Ise Vmath109

Mathematics for Engineers Part I (ISE) Version 1.
1/2007-01-29

9-1
9 Linear Algebra: Vectors, Matrices, Determinants, Linear
Systems of Equations.

9.1 Vector Algebra

Certain quantities in mathematics and physics and its engineering applications can-
not be characterized alone by a real number; such quantities are represented by ar-
rows. For example, a force is represented by an arrow; the direction of the arrow de-
scribes the direction in which the force is applied, and
the length or magnitude of the arrow indicates the
strength. The velocity is also represented by an arrow
which points in the direction of the motion (Fig. 9.1.1),
and whose length indicates the speed. We will call such
arrows vectors.

Note 9.1.1:
In mathematics and physics we use two kinds of quantities, scalars and vectors. An
arrow (vector) is a quantity that is determined by both its magnitude and its direc-
tion. Velocity, force and so on, which are represented by arrows, are vector quanti-
ties. A scalar is a quantity that is determined by its magnitude, its number of units
measured on a suitable scale. Speed, weight, time, temperature, distance and so on
are scalar quantities.

An arrow (A,A) has a tail A, called its initial
point, and a head A, called its terminal point.
By a parallel shifting of the arrow (A,A) we get a
new arrow (B,B) (Fig. 9.1.3) which has the same
length and the same direction. Therefore by par-
allel shifting we get an infinite set of arrows

M = {(A,A'),(B,B'),(C,C'),...}.

Fig. 9.1.1

Fig 9.1.2 Arrow (A,A)
Velocity
vector
A
A
Mathematics for Engineers Part I (ISE) Version 1.1/2007-01-29

9-2
Definition 9.1.1:
Arrows which have the same length and the same direction
are called parallel-equal.

Definition 9.1.2:
The set M = {(A,A'),(B,B'),(C,C'),...} of all parallel-equal arrows is called an arrow
class a
r
. An individual arrow of the class a
r
is called a representative of the arrow
class.

Since an arrow class consists of infinitely many arrows, it is impossible to draw such
a class. Therefore we designate a representative with the symbol of its arrow class
(Fig. 9.1.4).

Definition 9.1.3 (Vector space):
A non-empty set V of vectors is called a real vector space (or real linear space) if in
V there are defined two algebraic operations (called vector addition and scalar multi-
plication) as follows:

I. Vector addition associates with every pair of vectors u and v of V an unique vec-
tor w of V, called the sum of u and v and denoted by w = u + v, such that the follow-
ing axioms are satisfied:

V1 Commutativity:
For any two vectors u and v of V applies: u + v = v + u.

V2 Associativity:
For any three vectors u, v, w V applies: (u + v) + w = u + (v + w).

V3 Zero vector:
There is a unique vector in V, called the zero vector and denoted by 0, such that for
every v in V applies:

v + 0 = 0 + v = v.

Fig. 9.1.3

Fig. 9.1.4: Representative of the arrow class a
r


9-3
V4 Inverse element:
For every v V there is an unique vector in V denoted by -v such that

v + (-v) = (-v) + v = 0.

II. Scalar multiplication: The real numbers are called scalars. Scalar multiplication
associates with every v in V and every scalar r R an unique vector of V, called the
product of r and v and denoted by rv such that the following axioms are satisfied:

Distributivity:
For every scalar r, s R and vectors u and v in V applies:

V5 r(u + v) = ru + rv

V6 (r + s)v = rv + sv

Associativity:
For all scalars r,s R and every v in V applies:

V7 r(sv) = (rs)v = s(rv)

Neutral element:
For every v in V applies:

V8 1v = v

Remark 9.1.1:
A complex vector space is obtained if, instead of real numbers, we take complex
numbers as scalars.

Definition 9.1.4:
With R
n
we denote the set of all ordered n-tuple (x
1
, x
2
,...,x
n-1
, x
n
), x
i
R , i = 1,...,n;
i.e.:

R
n
= {(x
1
, x
2
,...,x
n-1
, x
n
): x
i
R, i = 1,...,n}.

Example 9.1.1:
R
n
is a real vector space. Suppose a
r
= (a
1
, a
2
,...,a
n
) and b
r
= (b
1
, b
2
,...,b
n
) are ele-
ments of R
n
. Then ( )
n n 1 1
b a ,..., b a + + + b a
r
r
R
n
and ( )
n 1
ra ,..., ra r a
r
R
n
for
each real number r. It is easy to show that this so defined vector addition and scalar
multiplication satisfy V1 V8.

9-4

Fig. 9.1.5 R
2
Fig. 9.1.6 R
3

Example 9.1.2:
If an experiment involves reading seven strategically placed thermometers in a given
timer interval (e.g. each hour) then a result can be recorded as a 7-dimensional vec-
tor ( )
7 2 1
T ,..., T , T R
7
.

Definition 9.1.5:
A non-empty subset of vector space V that itself forms a vector space with respect to
the two algebraic operations defined for the vector space V is called a subspace of
the vector space V.

Definition 9.1.6:
Given any set of m vectors
1
v ,
2
v , ,
m
v in a vector space V. A linear combina-
tion of these vectors is an expression of the form

c
1
1
v +c
2
2
v ++c
m
m
v ,

where c
1
, c
2
,, c
m
are any scalars.

Definition 9.1.7:
The vectors
1
v ,
2
v , ,
m
v are called linearly independent if

c
1
1
v +c
2
2
v ++c
m
m
v = 0

implies that c
1
= 0, c
2
= 0, , c
m
= 0.


9-5
Example 9.1.3:
The vectors (1,0,3) and (0,1,8) of the R
3
are linearly independent. Since for

(1,0,3) + (0,1,8) = (0,0,0)

we get: = 0 and = 0 and 3 + 8 = 0. Therefore these vectors are linearly in-
dependent.

Definition 9.1.8:
The vectors
1
v ,
2
v , ,
m
v are called linearly dependent if

c
1
1
v +c
2
2
v ++c
m
m
v = 0

holds with scalars not all zero.

Example 9.1.4:
The vectors (1,1,3) and (2,2,6) of the R
3
are linearly dependent, since

2(1,1,3) + (-1)(2,2,6) = (0,0,0).

Remark 9.1.2:
The empty set is defined to be linearly independent.

Definition 9.1.9:
Let S be a non-empty subset of a vector space V. S spans V if every vector in V can
be written as a linear combination of (finitely many) elements from S. S is then called
a spanning set or generating set of V. We can define the span of S to be the set of
all linear combinations of elements of S. Then S spans V if and only if V is the span
of S; in general, however, the span of S will only be a subspace of V.

Example 9.1.5:
The set {(1,0,0), (0,1,0), (1,1,0)} spans the space of all vectors in R
3
whose last
component is zero.

Definition 9.1.10:
A linearly independent set in V consisting of a maximum possible number of vectors
in V is called a basis for V.

Example 9.1.6:
(i) The set {(1,0,0), (0,1,0), (0,0,1)} forms a basis for R
3
(called the standard basis for
the vector space R
3
).

9-6
(ii) The real vector space R
3
has {(1,0,0), (0,1,0), (0,0,1)} as spanning set. This
spanning set is actually a basis. Another spanning set for the same space is given
by {(1,2,3), (0,1,2), (1,1/2,3), (1,1,1)}, but this set is not a basis, because it is line-
arly dependent.

Definition 9.1.11:
The dimension of a vector space V is the number of elements in a basis for V and is
denoted by dim(V).

Example 9.1.6:
(i) dim(R
n
) = n.

(ii) The set P
k
(R) of all polynomial functions p
k
: R R, k > 0, of the shape

( )
0 1
1 k
1 k
k
k k
a x a ... x a x a x p + + + +

, a
i
R, 1 i k,

is a vector space (proof this).

Vector addition:
Suppose ( )
0 1
1 k
1 k
k
k k
a x a ... x a x a x p + + + +

and ( )
0 1
1 k
1 k
k
k k
b x b ... x b x b x q + + + +

are elements of P
k
(R), then

( )( ) ( ) ( ) ( ) ( )
0 0 1 1
1 k
1 k 1 k
k
k k k k
b a x b a ... x b a x b a x q p + + + + + + + + +

is an element of P
k
(R).

Scalar multiplication:
Suppose ( )
0 1
1 k
1 k
k
k k
a x a ... x a x a x p + + + +

and r R, then

( )( )
0 1
1 k
1 k
k
k k
a x a ... x a x a x p r r r r r + + + +

is again in P
k
(R). With these algebraic operations P
k
(R) is a vector space.
The set {1, x,, x
k
} is a basis of P
k
(R). Therefore dim(P
k
(R)) = k + 1.

Definition 9.1.12:
Let U and V be two vector spaces and let T: U V be a mapping. Then T is called a
linear transformation (linear operator) if the following conditions are satisfied :

(i) T (u + w) = T (u) + T (w) , for all u , w

U
(ii) T (ru) = rT(u) , r is a scalar and u U

9-7
Example 9.1.7:
(i) Suppose C
1
(a;b) is the set of all real valued functions on the open Interval (a;b)
which have at least first order continuous derivatives.
This set is a vector space and the differential operator
dx
d
is a linear transformation.

(ii) Suppose I is the set of all functions having an antiderivative. This set again is a
vector space and the integral operator
dx is a linear transformation.

In particular our arrow classes defined above are vectors and the set of all arrow
classes embedded in a certain R
n
with the vector addition and scalar multiplication
geometrically defined below is a linear vector space.

Geometrical vector addition: Place the initial point of b
r
at the terminal point of a
r
;
then the sum a
r
+b
r
is the vector drawn from the initial point of a
r
to the terminal point
of b
r
.

Basic properties of vector addition:

(i) a
r
+b
r
= b
r
+a
r
(commutativity)
(ii) ( a
r
+b
r
) + c
r
= a
r
+ (b
r
+c
r
) (associativity)
(iii) a
r
+ 0 = 0 + a
r
= a
r

(iv) a
r
+ (- a
r
) = (- a
r
) + a
r
= 0

Where a
r
denotes the vector having the length | a
r
| and the direction opposite to that
of a
r
.
Fig 9.1.7

9-8

Fig 9.1.8

Geometrical scalar multiplication: Suppose c is a real number. If a
r
0, then ca
r

with c > 0 has the same direction of a
r
and with c < 0 the direction opposite toa
r
. The
length of ca
r
is c times the length of a
r
.

Basic properties of scalar multiplication:

Suppose c, d R, then applies:
(i) c( a
r
+b
r
) = c a
r
+ cb
r

(ii) (c + d) a
r
= ca
r
+ da
r

(iii) c(da
r
) = (cd) a
r

(iv) 1a
r
=a
r

Since it is not convenient and also mathematically not exact to work with these geo-
metrical operations we refer to arrow classes as elements of the vector space R
n

and introduce the components of an arrow (vector).

If a given arrow (vector) a
r
has initial point (tail) P:(x
1
,,x
n
) and terminal point (head)
Q:(y
1
,,y
n
), the n numbers

i i i
x y a , i = 1,,n,

are called the components of the vector a
r
with respect to that coordinate system,
and we write simply [ ]
n 1
a ,..., a a
r
.


9-9

Note 9.1.2:

vector components = head coordinates - tail coordinates

Fig. 9.1.9

Definition 9.1.13:
The length or magnitude of the vector a
r
, denoted by a
r
, is the distance between its
initial point and terminal point. Thus if a given vector a
r
has initial point P:(x
1
,,x
n
)
and terminal point Q:(y
1
,,y
n
), then

(9.1.1) ( ) ( )
2
n
2
1
2
n n
2
1 1
a ... a x y ... x y + + + + a
r

Example 9.1.8:
The vector a
r
with initial point P:(4, 0, 2) and terminal point P:(6, -1, 2) has the com-
ponents 4 6
1
a = 2, 0 1
2
a = -1, 2 2
3
a = 0. Hence a
r
= [2, -1, 0] and the
length of a
r
is given by a
r
= 5 0 ) 1 ( 2
2 2 2
+ + .

Definition 9.1.14:
The position vector r
r
of a point A:(x
1
,,x
n
) is the vector which has initial point at
the origin O:(0,,0) and terminal point at A. Thus r
r
= [x
1
,,x
n
].

point B:(2,4)

point A:(6,-1)
vectora
r
= [-4,5]

9-10

Fig. 9.1.10

Remark 9.1.3:
(i) We use the position vector as a standard representative of an arrow class.

(ii) If n > 3 then the n-dimensional vector [ ]
n 1
a ,..., a a
r
cannot be pictured geometri-
cally as an arrow or a point, but with the exception of the cross product in R
3
(see
Section 9.1.20) vector algebra will be the same whether the vector has 2, 3 or 100
components.

Definition 9.1.15:
A unit vector or normalized vector is a vector of unit magnitude. The unit vector in
the direction of a
r
is denoted by a . Thus a =
a
a
r
r
and 1 a .

The vectors
n 1
e ,..., e
r r
are the unit vectors in the positive directions of the axes of a
Cartesian coordinate system. Hence
{ 1
]
1
,...,0 1 0,...,
position ith
i
e
r
.

Fig. 9.1.11
x
y
z
a
1
1
e
r

a
2
2
e
r

a
3
3
e
r

point (x
1
,x
2
)
position vector [x
1
,x
2
]
y
x
z
2
e
r

1
e
r

3
e
r


9-11

Proposition 9.1.1:
Suppose a
r
R
n
and a
r
= [a
1
,...,a
n
], then applies:

a
r
= a
1
e
r
1
+ ... + a
n
e
r
n
=
n
i
i
a
1
i
e
r

This representation is unique.

Example 9.1.9:
(i) The vector [7,8] can be written as

7e
r
1
+ 8e
r
2
= 7[1,0] + 8[0,1] = [7,0] + [0,8] = [7 + 0,0 + 8] = [7,8]

(ii) The vector [4,3,9] can be written as

4e
r
1
+ 3e
r
2
+ 9e
r
3

= 4[1,0,0] + 3[0,1,0] + 9[0,0,1]

= [4,0,0] + [0,3,0] + [0,0,9]

= [4 + 0 + 0,0 + 3 + 0,0 + 0 + 9]

= [4,3,9].

Definition 9.1.16:
The sum a
r
+b
r
of two vectors a
r
= [a
1,,
a
n
] and b
r
= [b
1,
,
b
n
] is obtained by adding
the corresponding components. That is a
r
+b
r
= [ a
1
+b
1,
,
a
n
+b
n
].

Definition 9.1.17:
Scalar Multiplication: r a
r
= [ra
1
,...,ra
n
].

Example 9.1.10:
Let a
r
= [ 4, 0, 1 ] and b
r
= [ 2, 5, 3 ]. Then

(i) a
r
+ b
r
= [ 6, 5, 4]

9-12
(ii) a
r
- b
r
= a
r
+ (-b
r
) = [4, 0, 1] + [-2, -5, -3] = [2, -5, -2]

(iii) 2( a
r
- b
r
) = [4, -10,-4] = 2a
r
2b
r

Definition 9.1.18:
The scalar or dot product b a
r
r
of two vectors a
r
and b
r
is the product of their
lengths times the cosine of their angle; i.e.:

(i) b a
r
r
= a
r
b
r
cos() if a
r
0 and b
r
0
(ii) b a
r
r
= 0 if a
r
= 0 or b
r
= 0

The angle , 0 , between a
r
and b
r
is measured when the vectors have the
same initial point.

In Fig. 9.1.12 there is shown a triangle with the sides
| a
r
|, | b
r
| and | b a
r
r
|. By the law of cosines we get:

(9.1.2) ( ) cos b a 2 b a b a
2
2
r
r
r
r
r
r
+
2

If [ ]
n 1
a ,..., a a
r
and [ ]
n 1
b ,..., b b
r
then (9.1.2) becomes

(9.1.3) ( ) ( ) ( ) cos b a b a
n
i
i i
n
i
i i
b a
r
r
2
1
2 2
1
2
+

.

Formula (9.1.3) simplifies to

(9.1.4) ( ) cos b a
n
i
i i
b a
r
r
1

or, equivalently,

(9.1.5) ( )
b a
b a
r
r
r
r
cos .

Therefore we get

a
r

b
r

b a
r
r

Fig. 9.1.12

9-13
Proposition 9.1.2:
If [ ]
n 1
a ,..., a a
r
and [ ]
n 1
b ,..., b b
r
are vectors of the R
n
then the scalar product can be
computed by

(9.1.6)
n n 1 1
n
1 i
i i
b a ... b a b a + +

b a
r
r
.

Fig 9.1.13

The sign of ( ) cos determines whether is acute or obtuse. This sign is determined
by the sign of a
r
b
r
since the denominator in (9.1.5) is always positive. In particular
(see also Fig. 9.1.13):

(9.1.7) Sign of the scalar product:
(i) If a
r
b
r
> 0, then 0 < 90

(ii) If a
r
b
r
= 0, then = 90

(iii) If a
r
b
r
< 0, then 90 < 180

As a corollary of (9.1.7) we get for non-zero vectors

Corollary 9.1.1:
Suppose a
r
, b
r
R
n
, a
r
0 and b
r
0. Then applies:

a
r
b
r
= 0 if and only if a
r
and b
r
are perpendicular.

a
r

a
r

a
r

b
r
b
r
b
r

a
r
b
r
> 0 a
r
b
r
= 0 a
r
b
r
< 0

9-14
Example 9.1.11:
If e
r
i
and e
r
j
are the standard unit vectors of R
n
then applies:

(i) 0
j i
e e
r r
if i j

(ii) 1
j i
e e
r r
if i = j

(9.1.7) General Properties of the scalar product:
(i) a
r
b
r
=b
r
a
r

(ii) a
r
(b
r
+c
r
) =a
r
b
r
+a
r
c
r

(iii) a
r
a
r
= | a
r
|
2

(iv) The dot product is a scalar quantity.

Definition 9.1.19:
If a
r
and b
r
are non-zero vectors, then a
r
and b
r
are orthogonal iff a
r
b
r
= 0.

Example 9.1.12:
Find the angle between the vectors a
r
= [1,2,3] and b
r
= [-5,1,1].

Solution:
By Definition 9.1.18 we get: ( )
b a
b a
r
r
r
r
cos
Now a
r
b
r
= [1,2,3] [-5,1,1] = 1(-5) + 21 + 31 = 0 which shows that a
r
and b
r
are
orthogonal and so the angle between them is 90
.

Definition 9.1.20:
The vector product (cross product) a
r
b
r
of two vectors a
r
= [a
1,
a
2
a
3
] and
b
r
= [b
1,
b
2,
b
3
] is a vector c
r
= a
r
b
r
as follows:

(i) If a
r
and b
r
have the same or opposite direction or if one of these vectors is the
zero vector, then c
r
= a
r
b
r
= 0.

(ii) In any other case, c
r
= a
r
b
r
has the length c
r
= a
r
b
r
sin().

9-15
This is the area of the parallelogram with a
r
and b
r
as adjacent sides. The angle ,
0 180, between a
r
and b
r
is measured when the vectors have their initial point
coinciding.

(iii) The direction of c
r
= a
r
b
r
is perpendicular to both a
r
and b
r
and such that a
r
, b
r
,
c
r
, in this order, form a right-handed system; i.e. c
r
shows in the direction of the
thumb if the fingers of the right hand curl from a
r
to b
r
.

The components of the cross product vector
c
r
= [c
1
,c
2
,c
3
] = a
r
b
r
are given by

c
1
= a
2
b
3
a
3
b
2
,

c
2
= a
3
b
1
a
1
b
3
and

c
3
= a
1
b
2
a
2
b
1
.

Therefore:

(9.1.8) a
r
b
r
= [a
2
b
3
a
3
b
2
,

a
3
b
1
a
1
b
3
, a
1
b
2
a
2
b
1
].

The formula in (9.1.8) looks formidable to memorize, but there are simple routines
for finding the cross product. It is convenient to use the determinant notation (see
Section 1.4). We define

(9.1.9)
21 12 22 11
22 21
12 11
a a a a
a a
a a
,

whereby
ij
a , 2 1 i , 2 1 j , are real numbers.

Furthermore we set

(9.1.20)
32 31
22 21
13
33 31
23 21
12
33 32
23 22
11
33 32 31
23 22 21
13 12 11
a a
a a
a
a a
a a
a
a a
a a
a
a a a
a a a
a a a
+ .

whereby
ij
a , 3 1 i , 3 1 j , again real numbers.

Now the cross product can be written as a determinant and expanding this determi-
nant across the first row:

Fig 9.1.14 Vector product

9-16
(9.1.21) a
r
b
r
=
3 2 1
e e e
r r r
r r r
2 1
2 1
3 1
3 1
3 2
3 2
3 2 1
3 2 1
3 2 1
b b
a a
b b
a a
b b
a a
b b b
a a a
e e e
+ ,

where 3 1 i ,
i
e
r
, is the i-th unit basis vector of R
3
.

Note: There is a misusage in (9.1.21) of the determinant notation, since the ele-
ments of a determinant are real numbers and not vectors. But this notation is very
nice to compute the cross product.

The following example shows a short form of (9.1.21).

Example 9.1.13:
If a
r
= [2,1,-4] and b
r
= [3,-2,5] then

a
r
b
r
=
3 2 1
5 2 3
4 1 2
5 2 3
4 1 2
5 2 3
4 1 2
e e e
r r r
.

Therefore:

a
r
b
r
= ( ) ( ) ( ) [ ] 7 22 3 3 4 12 10 8 5
3 2 1
+ + , , e e e
r r r
.

(9.1.22) General properties of the vector product:
(i) b
r
a
r
= ( a
r
b
r
)
(ii) a
r
( b
r
+ c
r
) = ( a
r
b
r
) + ( a
r
c
r
)
(iii) ( a
r
+b
r
)c
r
= (a
r
c
r
) + ( b
r
c
r
)
(iv) a
r
( b
r
c
r
) ( a
r
b
r
)c
r
)
(v) a
r
( b
r
) = ( a
r
b
r
), where is any scalar
(vi) a
r
b
r
= 0 if a
r
and b
r
are parallel.

Definition 9.1.21:
The scalar quantity a
r
(b
r
c
r
) is called the scalar triple product and has the follow-
ing properties:

(i) If a
r
= [a
1,
a
2,
a
3
], b
r
= [b
1,
b
2,
b
3
] and c
r
= [c
1,
c
2,
c
3
] then


9-17
(9.1.23) a
r
(b
r
c
r
) =
3 2 1
3 2 1
3 2 1
c c c
b b b
a a a

(ii) a
r
(b
r
c
r
) = b
r
( c
r
a
r
) = c
r
( a
r
b
r
)

(iii) If any two vectors are equal or parallel, then a
r
(b
r
c
r
) = 0

Geometrically, the scalar triple product a
r
(b
r
c
r
) is the volume of the parallelepiped
with a
r
, b
r
, c
r
as edge vectors.

Fig 9.1.15 Parallelepiped


9-18
9.2 Matrices

A matrix is a rectangular array of real or complex numbers (or functions) enclosed
by square brackets. These numbers (or functions) are called entries or elements of
the matrix. We denote matrices by capital letters A, B, C, or by writing general en-
try in square brackets; i. e.

(9.2.1) A =[
ij
a ], A =[
ij
a ]
mxn
or A =[
ij
a ]
n j m,1 i 1
.

By a (mn)-matrix (read m by n matrix), we mean a matrix with m rows and n col-
umns. Thus a (mn)-matrix is of the form

(9.2.2) A = [
ij
a ] =
1
1
1
1
1
1
1
]
1
mn 1 mn m2 m1
2n 1 2n 22 21
1n 1 1n 12 11
a a ... a a
.
.
.
.
.
.
.
.
. . . .
a a ... a a
a a ... a a
, 1 i m, 1 j n.

If a matrix A has m rows and n columns then A is said to be a rectangular matrix of
order mn. If m = n, A is said to be a square matrix of order n. If A is a square ma-
trix of order n, then the entries a
11
, a
22
,, a
nn
is called the main diagonal of A.

Example 9.2.1:
(i) The matrix
,
_
4 0 2
5 3 1
A is of order 23.

(ii) B =
1
1
1
]
1
4 3 9
2 5 15
46 23 70
is a square matrix of order 3.

A matrix

(9.2.3) A = [a
1
, a
2
,.,a
n
]

with a single row is a (1n)-matrix and is called a row vector.

9-19
Similarly, a single column

(9.2.4) B =
1
1
1
1
1
1
1
]
1
m
2
1
b
.
.
.
b
b

is a (m1)-matrix and is called a column vector.

Definition 9.2.1:
(i) A square matrix in which each element of the main diagonal is an element and
all other elements are zero is called a scalar matrix. Thus

1
1
1
1
1
1
]
1
0 0 0 0
. . . .
. . . .
0 . . 0 0
0 . . 0 0

is a scalar matrix of order n.

(ii) If = 1 the scalar matrix is called the identity matrix of order n and is denoted by
I
n
or simply by I .

Definition 9.2.2:
An upper (lower) triangular matrix is a matrix whose elements below (above) the
main diagonal are all zero. Thus

1
1
1
1
1
1
]
1
mn
n 2 22
n 1 12 11
a . . . 0 0
. . .
. . .
a . . . a 0
a . . . a a

is an upper triangular matrix.

9-20
Definition 9.2.3
A square matrix, all of whose elements are zero except those in the main diagonal,
is called diagonal matrix. Thus

0 a ,
a ... 0 0 0
. . . .
. . . .
0 ... 0 a 0
0 ... 0 0 a
ii
nn
22
11
1
1
1
1
1
1
]
1
,

is a diagonal matrix.

Remark 9.2.1:
Every scalar matrix is a diagonal matrix.

Definition 9.2.4:
Two matrices A = [a
ij
] and B = [b
ij
] are equal, written A = B, if and only if they have
same order and the corresponding entries are equal, that is, a
ij
= b
ij
for each i and
for each j.

Example 9.2.2:
Let

A =
1
]
1
d c
b a
and B =
1
]
1
0 6
1 4

Then A = B if and only if a = 4, b = 1, c = 6 and d = 0.

Definition 9.2.5:
A matrix having all elements equals zero is called a zero matrix. If it has m rows
and n columns, we denote it by O
mn
or simply by O.

Definition 9.2.6:
Two matrices A and B are said to be conformable for addition if they have the
same number of rows and the same number of columns. Thus if A = [a
ij
] and B = [b
ij
]
are (mn)-matrices then their sum is the matrix

A + B = [a
ij
+ b
ij
]

9-21
of order (mn). That is, we add the corresponding elements of A and B to obtain the
sum A + B.

Definition 9.2.7:
(i) Given a (mn)-matrix A = [a
ij
], we define A = [-a
ij
]. Then A is a (mn)-matrix and
by definition

A + (-A) = (-A) + A = O

Thus -A is the inverse of A with respect to addition. That is, -A is the additive in-
verse of A.

(ii) Also O + A = A + O = A that is, O is the neutral element with respect to addition.

Definition 9.2.8:
We define subtraction of two matrices A and B of same order as

A - B = A + (-B)

Example 9.2.3;
Let

A =
1
]
1
1 0 4
3 1 2
and B =
1
]
1

0 1 0
2 1 1
.

Then A B = A + (-B) =
1
]
1
1 0 4
3 1 2
+
1
]
1

0 1 0
2 1 1

=
1
]
1
+ +
+
0 1 1 0 0 4
2 3 1 1 1 2

=
1
]
1
1 1 4
5 2 1

Definition 9.2.9:
Scalar Multiplication (Multiplication by a number): The product of any (mn)-matrix
A = [a
ij
] and any scalar r, written rA, is the (mn)-matrix

rA = [ra
ij
]

9-22
obtained by multiplying each entry in A by r.

Example 9.2.4:
Let
A =
1
]
1
5 3
1 2
.

Then 2A =
1
]
1
10 6
2 4
.

Proposition 9.2.1:
Let A, B be (mn)-matrices and r, s are scalars then:

(i) (r + s)A = rA + sA
(ii) r(A +B) = rA + rB
(iii) r(sA) = (rs)A
(iv) 1A = A

Remark 9.2.2:
The set of all (mn)-matrices together with the matrix addition and scalar multiplica-
tion form a vector space R
mxn
(or C
mxn
if the entries are complex numbers). The di-
mension of this vector space is given by dim(R
mxn
) = mn.

Definition 9.2.10:
The product C = AB of a (mn)-matrix A = [a
ij
] and a (pq)-matrix B = [b
ij
] is defined
if and only if p = n, that is,

number of rows of 2
nd
factor B = Number of columns of 1
st
factor A,

and is then defined as the (mq)-matrix C = [c
ij
] with entry

nj in
n
1 k
2j i2 1j i1 kj ik ij
b a ........ b a b a b a c

+ + + .

In other words: c
ij
is the sum of the products of the corresponding elements of the
i-th row of A and j-th column of B.

9-23
Example 9.2.5:

Let A =
1
1
1
]
1

0 3 5
7 4 2
1 0 1
and B =
1
1
1
]
1
3 2
4 0
1 6
.

Then AB=
1
1
1
]
1

0 3 5
7 4 2
1 0 1
1
1
1
]
1
3 2
4 0
1 6

=
1
1
1
]
1
+ + + +
+ + + +
+ +
) 3 ( 0 ) 4 ( 3 ) 1 ( 5 ) 2 ( 0 ) 0 ( 3 ) 6 ( 5
) 3 ( 7 ) 4 ( 4 ) 1 ( 2 ) 2 ( 7 ) 0 ( 4 ) 6 ( 2
) 3 ( 1 ) 4 ( 0 ) 1 ( 1 ) 2 ( 1 ) 0 ( 0 ) 6 ( 1

=
1
1
1
]
1
17 30
39 2
2 8

Hence A is (33)-matrix, B is a (32)-matrix and AB is (32)-matrix. Note that the
product BA is not defined.

Note 9.2.1:
Matrix multiplication is not commutative, AB BA in general.

Example 9.2.6:
Let A =
1
]
1
4 3
2 1
and B =
1
]
1
2 0
1 1
.

Then AB =
1
]
1
11 3
5 1
and BA =
1
]
1
8 6
6 4
.

Thus AB BA.

Note 9.2.2:
We may have AB = O when neither A = O nor B = O.

9-24
Example 9.2.7:
Let A =
1
1
1
]
1
0 4 1
0 1 1
0 2 1
and B =
1
1
1
]
1
9 4 1
0 0 0
0 0 0
.

Then AB =
1
1
1
]
1
0 0 0
0 0 0
0 0 0
.

But neither A = O nor B = O.

Note 9.2.3:
The cancellation laws do not hold for matrices. That is, we may have AB = AC (or
BA = CA) when B C.

Example 9.2.8:
Let A =
1
1
1
]
1
0 4 1
0 1 1
0 2 1
, B =
1
1
1
]
1
2 2 2
1 1 1
3 2 1
and C =
1
1
1
]
1
1 1 1
1 1 1
3 2 1
.

Then AB =
1
1
1
]
1
7 2 3
2 3 2
1 4 3
and AC =
1
1
1
]
1
7 2 3
2 3 2
1 4 3
.

Thus AB = AC but B C .

Proposition 9.2.2:
If the matrices A, B and C are conformable for the indicated sums and products
then:

(i) A(BC) = (AB)C (Associative Law)
(ii) A(B + C) = AB + AC (Left Distributive Law)
(iii) (A + B)C = AC + BC (Right Distributive Law)
(iv) r(AB) = (rA)B = A(rB), where r is any scalar.

Definition 9.2.11:
The transpose of a (mn)-matrix A = [a
ij
], denoted by A
T
, is a (nm)-matrix obtained
by interchanging rows and columns of A.

9-25
Example 9.2.9:
Let A =
1
1
1
]
1

0 3 5
7 4 2
1 0 1
and B =
1
1
1
]
1
3 2
4 0
1 6
.

Then A
T
=
1
1
1
]
1
0 7 1
3 4 0
5 2 1
and B
T
=
1
]
1

3 4 1
2 0 6
.

Proposition 9.2.3:
If the matrices A and B are conformable for the sum A + B and the product AB, then:

(i) (A + B)
T
= A
T
+ B
T

(ii) (A
T
)
T
= A
(iii) (rA)
T
= rA
T
, where r is a scalar.
(iv) (AB)
T
= B
T
A
T

Definition 9.2.12:
(i) A square matrix A for which A
k+1
= A, (k being a positive integer), is called peri-
odic. If k is the least positive integer for which A
k+1
= A, then A is said to be of pe-
riod k.

(ii) A square matrix A for which A
2
= A is called idempotent.

(iii) A square matrix A for which A
p
= O, (p being a positive integer), is called nilpo-
tent. If p is the least positive integer for which A
p
= O, then A is said to be nilpotent
of index p.

(iv) A square matrix A for which A
2
= I is called involutory matrix (an involutory ma-
trix is its own matrix inverse).

(v) A square matrix A for which A
T
= A is called symmetric matrix.

(vi) A square matrix A for which A
T
= -A is called skew-symmetric matrix.

(vii) If A is a matrix over C and its elements are replaced by their complex conju-
gates, then the resulting matrix is called conjugate of A denoted by A

(to be read A
conjugate).

(viii) A square matrix A such that ( A)
T
= A is called Hermitian matrix (the name
comes from an old French mathematician by the name of Hermite).


9-26
(ix) A square matrix A such that ( A)
T
= -A is called skew Hermitian matrix.
(x) The inverse of a (nn)-matrix A is denoted by A
-1
and is a (nn)-matrix such that
AA
-1
= A
-1
A = I , where I is the (nn)-unit matrix. If A has an inverse, then A is called
non-singular matrix. If A has no inverse, then A is called singular matrix.

(xi) A square matrix A is called orthogonal if A
T
= A
-1
.

(xii) A square matrix A is called unitary if ( A)
T
= A
-1
.

Note 9.2.4:
(i) Only square matrices can have inverses.

(ii) A real matrix is Hermitian iff it is symmetric.

Remark 9.2.2:
The inverse of a matrix, if it exists, is unique. Suppose A has two inverses, say B
and C. Then AB = BA = I and AC = CA = I , so that we obtain the uniqueness from
B = I B = (CA)B = C(AB) = CI = C.

Example 9.2.8:
(i) If A =
1
1
1
]
1
+

1 0
0 2 1
0 2 1 1
j
j j
j
, then A =
1
1
1
]
1
+
1 0
0 2 1
0 2 1 1
j
j j
j
and

( A)
T
=
1
1
1
]
1
+

1 0
0 2 1
0 2 1 1
j
j j
j
. Therefore A is a Hermitian matrix.

(iii) Let A =
1
1
1
]
1

1 1 1
0 1 1
3 2 6
and B =
1
1
1
]
1
4 2 1
2 3 2 3 2 1
2 3 2 1 2 1
.

Now AB =
1
1
1
]
1
1 0 0
0 1 0
0 0 1
= I
3
and BA =
1
1
1
]
1
1 0 0
0 1 0
0 0 1
= I
3
.


9-27
Hence B = A
-1
. Also A = B
-1

(iv) A =
1
1
1
]
1
2 1 2
2 2 1
1 2 2
3
1
is an orthogonal matrix and

B =
1
1
1
1
1
1
1
]
1
j
j j
0 0
0
2
1
2
1
0
2
1
2
1
is a unitary matrix.

Proposition 9.2.4:
Let A and B be non-singular matrices of the same order n, then AB is non-singular
and (AB)
-1
= B
-1
A
-1

Proof:
If we show that (AB) (B
-1
A
-1
) = I = (B
-1
A
-1
) (AB) then we prove that AB is non-singular
and that its inverse is B
-1
A
-1
. Now with Proposition 9.2.2 (AB) (B
-1
A
-1
) = A(B B
-1
)A
-1
=
AI A
-1
= A A
-1
= I and (B
-1
A
-1
)(AB) = B
-1
(A
-1
A)B = B
-1
I B = B
-1
B = I . Thus the product
AB of two non-singular matrices A and B is non-singular and (AB)
-1
= B
-1
A
-1
.

(9.2.5) Elementary Row Operations:
The following operations on a matrix are called elementary row operations:

(i) Interchange of any two rows
(ii) Multiplication of a row by any non-zero scalar
(iii) Addition of any multiple of one row to another row.

Any (mn)-matrix B is called row equivalent to a (mn)-matrix A if B is obtained
from A by performing a finite sequence of elementary row operations on A. We write

B
R
~ A

to denote B is row equivalent to A.


9-28
Definition 9.2.13:
A (mn)-matrix A is said to be an echelon matrix if it has the following properties:

(i) The first k rows of A are non-zero; the remaining (m - k) rows are zero (k m).
(ii) The number of zeros preceding the first non-zero entry in each non-zero row is
larger than the number of zeros that appear before the first non-zero element in
any preceding row. In other words: Each non-zero row in the matrix starts with
more zeros then the previous row.

In an echelon matrix, the first non-zero entry of a row is called pivot. A column con-
taining pivot is called a pivot column. If a matrix in echelon form has the additional
property that each pivot is 1 and every other entry of the pivot column is zero, then it
is said to be in reduced echelon form.

Note 9.2.5:
There is at least one pivot in each row and in each column of an echelon matrix.

Example 9.2.9:
The matrices

A =
1
1
1
]
1
6 0 0 0
1 9 0 0
3 2 5 0
and B =
1
1
1
]
1
1 0 0 0
0 1 0 0
0 0 3 1

are in echelon form. The second matrix is in reduced echelon form.

Notation:
The following notations prove useful in numerical problems:

(i) R
ij
denotes interchange of ith and jth rows of a matrix
(ii) rR
i
denotes the multiplication of ith row of a matrix by a non-zero scalar r
(iii) R
i
+rR
j
denotes addition of r times the elements of jth row in the corresponding
elements of ith row.

Theorem 9.2.1:
Any matrix is row equivalent to a matrix in echelon form (reduced echelon form).

Proposition 9.2.5:
A square matrix A of order n is non-singular if and only if it is row equivalent to the
identity matrix I
n
.

9-29
Proposition 9.2.6:
If a square matrix A is reduced to the identity matrix by a sequence of elementary
row operations, the same sequence of operations performed on the identity matrix
produces the inverse matrix A
-1
of A.

Remark 9.2.3:
Proposition 9.2.6 explains a method to evaluate the inverse of a given matrix. This
is illustrated in the following example.

Example 9.2.10:
Find the inverse of the matrix

A =
1
1
1
]
1
0 3 1
1 4 2
3 0 1

Solution:
Performing single row operation at a time and exhibit the successive matrices row
equivalent to A in left hand column and successive matrices row equivalent to I in
the right hand column.

A I
3

1
1
1
]
1
0 3 1
1 4 2
3 0 1

1
1
1
]
1
1 0 0
0 1 0
0 0 1

1
1
1
]
1
0 3 1
5 4 0
3 0 1

1
1
1
]
1
1 0 0
0 1 2
0 0 1
R
2
2R
1

1
1
1
]
1
3 3 0
5 4 0
3 0 1

1
1
1
]
1
1 0 1
0 1 2
0 0 1
R
3
R
1

1
1
1
]
1
1 1 0
5 4 0
3 0 1

1
1
1
]
1
3 1 0 3 1
0 1 2
0 0 1
(-1/3)R
3


9-30
1
1
1
]
1
1 1 0
0 1 0
3 0 1

1
1
1
]
1

3 1 0 3 1
3 5 1 3 1
0 0 1
R
2
+ 5 R
3

1
1
1
]
1
1 0 0
0 1 0
3 0 1

1
1
1
]
1

3 4 1 3 2
3 5 1 3 1
0 0 1
R
3
R
2

1
1
1
]
1
1 0 0
0 1 0
0 0 1

1
1
1
]
1

3 4 1 3 2
3 5 1 3 1
4 3 1
R
1
3R
3

1
1
1
]
1
1 0 0
0 1 0
0 0 1

1
1
1
]
1

3 4 1 3 2
3 5 1 3 1
4 3 1
(1)R
2

Hence A
-1
=
1
1
1
]
1

3 4 1 3 2
3 5 1 3 1
4 3 1
.

Definition 9.2.14:
The maximum number of linearly independent row vectors of a matrix A=[a
ij
] is called
the rank of A and is denoted by rank(A).

Proposition 9.2.7:
The non-zero rows of a matrix in echelon form are linearly independent.

Remark 9.2.4:
(i) The rank of a matrix A equals the maximum number of linearly independent col-
umn vectors of A. Hence A and its transpose A
T
have the same rank.

(ii) Row equivalent matrices have the same rank. The rank of a matrix is equal to the
number of non-zero rows in its echelon form.

9-31
Example 9.2.11:
Find the rank of

A =
1
1
1
]
1
3 5 1
6 5 3
3 9 5
.

Solution:
A
R
~
1
1
1
]
1

3 9 5
6 5 3
3 5 1
R
13

R
~
1
1
1
]
1
3 9 5
6 5 3
3 5 1
(1)R
1

R
~
1
1
1
]
1
12 16 0
15 20 0
3 5 1
R
2
+ 3R
1
and R
3
5R
1

R
~
1
1
1
]
1
12 16 0
4 3 1 0
3 5 1
(1/20)R
2

R
~
1
1
1
]
1
0 0 0
4 3 1 0
3 5 1
R
3
+ 16R
2

This is an echelon form of matrix A and the number of its non-zero rows is 2.
Hence rank(A) = 2.

Note: A general algorithm to reduce a given Matrix to an echelon form is the Gauss
elimination method which will be introduced in the next section.


9-32
In Example 9.2.11 the rank of a matrix has been found by reducing it to an echelon
form. To reduce a matrix A to an echelon form many row operations are needed and
in the process fractions creep in which make the computations awkward and cum-
bersome (see the above example) excluded you use a computer program. The fol-
lowing new method is very elegant involving no fractions and it also yields an eche-
lon matrix that is row equivalent to A and whose non-zero rows constitute a basis for
the row space of a given matrix A.

Proposition 9.2.8:
Let A = [a
ij
] be a (mn)-matrix with a
11
0. We define determinants d
ij
of order 2 as
follows:

For 2 i m and 2 j n, we set

d
ij
=
i1 1j ij 11
ij i1
1j 11
a a a a
a a
a a
.

Then applies:

( )
,
_
1
1
1
1
1
1
]
1
+
mn 3 m 2 m
n 3 33 32
n 2 23 22
d ... ... d d
. . .
. . .
d ... ... d d
d ... ... d d
rank rank 1 A .

This method is illustrated by the following example.

Example 9.2.12:
Find the rank of the matrix

A =
1
1
1
]
1
6 1 3 8
4 0 2 7
2 1 5 3
.

Also write an echelon matrix row equivalent to A.

9-33
Solution:

rank(A) = 1 + rank
,
_
1
1
1
1
]
1

6 8
2 3
1 8
1 3
3 8
5 3
4 7
2 3
0 7
1 3
2 7
5 3

= 1 + rank

,
_
1
]
1

2 5 31
2 7 41

= 2 + rank

,
_
1
]
1

2 31
2 41
5 31
7 41

= 2 + rank([12 144])

= 3

An echelon matrix row equivalent to A is

1
1
1
]
1
144 12 0 0
2 7 41 0
2 1 5 3
.

9-34
9.3 Systems of linear equations

The theory of matrices has been usefully employed in various branches of pure as
well as applied mathematics. In what follows, we shall apply this theory to solution of
m linear equations in n unknowns. Consider the m equations:

(9.3.1)
'
+ + +
+ + +
+ + +
m n mn 2 2 m 1 1 m
2 n n 2 2 22 1 21
1 n n 1 2 12 1 11
b x a ... x a x a
.
.
.
b x a ... x a x a
b x a ... x a x a

in n unknowns x
1
, x
2
, ., x
n
, where a
ij
and b
i
are scalars, i = 1, 2,, m;
j = 1, 2, , n. Using the matrix notation, the system (9.3.1) can be written as

(9.3.2) Ax = b,

where A, x, b are the following matrices.:

(9.3.3) A =
1
1
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
, x =
1
1
1
1
1
1
]
1
n
2
1
x
.
.
x
x
, b =
1
1
1
1
1
1
]
1
m
2
1
b
.
.
b
b
.

The matrix A is called the matrix of the coefficients of the system of equations, the
columns of constants b
i
forms a column vector b of order m and the unknowns x
j

form the column vector x of order n. Thus equation (9.3.2) can be written as:

(9.3.4)
1
1
1
1
1
1
1
]
1
1
1
1
1
1
1
1
]
1
1
1
1
1
1
1
1
]
1
m
2
1
n
2
1
mn 2 m 1 m
n 2 22 21
n 1 12 11
b
.
.
.
b
b
x
.
.
.
x
x
a ... a a
.
.
.
.
.
.
.
.
.
a ... a a
a ... a a


9-35
Definition 9.3.1:
An equation of the type given in (9.3.1) above with b 0, is called a system of non-
homogeneous linear equations. If b = 0, then the system of equations (9.3.1) is
known as a system of homogeneous linear equations.

Definition 9.3.2:
(i) A solution of (9.3.1) is a set of real numbers {x
1
, x
2
, , x
n
} that satisfies all the m
equations.

(ii) A solution depending on parameters
i
R, i = 1, ,k, is called a general solu-
tion of (9.3.1). For instance

( ) ( ) ( ) ( ) 3,0,1,4 2,1,0,0 ,4 , , 3 2 x , x , x , x
2 1 2 2 1 2 1 4 3 2 1
+ ,
1
,
2
R

is a general solution of a linear system in 4 unknowns. Each special choice of the
parameters leads to a particular solution. Setting
1
= 1 and
2
= 0 in our example
we get the particular solution 0 x x 1, x 2, x
4 3 2 1
.

Example 9.3.1:
(i) The linear system

'

+

+
0
0 5 7 3
0 5 3 2
0 2 4
4 3 2
4 3 2 1
4 3 2 1
3 2 1
x x x
x x x x
x x x x
x x x

has the general solution

( ) ( ) ( ) 4,1,0,1 2,1,1,0 x , x , x , x
2 1 4 3 2 1
+ ,
1
,
2
R.

For
1
= 1 and
2
= 0 we get the particular solution 0 x 1, x 1, x 2, x
4 3 2 1
. Us-
ing other values for the both parameters you can find infinitely many particular solu-
tions.

(ii) The matrix notation of the system of equations

'
+ +
+
+
3 x 7 x 5 x 3
2 x x
1 x x 3 x 2
3 2 1
2 1
3 2 1

is given by

9-36

1
1
1
]
1
1
1
1
]
1
1
1
1
]
1
3
2
1
7 5 3
0 1 1
1 3 2
3
2
1
x
x
x
.

Definition 9.3.3:
The matrix

(1.3.5) A
b
=
1
1
1
1
]
1
m mn 2 m 1 m
2 n 2 22 21
1 n 1 12 11
b a ... a a
b a ... a a
b a ... a a
M M M M

is called the augmented matrix of the system (9.3.1).

The augmented matrix is quite important in the sense that its rank and rank of A (the
matrix of coefficients) determine whether the system of equations Ax = b does or
does not have a solution.

The Gauss elimination (named after its inventor, the German mathematician C. F.
Gauss (1777-1855)) is a standard method for solving linear systems. It is a system-
atic elimination process, a method of great importance that works in practice.

For instance, to solve the system

'
+
+
18 3x 4x
2 5x 2x
2 1
2 1

we multiply the first equation by (-2) and add it to the second, obtaining

'
+
14 7x -
2 5x 2x
2
2 1

This is Gauss elimination for 2 two equations. The solution now follows by back sub-
stitution:

2 - x
2
and ( ) ( ) 6 /2 2 5 2 x
1


9-37
Remark 9.3.1:
Since a linear system is completely determined by its augmented matrix, the elimina-
tion process can be done by merely considering this matrix. We apply row opera-
tions on the augmented matrix to reduce it to an echelon form from which we shall
then readily obtain the values of the unknowns by back substitution.

The general Gauss-elimination-method
First elimination step: If
11
a = 0 interchange rows until 0 a
11
. Multiply the first row
by
11
21
a
a
and add it to the second row. Multiply the first row by
11
31
a
a
and add it to
the third row etc. By these elementary row operations we get a new matrix that is
row equivalent to the starting matrix (9.3.5) and has the shape

,
_
m mn 2 m
2 n 2 22
1 n 1 12 11
c d ... d 0
c d ... d 0
b a ... a a
M M M M
.

Second elimination step: For the new submatrix

,
_
m mn 3 m 2 m
3 n 3 33 32
2 n 2 23 22
c d ... d d
c d ... d d
c d ... d d
M M M M

we perform again the elementary row operations described above. That is: If
22
d = 0
interchange rows until 0 d
22
. If such a row does not exist the second step is fin-
ished. If yet, multiply the first row by
22
32
d
d
and add it to the second row etc. After the
second step the starting matrix (9.3.5) has the shape

,
_
m mn 3 m
2 n 2 23 22
1 n 1 13 12 11
c
~
e ... e 0 0
...
0
c
~
e ... e e 0
b a ... a a a
M M M
M
M

After at most (m -1) steps we get matrix of the shape

9-38
(9.3.7) ( )
r m
r
d
d
d
d
d
... ...
...
... ...
* ... * ...
* * * *
* * * * * *
* ... * * * * * *
A G
m
r
r
b
1
1
1
1
1
1
1
1
1
1
1
]
1
+
M M M M M
M M M
1
2
1
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0
0 0
0
,

which is row equivalent to the augmented matrix A
b
. The places marked by * are real
numbers calculated by the Gauss elimination method and r = rank(A). If m > r and a
scalar
m 1 r
d ,..., d
+
in (9.3.7) is different from zero then the system of equations Ax = b
has no solution. If e.g. 0 d
1 r

+
, then we get the contradiction

0 d x 0 ... x 0 x 0 0
1 r n 2 1
+ + +
+
.

If 0 d .... d
m 1 r

+
, we can compute the solution of (9.3.1) from (9.3.7) by means of
back substitution. If k coefficients in the r-th equation are different from zero, then
there are (k -1) degrees of freedom; i.e. we can choose (k - 1) unknowns freely. After
choosing these free variables the k-th variable can be computed. Again by means of
back substitution we get the solution. If (k 1) > 0, then the system (9.3.1) has infi-
nitely many solutions.

Proposition 9.3.1:
Let Ax = b be a system of m linear equations in n unknowns.

(i) If m = n, and the matrix A is non-singular, then the system has the unique solution

(9.3.8) x = A
-1
b.

(ii) The non-homogeneous system has a solution if and only if rank(A) = rank(A
b
);
i.e. the rank of the coefficient matrix A is equal to the rank of augmented matrix A
b
.

(iii) The homogeneous system Ax = 0 has a non-trivial solution if m < n.

(iv) The homogeneous linear system Ax = 0 has the unique solution
0 x ... x x
n 2 1
if and only if rank(A) = n.

(v) The homogeneous system Ax = 0 has a non-trivial solution if and only if
rank(A) < n.

9-39
(vi) The solution of the homogeneous linear system Ax = 0 has (n rank(A)) un-
knowns that can be chosen freely. In other words: The general solution of Ax = 0
has (n rank(A)) parameters.

(vii) If the system Ax = b has a solution then the general solution is given by

( ) ( ) ( )
n 2 1 n 2 1
u ,..., u , u x ,..., x , x x ,..., x , x
n 2 1
+
0 0 0

where ( )
0 0 0
n 2 1
x ,..., x , x is a particular solution of the inhomogeneous system Ax = b
and ( )
n 2 1
u ,..., u , u is the general solution of the assigned homogeneous system
Ax = 0.

Proof:
We give only the proof for (i). Since A is non-singular, its inverse A
-1
exists. So

Ax = b

can be written as

A
-1
Ax = A
-1
b.

Since A
-1
A = I
n
, the above equation becomes

I
n
x = x = A
-1
b,

giving the solution vector x = A
-1
b. To show the uniqueness, let y be a second solu-
tion. Then

Ay = b and so y = A
-1
b = x.

The Gauss-elemination-method is illustrated with the help of the following example.

Example 9.3.2:
(i) Solve the systems of equations:

(9.3.6)
'
+ +
+ +
+
1 - x x x
1 2x x 4x
0 2x x - x
3 2 1
3 2 1
3 2 1


9-40
Solution:
Here
A =
1
1
1
]
1

1 1 1
2 1 4
2 1 1
and A
b
=
1
1
1
]
1
1 1 1 1
1 2 1 4
0 2 1 1
.

We apply row operations on A
b
to reduce it to an echelon form. Thus

A
b
=
1
1
1
]
1
1 1 1 1
1 2 1 4
0 2 1 1

R
~
1
1
1
]
1
1 1 2 0
1 6 5 0
0 2 1 1
R
2
+

(-4)R
1
and R
3
+ (-1)R
1

R
~
1
1
1
]
1
5 7 5 7 0 0
1 6 5 0
0 2 1 1
/ /
R
3
+ (-2/5)R
2

By back substitution we get:

x
3
= - 1,

5x
2
6x
3
= 5x
2
6(-1) =

1 and

x
1
- x
2
+ 2x
3
= x
1
- (-1)

+ 2(-1) = 0.

Therefore x
1
= 1, x
2
= 1, x
3
= 1 is a solution of (9.3.6).

The following row operations transform the augmented matrix to a matrix into re-
duced echelon form without fractions. These transformations are based on experi-
ence and are not suitable for a computer program (but a computer has no problems
with fractions). The solution remains of course the same.

A
b
=
1
1
1
]
1
1 1 1 1
1 2 1 4
0 2 1 1


9-41
R
~
1
1
1
]
1
1 1 2 0
1 6 5 0
0 2 1 1
R
2
+

(-4)R
1
and R
3
+ (-1)R
1

R
~
1
1
1
]
1
1 1 2 0
3 4 1 0
0 2 1 1
R
2
+

(-2)R
3

R
~
1
1
1
]
1
7 7 0 0
3 4 1 0
0 2 1 1
R
3
+

(-2)R
2

R
~
1
1
1
]
1
1 1 0 0
3 4 1 0
0 2 1 1
(1/7)R
3

By back substitution we get:

x
3
= -1,

x
2
4x
3
= x
2
4(-1) =

3 and

x
1
- x
2
+ 2x
3
= x
1
- (-1)

+ 2(-1) = 0.

Therefore we get the same solution x
1
= 1, x
2
= 1, x
3
= 1.

Definition 9.3.4:
If the equation Ax = b has a solution, then it is called a consistence system other-
wise it is termed as inconsistence system. A linear system Ax = b is called over
determined if m > n, that is, it has more equations than unknowns. A linear system
Ax = b is called determined if m = n, that is, the number of equations equals to the
number of unknowns. A linear system Ax = b is called under determined if
m < n, that is, it has fewer equations than unknowns.

Example 9.3.3:
Examine the following system for a non-trivial solution:

(9.3.9)
'
+ + +
+ +
+ +
0 2x 2x x 4x
0 x 2x 3x
0 x 2x x - x
4 3 2 1
4 2 1
4 3 2 1


9-42
Solution:
The matrix of coefficients is

A =
1
1
1
]
1

2 2 1 4
1 0 2 3
1 2 1 1
.

We bring A into the echelon form by the following elementary row operations:

R
~

1
1
1
]
1
1 2 1 1
1 0 2 3
2 2 1 4
R
13

R
~
1
1
1
]
1
1 2 1 1
1 0 2 3
2 1 2 1 4 1 1
(1/4)R
1

R
~
1
1
1
]
1

2 1 2 3 4 5 1
2 1 2 3 4 5 0
2 1 2 1 4 1 1
R
2
- 3R
1
and R
3
- R
1

R
~

1
1
1
]
1

2 1 2 3 4 5 0
5 2 5 6 1 0
2 1 2 1 4 1 1
(4/5)R
2

R
~
1
1
1
]
1

0 0 0 0
5 2 5 6 1 0
5 3 5 4 0 1
R
1
- (1/4)R
2
and R
3
+ (5/4)R
2

The rank of the matrix A is 2 which is less than number of unknowns 4. Hence the
system has a non-trivial solution. The first two rows of the last matrix give the follow-
ing relations. In equation 2 we have 2 degrees of freedom. Assigning arbitrary val-
ues to x
3
and x
4
, we find the corresponding values of x
1
and x
2
. Therefore system
(9.3.9) has infinitely many solutions.


9-43
9.4 Determinants

Determinants were originally introduced for solving linear systems. Although imprac-
tical in computations, they have important engineering applications in eigenvalue
problems, differential equations, vector algebra etc. They can be introduced in sev-
eral equivalent ways. Our definition is particularly practical in connection with linear
systems.

An n-th-order determinant is an expression associated with an (nn)-matrix (square
matrix) A = [a
ij
], as we now explain, beginning with n = 2.

Definition 9.4.1:
A determinant of second order is defined by

(9.4.1) ( )
21 12 22 11
22 21
12 11
a a a a
a a
a a
A det

So here we have bars, whereas a matrix has brackets.

Remark 9.4.1:
The determinant det(A) of order 2 is a real number associated with a square matrix
A of order 2. So, in this case, we may regard det:R
2x2
R as a function whose do-
main is the set of all square matrices of order 2 and whose range is the set of the
real numbers.

Example 9.4.1:
(i) 1 6 5 2 3 5 1
5 2
3 1
det
,
_
1
]
1

(ii) 10 14 24 2 7 8 3
8 2
7 3

Definition 9.4.2:
A determinant of third order can be defined by

( )
23 22
13 12
31
33 32
13 12
21
33 32
23 22
11
33 32 31
23 22 21
13 12 11
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 det .

9-44
Definition 9.4.3:
(i) Let A be a square matrix of order n. A matrix obtained from A by deleting
its i-th row and j-th column is again a matrix A
ij
of order (n -1). A
ij
is called the ij-th
minor of A.

(ii) Let A
ij
be the ij-th minor of a square matrix A of order n. Then C
ij
= (-1)
i+j
det(A
ij
) is
called the ij-th cofactor of A.

Example 9.4.2:

Let det(A) =
33 32 31
23 22 21
13 12 11
a a a
a a a
a a a
be a determinant of order 3. Then applies:

det(A
11
)
=
33 32
23 22
a a
a a
and det(A
23
)=
32 31
12 11
a a
a a
.

Definition 9.4.4 (Cofactor expansion for a determinant of order n):
For the matrix

A = [
ij
a ] =
1
1
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
, i = 1,2,,n and j = 1,2,,n

of order n, we define det(A) for a given i {1,,n} by

(9.4.2) det(A) = a
i1
C
i1
+ a
i2
C
i2
+.+ a
in
C
in
.

This expression is called an expansion of det(A) by cofactors of ith row of A.
We can also write (9.4.2) in terms of minors as

(9.4.3) det(A) = ( )
+
n
1 j
ij
j i
A 1 -
ij
a , i {1,,n}.


9-45
Remark 9.4.2:
(i) Although the above technique to evaluate determinant of a (nn)-matrix seems
quite straight forward, yet, in practice, it is very laborious to work with when n > 3
because it involves a lot of calculations. We shall later discuss methods which sim-
plify these calculations.

(ii) For some special type of matrices, their determinants can be easily evaluated.
For instance, if a matrix A is triangular that is, all its entries above or below the main
diagonal are zero, then its determinant is the product of the elements on the main
diagonal.

(iii) Analog to expanding a determinant of a square matrix A across the i-th row it can
be expanded across the j-th column. In this case we get:

(9.4.4) det(A) = ( )
+
n
1 i
ij
j i
A 1 -
ij
a , j {1,,n}.

Example 9.4.3:
(i) We compute the determinant of the matrix

1
1
1
1
]
1
1 3 4 2
2 3 6 0
0 1 2 3
1 0 2 1
A .

Expanding det(A) by cofactors of the first column leads to

1
1
1
]
1
1 3 4
2 3 6
0 1 2
A
11
,
1
1
1
]
1

1 3 4
2 3 6
1 0 2
A
21
,

1
1
1
]
1

1 3 4
0 1 2
1 0 2
A
31
,
1
1
1
]
1

2 3 6
0 1 2
1 0 2
A
41
.

Therefore we get:

( ) ( ) ( ) ( ) ( )
( )
72
16 2 24 3 32

+
41 31 21 11
A det 2 A det 0 A det 3 A det 1 A det


9-46
(ii) The determinant of the matrix

1
1
1
1
]
1
7 0 0 0
3 0 0
0 1 0
1 0 2 1
A
7
2

is given by det(A) = 1237 = 42.

(9.4.5) General properties of determinants:
(i) Interchange of any two rows (columns) multiplies the value of determinant
by (-1).
(ii) Addition of a multiple of a row (column) to another row does not alter the value
of the determinant.
(iii) Multiplication of a row (column) by a scalar r multiplies the value of the determi-
nant by r.
(iv) If the two rows (columns) are identical then value of determinant will be zero.
(v) A zero row (column) renders the value of a determinant zero.
(vi) For any (nn)-matrices A and B, det(AB) = det(BA) = det(A)det(B).

(9.4.6) An Algorithm to evaluate det(A):
By an algorithm we mean a sequence of a finite number of steps to get a desired re-
sult. The step by step evaluation of det(A) of order n is obtained as follows:

Step 1:
By an interchange of rows of A bring a non-zero entry to position (1,1).

Step 2:
By adding suitable multiples of the first row to all the other rows, reduce the (n -1)
entries, except (1,1) in the first column, to zero. Expand det(A) by its first column.
Repeat this process for the minor A
11
or continue the following steps.

Step 3:
Repeat step 1 and step 2 with the last remaining rows concentrating on the second
column.

Step 4:
Repeat step 1, step 2 and step 3 with the remaining (n -2) rows, (n -3) rows and so
on, until a triangular matrix is obtained.

9-47
Step 5:
Multiply all the diagonal entries of the triangular matrix and then multiply it by its sign
to get det(A).

Remark 9.4.3:
For (3x3)-matrices A there is a simple rule called Sarrus Rule to compute the
number det(A).

(9.4.7)

. 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 a
det(A)

12 21 33 11 23 32 13 22 31 32 21 13 31 23 12 33 22 11
32
22
12
31
21
11
33 32 31
23 22 21
13 12 11
+ +

Example 9.4.4:

Suppose
1
1
1
]
1
5 6 3
2 1 2
3 2 0
A , then we get by Sarrus Rule:

det(A) = 53 20 0 9 36 12 0
6
1
2
3
2
0
5 6 3
2 1 2
3 2 0
+
.

Definition 9.4.5:
Let A be an (nxn)-matrix and C
ij
be the cofactor of a
ij
in A. We set

B =
1
1
1
1
1
1
]
1
nn n2 n1
2n 22 21
1n 12 11
C . . . C C
. . .
. . .
C . . . C C
C . . . C C
.

Then the adjoint of A, written adj(A), is the transpose of B. Thus
+ + +
- - -

9-48
adj(A) = B
T
=
1
1
1
1
1
1
]
1
nn 2n 1n
n2 22 12
n1 21 11
C . . . C C
. . .
. . .
C . . . C C
C . . . C C
.

Proposition 9.4.1:
Let A be a (nn)-matrix. Then applies:

(i) A is non-singular if and only if det(A) 0

(ii) A is singular if and only if det(A) = 0

(iii) Aadj(A) = det(A)I
n

(iv) ( ) ( )
T
A det A det

(v) If det(A) 0, then det(A
-1
) =
( ) A det
1
.

Proof:
We prove only (iv). By the product of determinants, we have

det(A
-1
A) = det (A
-1
)det (A).

But det(A
-1
A) = det (I ) = 1. Hence det(A
-1
)det(A) = 1. Therefore

det(A
-1
) 0 and det(A) 0.

So det(A
-1
) =
( ) A det
1
.

Corollary 9.4.1:
(i) Let A be an (nn)-matrix.Then A is invertible if and only if det(A) 0.
(ii) If det(A) 0, then A
-1
=
( ) A det
1
adj(A).

Proof:
If A is invertible, then A is non-singular and therefore det(A) 0.
Next if det(A) 0, then

9-49
A
( ) A det
1
adj(A) =
( ) A det
1
A adj(A) =
( )
( ) A det
A det
I = I .

Hence A is invertible and

A
-1
=
( ) A det
1
adj(A).

Example 9.4.5:
Find, by adjoint method, the inverse of

A =
1
1
1
]
1
7 2 5
8 1 2
5 4 3
.

Solution:
Here, the cofactors are

C
11
= (-1)
1+1

7 2
8 1
= 9, C
12
= (-1)
1+2

7 5
8 2
= 26,

C
13
= (-1)
1+3

2 5
1 2
= 1, C
21
= (-1)
2+1

7 2
5 4
= -38,

C
22
= (-1)
2+2

7 5
5 3
= -4, C
23
= (-1)
2+3

2 5
4 3
= 26,

C
31
= (-1)
3+1

8 1
5 4
= 37, C
32
= (-1)
3+2

8 2
5 3
= -14,

C
33
= (-1)
3+3

1 2
4 3
= -11.

Now

det(A) = a
11
C
11
+ a
12
C
12
+ a
13
C
13
= 3 9 + 4 26 + 5 1 = 136.

Therefore

9-50
A
-1
=
( ) A det
1
adj(A) =
136
1
1
1
1
]
1
11 26 1
14 4 26
37 38 9
.

In the following we discuss a classical method of solving a system of n equations in
n unknowns using determinants. Consider the system of equations

(9.4.8)
'
+ + +
+ + +
+ + +
n n nn 2 2 n 1 1 n
2 n n 2 2 22 1 21
1 n n 1 2 12 1 11
b x a ... x a x a
.
.
.
b x a ... x a x a
b x a ... x a x a

which can be written as Ax = b, where

A = [
ij
a ] is a (nn)-matrix, x = [x
1,
x
2,
...
,
x
n
]
T
and b = [b
1,
b
2,,
b
n
]
T
.

Suppose that det(A) 0. Then (9.4.8) has the unique solution x = A
-1
b where the in-
verse may be calculated by any of the methods discussed earlier. However, in the
following we describe a method of finding a solution with row (column) reduction.

(9.4.9) Cramers Rule:
Let [ ]
n 1
a ,..., a A R
nxn
be a (nn)-matrix, where [ ]
T
nj j 1
a ,..., a
j
a is the j-th column
of A, with det(A) 0. Then the unique solution of the linear system Ax = b is given by

(9.4.9)
( ) ( )
( )
n 1 j 1 j 1 j j
,..., , , ,..., det
A det
1
D
A det
1
x a a b a a
+

where D
j
= ( )
n 1 j 1 j 1
,..., , , ,..., det a a b a a
+
, j = 1,2,,n, is the determinant obtained from
A by replacing in A the jth column by the column with entries b = [b
1,
b
2,,
b
n
]
T
.

Remark 9.4.4:
If the system (9.4.8) is homogeneous and det(A) 0, then it has only the trivial solu-
tion x
1
= x
2
= = x
n
= 0. If det(A) = 0, the homogeneous system also has non-trivial
solutions.

9-51
Example 9.4.6:
Solve, by Cramers Rule, the system of equations:

1 3x x 2x
3 2 1
+
2 x 2x x
3 2 1
+
3 2x 2x 3x
3 2 1
+ + .

Solution:
Here

A =
1
1
1
]
1
2 2 3
1 2 1
3 1 2

and therefore det(A) =
2 2 3
1 2 1
3 1 2
= 5,

Furthermore:

D
1
=
2 2 3
1 2 2
3 1 1
= 7, D
2
=
2 3 3
1 2 1
3 1 2
= 0 and D
3
=
3 2 3
2 2 1
1 1 2
= -3.

Hence x
1
=
( ) A det
D
1
=
5
7
, x
2
=
( ) A det
D
2
= 0 and x
3
=
( ) A det
D
3
=
5
3
.


9-52
9.5 Eigenvalues and Eigenvectors

From the standpoint of engineering applications, eigenvalue problems are among
the most important problems in connection with matrices. Let A = [a
ij
] be a given
(nn)-matrix and consider the vector equation

(9.5.1) A x = x.

Here, x = [x
1,
x
2,
...
,
x
n
]
T
is an unknown vector and is an unknown scalar and we want
to determine both. Clearly, the zero vector x = 0 is a solution of (9.5.1) for any value
of . This is of no practical interest.

Definition 9.5.1:
(i) A value of C for which Ax = x has a non-trivial solution x 0 is called an ei-
genvalue or characteristic value of the matrix A.

(ii) The corresponding solutions x 0 of Ax = x are called eigenvectors or charac-
teristic vectors of A corresponding to that eigenvalue .

(iii) The set of eigenvalues is called the spectrum of the matrix A.

(iv) The largest of the absolute values of the eigenvalues of A is called spectral ra-
dius of A.

(v) The set of all eigenvectors corresponding to an eigenvalue of A, together with 0,
forms a vector space, called the eigenspace of A corresponding to this eigenvalue.

The problem of determining the eigenvalues and eigenvectors of a matrix is called
an eigenvalue problem. Problems of this type occur in connection with physical,
technical, geometrical, and other application.

The equation Ax = x written in components is

(9.5.2)
'
+ + +
+ + +
+ + +
n n nn 2 2 n 1 1 n
2 n n 2 2 22 1 21
1 n n 1 2 12 1 11
x x a ... x a x a
.
.
.
x x a ... x a x a
x x a ... x a x a

Transferring the terms on the right side to the left side, we have

9-53
(9.5.3)
( )
( )
( )
'
+ + +
+ + +
+ + +
0 x a ... x a x a
.
.
.
0 x a ... x a x a
0 x a ... x a x a
n nn 2 2 n 1 1 n
n n 2 2 22 1 21
n n 1 2 12 1 11
.

In matrix notation:

(1.5.4) (A I )x = 0.

By Cramers Rule, this homogeneous linear system of equations (9.5.4) has a non-
trivial solution if and only if the corresponding determinant det(A I ) of the coeffi-
cients is zero; i.e.:

(9.5.5) det(AI ) =
( )
( )
( ) a ... ... a a
. . .
. . .
a ... ... a a
a ... ... a a
nn 2 n 1 n
n 2 22 21
n 1 12 11
= 0.

Definition 9.5.2:
The determinant det(AI ) is called the characteristic determinant. Equation
(9.5.4) is called the characteristic equation of the matrix A. By developing
det(AI ) we obtain a polynomial of n-th degree in . This is called the characteris-
tic polynomial of A.

This proves the following important theorem.

Theorem 9.5.1:
The eigenvalues of a square matrix A are the roots of the corresponding characteris-
tic equation (9.5.4). Hence a (nxn)-matrix has at least one eigenvalue and at most n
numerically different eigenvalues.

Remark 9.5.1:
The eigenvalues must be determined first. Once these are known, corresponding
eigenvectors are obtained from the system (9.5.3), for instance, by the Gauss elimi-
nation, where is the eigenvalue for which an eigenvector is wanted.

9-54
Example 9.5.1:
Find the eigenvalues and the corresponding eigenvectors of the matrix

A =
1
]
1
2 2
3 1
.

Solution:
The eigenvalues of A are given by its characteristic equation:

det(AI
2
) =

2 2
3 1
= 0

(1 )(2 ) 6 = 0

2
- 3 - 4 = 0

( 4)( + 1) = 0.

Hence the eigenvalues of A are
1
= -1 and
2
= 4.

Let the corresponding eigenvectors be v
1
=
1
]
1
2
1
x
x
and v
2
=
1
]
1
2
1
y
y
.
Therefore

(A
1
I
2
) v
1
= 0

1
]
1
+
+
1 2 2
3 1 1

1
1
]
1
2
x
1
x
=
1
]
1
0
0

'
+
+
0 3x 2x
0 3x 2x
2 1
2 1
.

The eigenvectors of the eigenvalue
1
= -1 are v
1
=
1
1
]
1
3
2 for any scalar . If we
set = 3, then an eigenvector is v
1
=
1
]
1
2
3
.

Similarly (A
2
I
2
)v
2
= 0 becomes

9-55

1
]
1
4 2 2
3 4 1
1
]
1
2
1
y
y
=
1
]
1
0
0

'

+
0 2y 2y
0 3y 3y
2 1
2 1

A solution is easy in this case and can be expressed as y
1
= y
2
= for any
non-zero scalar . Therefore the eigenvectors of the eigenvalue
2
= 4 are
v
2
=
1
]
1
. If we put = 1, then an eigenvector is v

2
=
1
]
1
1
1
.

Remark 9.5.2:
(i) The eigenvalues of a Hermitian matrix (and thus of a symmetric matrix) are real.

(ii) The eigenvalues of a skew-Hermitian matrix (and thus of a skew-symmetric ma-
trix) are pure imaginary or zero.

(iii) The eigenvalues of a unitary matrix (and thus of an orthogonal matrix) have abso-
lute value 1.

Proposition 9.5.1:
If v is an eigenvector of a matrix A corresponding to an eigenvalue , so is rv with an
r 0.

Proof:
Av = v implies r(Av) = A(rv) = r(v) = (rv).

Example 9.5.1:
Tank T
1
in Figure 9.5.1 contains initially 100 gal of
pure water. Tank T
2
contains initially 100 gal of
water in which 150 lb of fertilizer are dissolved.
Liquid circulates through the tanks at a constant
rate of 2 gal/min, and the mixture is kept uniform by
stirring. Find the amounts of fertilizer y
1
(t) and y
2
(t)
in T
1
and T
2
, respectively, where t is the time.

T
2

2 gal/min
2 gal/min
T
1

Fig. 9.5.1 Fertilizer content in Tanks

9-56
Solution:
(i) The time rate of change ( ) t y
1
of y
1
(t) (amount of fertilizer in T
1
) equals inflow
minus outflow. Similar for tank T
2
. Thus we get:

1
y = infolw/min outflow/min =
1 2
y
100
2
y
100
2
(Tank T
1
)

2
y = infolw/min outflow/min =
2 1
y
100
2
y
100
2
(Tank T
2
).

The mathematical model of the mixture problem is the following system of first-order
differential equations:

(9.5.6)
'

+
2 1 2
2 1 1
y 0.02 y 0.02 y
y 0.02 y 0.02 y

As a vector equation with vector
1
]
1
2
1
y
y
y and matrix A this becomes

(9.5.7) Ay y where
1
]
1
02 0 02 0
02 0 02 0
. .
. .
A

(ii) We try an exponential function of t:

(9.5.8)
t
e x y
t
e x y

With (9.5.7) we get:

(9.5.9)
t
e x Ay
t t
e e x Ax

Dividing by
t
e we obtain

(9.5.10) x Ax .

We need nontrivial solutions and hence we have to look for eigenvalues and
eigenvectors of the matrix A. The eigenvalues are the solutions of the characteristic
equation

(9.5.11) ( ) ( ) 0 0.04
0.02 0.02
0.02 0.02
det +

I A .

9-57
We get 0
1
and 0.04
2
and therefore

(9.5.12)
'

+
0 x 0.02 x 0.02
0 x 0.02 x 0.02
2 1
2 1
and
( )
( )
'
+ +
+ +
0 x 0.04 0.02 x 0.02
0 x 0.02 x 0.04 0.02
2 1
2 1
.

As easily seen we need only the first equations of the systems in (9.5.12)

(9.5.13) 0 x 0.02 x 0.02
2 1
+ and ( ) 0 x 0.02 x 0.04 0.02
2 1
+ + .

Hence
2 1
x x and
2 1
x x and we can take the eigenvectors

(9.5.14)
( )
1
]
1
1
1
1
x and
( )
1
]
1
1
1
2
x .

From (9.5.8) and the superposition principle we thus obtain a solution

(9.5.15) ( )
( ) ( ) t . t t
e c c e c e c t
04 0
2 1 2 1
1
1
1
1
2 1

1
]
1
+
1
]
1
+
2 1
x x y

where c
1
and c
2
are arbitrary constants. Later this will be called a general solution.

(iii) The initial conditions are y
1
(0) = 0 (no fertilizer in tank T
1
) and y
2
(0) = 150. From
this and (9.5.15) we get

(9.5.16) ( )
1
]
1
1
]
1
1
]
1
+
1
]
1
150
0
c c
c c
1
1
c
1
1
c 0
2 1
2 1
2 1
y

In components 0 c c
2 1
+ and 150 c c
2 1
. The
solution is 75 c 75, c
2 1
. This gives the
answer of our problem

(9.5.17) ( )
t 0.04
e
1
1
75
1
1
75 t

1
]
1
1
]
1
y .

In components:

( )
t 0.04
1
e 75 75 t y

(Tank T
1
, lower curve in Fig. 9.5.2)

( )
t 0.04
e 75 75 t y

+
2
(Tank T
2
, upper curve in Fig. 9.5.2)

Fig. 9.5.2 Solution of the mixing problem

Ise Vmath109

Uploaded by

Copyright:

Available Formats

Ise Vmath109

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ise Vmath109

Uploaded by

Copyright:

Available Formats

Mathematics for Engineers Part I (ISE) Version 1.

. If we put = 1, then an eigenvector is v

You might also like