + A
T
is symmetric.
Solution
Form B = A + A
T
B
T
= (A + A
T
)
T
= A
T
+ (A
T
)
T
= A
T
+ A
= A + A
T
which shows that B is symmetric, i.e. A + A
T
is symmetric.
5.7 The Zero Matrix
The zero matrix has all its elements equivalent to zero and can vary in size. It is also called
the null matrix and it is written as:
(0)
m n
, which indicates it can have any size.
Given the definition of 0, then A = B, means that A  B = 0. Also A + 0 = A and A  0 = A.
Note also that 0A = A0 = 0. It is important to recognize that AB = 0 does not mean that A or
B equals zero i.e. the cancellation law of multiplication is inapplicable to matrices.
Recall the previous example:
,
_
0 0
y x
,
_
x x
y y
=
,
_
0 0
0 0
5.8 Identity and Diagonal Matrices
An identity matrix of n x n is given by I
n
.
I
n
=
1
1
1
1
]
1
1 0 0
0 1 0
0 0 1
The identity and zero matrices play roles in matrix algebra parallel to scalar algebra. Thus:
(i) 0 + A = A + 0 = A
(ii) A  A = 0
(iii) 0 * B = 0
(iv) B * 0 = 0
(v) I
m m
A
m n
= A
m n
(vi) A
m n
I
n n
= A
m n
Example 5.10
1
1
1
]
1
1 0 0
0 1 0
0 0 1
1
1
1
]
1
33 32 31
23 22 21
13 12 11
a a a
a a a
a a a
=
1
1
1
]
1
33 32 31
23 22 21
13 12 11
a a a
a a a
a a a
A square matrix which has zeroes as the data items in every cell except in the leading
diagonal is called a diagonal matrix. More specifically, the matrix A, in which the typical
element is a
ij
, is a diagonal matrix if a
ij
= 0 for i j.
A typical 4 x 4 diagonal matrix is
,
_
4 0 0 0
0 3 0 0
0 0 2 0
0 0 0 1
A
A diagonal matrix is similar to the identity matrix with all offdiagonal elements equal to zero
but with entries along the principal (leading) diagonal, such as the n n example shown
below:
D =
1
1
1
1
]
1
nn
22
11
d 0 0
0 d 0
0 0 d
written as D = diag. [d
11
d
nn
]
Example 5.11
Let D be the diagonal matrix
1
1
1
]
1
3
2
1
d 0 0
0 d 0
0 0 d
i.e. D = diag (d
1
, d
2
, d
3
)
Prove by induction that
D
n
= diag (d
1
n
, d
2
n
, d
3
n
)
Solution
Step 1: Show the result is true when n = 1
D
1
= D = diag (d
1
, d
2
, d
3
) = diag(d
1
1
, d
2
1
, d
3
1
)
Thus the result is true for n = 1.
Step 2: Assume the statement is true for n = k
D
K
= diag (d
1
K
, d
2
K
, d
3
K
)
Step 3: Show statement is true for n = k + 1
D
K + 1
= D
K
. D
=
1
1
1
1
]
1
K
3
K
2
K
1
d 0 0
0 d 0
0 0 d
1
1
1
]
1
3
2
1
0 0
0 0
0 0
d
d
d
=
1
1
1
1
]
1
+
+
+
1 K
3
1 K
2
1 K
1
d 0 0
0 d 0
0 0 d
= diag (d
1
K+1
, d
2
K+2
, d
3
K+3
)
The expression holds for n = k+1.
Step 4: The expression holds for n = 1, then it is true for n = 1+1 = 2; then it is true for n = 2
+ 1 = 3; and so on. The expression is therefore true for all positive integers n.
5.8.1 Upper Triangular (UTM), Lower Triangular (LTM) and Diagonal Matrices
A matrix A
n n
is said to be upper triangular if the a
ij
element is zero for all elements where
i>j. Correspondingly, a matrix B
n n
is said to be lower triangular if the b
ij
element is zero
for all elements where i < j. A diagonal matrix is both upper and lower triangular.
E.g. of a UTM
1
1
1
1
]
1
10 0 0 0
9 8 0 0
7 6 5 0
4 3 2 1
E.g. of a LTM
1
1
1
1
]
1
10 9 8 7
0 6 5 4
0 0 3 2
0 0 0 1
5.9 Orthogonal Matrices
A matrix A is said to be orthogonal if its transpose is equivalent to its inverse. Also, a square
matrix is said to be orthogonal if the product of the matrix with its transpose yields the
identity matrix.
A
T
= A
1
i.e. if AA
T
= A
T
A = I
Example 5.12
Demonstrate that the following matrix is orthogonal:
1/9
1
1
1
]
1
4 1 8
7 4 4
4 8 1
Solution
AA
T
=
,
_
+ +
81
81
0 0
0
81
81
0
0 0
81
16
81
64
81
1
= I
3
Example 5.12
Show that for a symmetric matrix A such that A
T
= A
1
that A
2
= I.
Solution
A
2
= A . A = A . A
T
(since A = A
T
for symmetric matrices)
= A . A
1
(A is orthogonal)
= I
5.9.1 Some Properties of Orthogonal Matrices
(1) If A and B are orthogonal matrices, then AB is an orthogonal matrix
(AB)
T
= B
T
A
T
(AB)
T
AB = B
T
A
T
AB
= B
T
IB since A
T
A = I
= B
T
B
= I
Hence AB is orthogonal
(2) For A
n n
an orthogonal matrix, the determinant is t 1.
i.e. A = t 1
Note that A
1
 =
1
/
A
(1)
A
T
 = A (2)
and A
T
= A
1
(3)
so that A = A
T
 = A
1
 =
1
/
A
A
2
= 1
A = t 1
(3) A
1
and A
T
are each orthogonal if A itself is orthogonal
Since A
1
= A
T
and (A
T
)
T
= A
then (A
T
)
T
A
T
= AA
T
= AA
1
= I
which shows that A
T
is orthogonal
5.10 Invertible Matrices
A square matrix A is invertible if there exists a matrix B such that
AB = BA = I
The matrix B is a unique matrix. There is a link between an invertible matrix and the solution
to a system of linear equations. More specifically, a square matrix A is invertible if the
system of equations represented by
Ax = b
has a unique solution.
Example 5.13
Prove for any non singular matrices A and B that
(AB)
1
= B
1
A
1
Solution
We are told that A and B are non singular so that both A
1
and B
1
exists.
We have AB. B
1
A
1
= A IA
1
= AA
1
= I
Also B
1
A
1
AB = B
1
IB = B
1
B = I
Hence (AB)
1
= B
1
A
1
5.11 Power of Matrices
For a square matrix A of order n n, we can write the product of A A (p times) as A
p
.
In the mould of reasoning consistent with the associative law, we can note that
A
x
A
y
= A
y
A
x
= A
x+y
and (A
x
)
y
= (A
y
)
x
= A
xy
Example 5.14
Let A =
,
_
1 0
2 1
Find A
2
, A
3
and A
4
Solution
A
2
=
,
_
1 0
4 1
, A
3
=
,
_
1 0
6 1
and A
4
=
,
_
1 0
8 1
Example 5.15
Show using proof by induction that (A
n
)
1
= (A
1
)
n
, where A is a nonsingular matrix and n is a
positive number.
Solution
Step 1: Show that the expression is true when n = 1
(A
1
)
1
= A
1
= (A
1
)
1
so that the result holds when n = 1
Step 2: Assume the result is true when n = k i.e. Assume that (A
k
)
1
= (A
1
)
k
.
Step 3: Show that the result is true for n = k + 1
(A
k+1
)
1
= (A
1
)
k+1
(A
k+1
)
1
= (AA
k
)
1
= (A
k
)
1
A
1
= (A
1
)
k
(A
1
) = (A
1
)
k+1
Thus the result holds for n = k+1
Step 4: Since the result holds for n=1 and we have shown that if it holds for n = k it will hold
for n=k+1, by induction it is true for all positive integers.
5.12 Differences between Scalars and Matrices
For scalars a and b and matrices A and B, their main points of distinction are:
(i) Whilst with scalars ab = ba, this is not generally the case with matrices as AB = BA
only in a specific special case.
(ii) With scalars, ab = 0 means either a and/or b = 0. With matrices, AB = 0 does not
necessarily mean that A and/or B = 0.
Lebih dari sekadar dokumen.
Temukan segala yang ditawarkan Scribd, termasuk buku dan buku audio dari penerbitpenerbit terkemuka.
Batalkan kapan saja.