Chapter 1

Chapter 1
Linear Systems of
Algebraic Equations
and Matrices
1.1 Historical Remarks on Algebraic Equations
Muhammad Ibn Musa al-Khwarizmi to Whom we Owe the Word Algebra

His name indicates he might have came from Khwarizm (Khiva) in the Khorasan
province of the Abbasid empire, now Xorazm Province of Uzbekistan.
His name is often given as either Abu or Abu Jafar Abd Allah Muhammad ibn Musa
al-Khwarizmi. The pious preface to al-Khwarizmi's Algebra shows that he was a
keen Muslim. Al-Khwarizmi accomplished most of his work in the period between
813 and 833. After the Islamic conquest of Persia, Baghdad became the centre of
scientific studies and trade, and many merchants and scientists, from as far as China
and India traveled to this city, apparently so did Al-Khwarizmi. He worked in
Baghdad as a scholar at the House of Wisdom established by Caliph al-Mamun,
where he studied the sciences and mathematics, which included the translation of
Greek and Sanskrit scientific manuscripts.
His major contributions to mathematics, astronomy, astrology, geography and
cartography provided foundations for later and even more widespread innovation in
algebra, trigonometry, and his other areas of interest. His systematic and logical
approach to solving linear and quadratic equations gave shape to the discipline of
algebra, a word that is derived from the name of his 830 book on the subject, al-
Kitab al-mukhtasar fi hisab al-jabr wa'l-muqabala or: "The Compendious Book on
Calculation by Completion and Balancing". The book was first translated into Latin
in the twelfth century.
His book On the Calculation with Hindu Numerals written about 825, was
principally responsible for the diffusion of the Indian system of numeration in the
4 Linear Systems of Algebraic Equations and Matrices
Middle-East and then Europe. This book also translated into Latin in the twelfth
century, as Algoritmi de numero Indorum. From the name of the author, rendered in
Latin as algoritmi, originated the term algorithm.
Some of his contributions were based on earlier Persian and Babylonian Astronomy,
Indian numbers, and Greek sources.
Al-Khwarizmi systematized and corrected Ptolemy's data in geography as regards to
Africa and the Middle east. Another major book was his Kitab surat al-Ard , "The
Image of the Earth"; translated as Geography, which presented the coordinates of
localities in the known world based, ultimately, on those in the Geography of
Ptolemy but with improved values for the length of the Mediterranean Sea and the
location of cities in Asia and Africa.
He also assisted in the construction of a world map for the caliph al-Ma'mun and
participated in a project to determine the circumference of the Earth, supervising the
work of 70 geographers to create the map of the then "known world".
When his work was copied and transferred to Europe through Latin translations, it
had a profound impact on the advancement of basic mathematics in Europe. He also
wrote on mechanical devices like the astrolabe and sundial.
al-Kitāb al-mukhtasar fī hisāb al-jabr wa-l-muqabala is considered to have defined
the branch algebra of mathematics. The word algebra is derived from the name of
one of the basic operations with equations (al-jabr) described in this book. The book
was translated in Latin as Liber algebrae et almucabala by Robert of Chester, in
Segovia, in the year 1145.
Al-Khwarizmi's method of solving linear and quadratic equations worked by first
reducing the equation to one of six standard forms below, where b and c are positive
integers:
• squares equal roots (ax2 = bx)
• squares equal number (ax2 = c)
• roots equal number (bx = c)
• squares and roots equal number (ax2 + bx = c)
• squares and number equal roots (ax2 + c = bx)
• roots and number equal squares (bx + c = ax2)
by dividing out the coefficient of the square and using the two operations al-jabr
“restoring” or “completion”, and al-muqabala, "balancing". Al-jabr is the process of
removing negative units, roots and squares from the equation by adding the same
quantity to each side. For example, x2 = 40x - 4x2 is reduced to 5x2 = 40x. Al-
muqabala is the process of bringing quantities of the same type to the same side of
the equation. For example, x2+14 = x+5 is reduced to x2+9 = x.
1.2 Introduction to the Systems of Linear Equations
During the first years of your primary school education, you might be encountered
with the problems as in the following example:
EXAMPLE 1.1 For five kilo of rice I paid 20 liras. What was the price of the rice?
We can express this question with an algebraic equation of the form
5 20, ?
We know that, when we divide both sides of an equation by the same nonzero
number, the equality still holds:
5 /5 20/5, 4.
Generalizing the above question, we may search for the solution of the equation
ax  b  0 (1.1)
where a, b are real constants and x is the unknown. To find the value of the
unknown x , you add  b to both sides knowing that adding to the both sides of
(1.1) the same number, the equality does not disturbed:
ax  b  (b)  b (1.2)
and get
ax  b . (1.3)
Then you multiply both sides of (1.3) with 1 / a , knowing that multiplying both sides
with the same number, the equality does not disturbed:
1 1
.a.x  b.
a a
and simplify to
b
x .
a
Real life problems mostly deal with more than one unknowns.
EXAMPLE 1.2 In our farm we have hens and rabbits. When we count the heads, we get the number
18. Counting the legs, we get 64. How many of hens and rabbits we have?
Let the number of hens and rabbits be x and y respectively. The first information
about the heads yields us the equation
x  y  18 (1.4)
while the second information about the legs lead to
4 x  2 y  64. (1.5)
To find x and y , we solve y from (1.4)
y  18  x (1.6)
and substitute in (1.5):
4 x  218  x   64. (1.7)
Simplifying (1.7) one gets
4  2x  64  36, 2 x  28, x  14. (1.8)
Substitution of the value of x in (1.8) into (1.4), the value of y is found to be
y  18  14  4
It is understood that we have 14 hens and only 4 rabbits.
1.2.1 The method of Gauss1-Jordan2 elimination

In the cases of more than two unknowns, it is more convenient to use the method of
elimination. Assume we have three unknowns x, y , and z , and three equations to
specify these three unknowns:
a11 x  a12 y  a13 z  b1
a 21 x  a 22 y  a 23 z  b2 (1.9)
a 31 x  a 32 y  a33 z  b3 .
1
Johann Carl Friedrich Gauss, Born: 30 April 1777 in Brunswick, Duchy of Brunswick (now Germany),
died: 23 Feb 1855 in Göttingen, Hanover (now Germany), worked in a wide variety of fields in both mathematics and
physics including number theory, analysis, differential geometry, geodesy, magnetism, astronomy and optics. His work
has had an immense influence in many areas.
2
Camille Jordan (1838-1922), French mathematician.
1.2 Introduction to the Systems of Linear Equations 5
To find the value of unknowns, let us multiply both sides of the three equations by
the inverse of the coefficients of x in each equation, assuming that this coefficient
is not zero:
a12 a b
x y  13 z  1
a11 a11 a11
a 22 a b
x y  23 z  2
a 21 a 21 a 21
a32 a b
x y  33 z  3 .
a31 a31 a31
Now let us multiply the first equation by  1 and add on the second and third
equations knowing that the solution set of the three equations is not change under
these operations:
a12 a b
x y  13 z  1
a11 a11 a11
 a 22 a12  a a  b b
   y   23  13  z  2  1 (1.10)
 a 21 a11   a 21 a11  a 21 a11
 a32 a12  a a  b b
   y   33  13  z  3  1 .
 a31 a11   a 31 a11  a31 a11
In (1.10) the last two equations contain only two unknowns y and z . Applying the
same procedure again we have only one equation with one unknown z from which
this unknown is solved. Substituting this value of z in any of the last two equations
in (1.10), we can solve y . The values of y and z are substituted in the first
equation in (1.10) to find x .
EXAMPLE 1.3 In a confection factory, three types of shirts are produced. First type of shirts needs 4
buttons, 80 centimeters of textile and 20 centimeters of ribbon. For the second kind
of shirts this are 6, 90, and 30. For the third kind the same values are 8, 120, and 50.
If in stocks there are 2060 buttons, 325 meters of textile and 180 meters of ribbons,
what is a possible production plan?
Let the respective numbers of shirts of each kind be x, y , and z . The three
equations in the below relate these three unknowns:
4 x  6 y  8 z  2060
80 x  90 y  120 z  32500
20 x  30 y  50 z  12100.
After simplifications one has
8 x  12 y  16 z  4120
8 x  9 y  12 z  3250 (1.11)
8 x  12 y  20 z  4840.
multiplying the first equation of (1.11) by  1 and adding on the second and third
equations one has
8 x  12 y  16 z  4120
 3 y  4 z  870 (1.12)
4 z  720.
From the last equation in (1.12), the unknown z is solved as z  720 / 4  180 .
Substituting this value of z in the second equation in (1.12), y is found as
 3 y  720  870,  3 y  150, y  50.
The values of y and z are substituted in the simplified version of the first equation
in (1.12) to find x :
x  3 y / 2  2 z  515, x  75  360  515, x  515  435  80.
Hence in an optimum production plan, the number of each kind of shirts is 80, 50,
and 180.
1.2.2 The method of echelon forms

The elimination method used in the previous subsection can made more systematic
making use of matrices.
DEFINITION 1.1 An array of numbers with m rows and n columns is called an m  n matrix.
 a11 a12 ... a1n 

a a 22 ... a 2 n 
 21 (1.13)
 ... ... ... ... 
 
a m1 am2 ... a mn 
The coefficients of a linear system of equations

a11 x  a12 y  a13 z  b1
a 21 x  a 22 y  a 23 z  b2 (1.14)
a31 x  a32 y  a33 z  b3 .
can be stored in a matrix
 a11 a12 a13 

A  a 21 a 22 a 23  (1.15)
a31 a32 a33 
which is called the coefficient matrix of the linear system of equations in (1.14).
We may add the numbers in the right hand sides of equations in (1.14) as a last
column to get the so called augmented matrix for the system (1.14):
 a11 a12 a13 b1 

a a 22 a 23 b2  (1.16)
 21
a31 a32 a33 b3 
It is easy to show that the following operations which we use during the application
of the elimination method, do not alter the solution set of the linear system of
algebraic equations in (1.14):
1. Multiplying any of the equations throughout by a nonzero number,
2. Multiplying any of the equations throughout by a nonzero number, and adding
on another equation.
3. Replacing equations.
The above operations, when observed through augmented matrix, are called
elementary row operations:
1. Multiplying any of the rows by a nonzero number,
2. Multiplying any of the rows by a nonzero number, and adding on another row.
3. Replacing rows
The resulting augmented matrix represents a new linear algebraic system which has
exactly the same solution set as the original one.
The resulting system can easily be solved if it is in row echelon form.
DEFINITION 1.2 An augmented matrix in row echelon form if

a. All zero rows, if there are any, appear at the bottom of the matrix,
b. The first nonzero entry on the left of a nonzero row is one,
c. The leading one appears to the right and below any leading ones in preceding
rows,
In addition to the above properties, if a matrix satisfy
d. If a column contains a leading one, then all other entries in this column are zero,
then it is called in reduced row echelon form.
EXAMPLE 1.4 Solve the system of algebraic equations by the use of Gaussian elimination;
4 x  6 y  8 z  2060
80 x  90 y  120 z  32500
20 x  30 y  50 z  12100.
The corresponding augmented matrix is
4 6 8 2060 
80 90 120 32500
 
20 30 50 12100 
Multiplying rows by appropriate numbers we simplify the augmented matrix into
1 3 / 2 2 515 
8 9 12 3250
 
2 3 5 1210 
Now let us perform the following elementary row operations: multiply first row by
 8 , and add on the second row  8 R1  R 2  R 2 , multiply first row by  2 , and
add on the third row  2 R1  R3  R3 ,
1 3 / 2 2 515  1 3 / 2 2 515
0  3  4  870 1 0 1 4 / 3 290 .
   R2  R2  
3
0 0 1 180  0 0 1 180 
The last matrix is in echelon form. If we interpret it as a system of linear equations
we get
x  3 y / 2  2 z  515
y  4 z / 3  290
z  180.
From the last equation we get z  180. Substituting this value of z in the second
equation y is found as y  50. The values of y and z are substituted in the first
equation to find x :
x  3 y / 2  2 z  515, x  75  360  515, x  515  435  80.
If we continue to achieve the reduced row echelon form, we have
1 3 / 2 2 515 1 3 / 2 0 155
0 1 4 / 3 290  4 R3 / 3  R 2  R 2,  
   2 R3  R1  R1, 0 1 0 50 
0 0 1 180  0 0 1 180
and finally the operation  3R 2 / 2  R1  R1 leads to the reduced row echelon

matrix
1 0 0 80 
0 1 0 50 
 
0 0 1 180
whose last column exposes the solution. Indeed interpreted as a system of linear
equation the above augmented matrix yields
x  80, y  50, z  180.
DEFINITION 1.3 In a row echelon form, variables corresponding to leading ones are called leading
variables. The others are called the free variables.
THEOREM 1.1 Let e, and f be the number of leading and free variables in the row echelon matrix
of a linear algebraic system. Depending on e, the number r of the nonzero rows, and
the number n of unknowns, one of the following three cases does occur:
1. r  e  n , the linear system of algebraic equations has a unique set of solutions.
2. r  e  n , the linear system of algebraic equations has infinitely many sets of
solutions.
3. r  e , then the linear system of algebraic equations is contradictory and has no
solutions.
EXAMPLE 1.5 Find all solutions of the below system of algebraic equations by the use of Gaussian
elimination;
x  y  2 z  1
x  2 y  z  5
3 x  y  z  3.
The corresponding augmented matrix under two elementary row operations becomes
1 1 2  1 1 1 2  1
1  2 1  5  R1  R 2  R 2, 0  3  1  4 
   3R1  R3  R3  
3 1 1 3  0  2  5 6 
Multiplying second row by –1/3, and applying another elementary row operation
one has
1 1 2 1  1 1 2 1 
0 1 1 / 3 4 / 3 2 R 2  R3  R3 0 1 1 / 3 4 / 3 
  
0  2  5 6  0 0  13 / 3 26 / 3
Multiplying third row by –3/13, and applying other two elementary row operations
we get
1 1 2 1  1 1 0 3 
0 1 1 / 3 4 / 3  R3 / 3  R 2  R 2, 0 1 0 2 
   2 R3  R1  R1  
0 0 1  2  0 0 1  2
Multiplying second row by –1, and adding to the first row we get the reduced
echelon form of the augmented matrix:
1 0 0 1 
0 1 0 2  .
 
0 0 1  2
We observe that r  m  n  3 , according to Theorem 1.1, the linear system of

algebraic equations of this example has a unique set of solutions
{x, y , z}  {1, 2,  2} .
EXAMPLE 1.6 In the following linear system, find all values of the parameter a such that the
system has
a) No solution,
b) A unique solution,
c) Infinitely many solutions;
x yz 2
x  2y  z  3
x  y  a 2  5 z  a.
The corresponding augmented matrix under two elementary row operations becomes
1 1 1 2 1 1 1 2 
1 2  R1  R 2  R 2,
 1 3 0 1
 2 1 
 R1  R3  R3
1 1 a 2  5 a  0 0 a 2  4 a  2
a) For a  2 the echelon form is
1 1  1 2 
0 1 2 1  .

0 0 0  4
When interpreted as system of algebraic equations, the last row is interpreted as the
equation
0 z  4
which cannot be satisfied by a definite number z . For a  2 , r  3  e  2 , and
according to Theorem 1.1, the linear system of algebraic equations in this example
has no solutions, the system is contradictory.
b) For a   2 , say a  1, after the application of two elementary row
operations the reduced echelon form is
1 1  1 2   2 R3  R 2  R 2, 1 0 0 2 
0 1 2 1  R3  R1  R1 , 0 1 0 1 / 3 .
  
0 0 1 1 / 3  R 2  R1  R1 0 0 1 1 / 3
We observe that r  e  n  3 , according to Theorem 1.1, the linear system of

algebraic equations of this example for a  1, has a unique set of solutions
{x, y , z}  {2,1 / 3, 1 / 3} .
c) For a  2 after the application of one elementary row operations the reduced
echelon form is
1 1  1 2 1 0  3 1
0 1 2 1   R 2  R1  R1 0 1 2 1  .
   
0 0 0 0 0 0 0 0
We observe that r  e  2  n  3 , according to Theorem 1.1, the linear system of

algebraic equations of this example for a  2 has infinitely many sets of solutions.
The variables x, y which correspond to the leading ones are leading variables, and
z is a free variable. x, y are expressed in terms of z:
x  3z  1, x  3z  1
y  2 z  1, y  1  2z
z is free
the linear system of algebraic equations of this example for a  2 has infinitely
many sets of solutions of the form:
{x, y , z}  {3 z  1, 1  2 z , z}
as an example, for z  3 one has { x, y , z}  {10,  5, 3} .
EXAMPLE 1.7 Find all solutions of the below system of algebraic equations by the use of
Gaussian-Jordan reduction;
2 x  y  z  2w  1
3 x  2 y  z  6 w  2
x  y  z  w  1
5 x  y  2 z  8 w  3.
The corresponding augmented matrix interchanging first and third rows becomes
2 1 1 2 1  1 1  1  1  1
3  2 1  6  2 3  2 1  6  2
 R1  R3 
1 1  1  1  1 2 1 1 2 1 
   
5  1 2 8 3  5  1 2 8 3 
Applying the three elementary row operations below we get

1 1  1  1  1 1 1  1  1  1
3  2 1 3R1  R 2  R 2,
  6  2 0  5 4  3 1 
 
 2 R1  R3  R3
2 1 1 2 1  0  1 3 0 3
   5 R1  R 4  R 4  
5  1 2 8 3  0  6 7  3 8 
Multiplying third row by –1, interchanging second and third rows and applying the
two elementary row operations below we get
1 1  1  1  1 1 1 1 1 1 
0 1  3 0  3 5 R 2  R3  R3, 0 1 3 0  3 
  
0  5 4  3 1  6R2  R4  R4 0 0  11  3  14
   
0  6 7  3 8  0 0  11  3  10
Multiplying second and third rows by –1/11, and applying the elementary row
operation below we get the echelon form:
1 1 1 1 1 
0 1 3 0  3 
  R3  R 4  R 4
0 0 1 3 / 11 14 / 11
 
0 0 1 3 / 11 10 / 11
1 1 1 1 1 
0 1 3 0  3 
 .
0 0 1 3 / 11 14 / 11 
 
0 0 0 0  4 / 11
We observe that r  4  m  3 , according to Theorem 1.1, the linear system of
algebraic equations in this example has no solutions, the system is contradictory.
1.2.3 Homogeneous systems

An algebraic system of m equations for n unknowns
a11 x1  a12 x 2  ...  a1n x n  b1
a 21 x1  a 22 x 2  ...  a 2 n x n  b2
(1.16)
.........
a m1 x1  a m 2 x 2  ...  a mn x n  bm
is called homogeneous if all right hand sides are zero: b1  b2  .......  bm  0 .
Since the last column is zero in the augmented matrix, for homogeneous equations,
Theorem 1.1 is modified as follows
THEOREM 1.2 Let e, and f be the number of leading and free variables in the row echelon matrix of
a homogeneous linear algebraic system. The number of leading variables is always
equal to the number of the nonzero rows, e  r . Depending on the number n of

unknowns, e and r, one of the following two cases do occur:
1. r  e  n , the linear system of algebraic equations has only the trivial solution
x1  x 2  .......  x n  0 .
2. r  e  n , the linear system of algebraic equations has infinitely many sets of
solutions.
Gaussian-Jordan elimination;
2x  9 y  z  0
2x  6 y  7z  0
x  3 y  6 z  0.
The corresponding augmented matrix under three elementary row operations
becomes
 2 9 1 0 R1  R3 1 3  6 0
 2 6 7 0  2 R1  R 2  R 2, 0 0 19 0
   
1 3  6 0  2 R1  R3  R3 0 3 13 0
three other elementary row operations give
R 2 / 19  R3 1 3  6 0 1 0 0 0
 3R 2  R1  R1
13R3  R 2  R 2, 0 1 0 0 0 1 0 0 
  6 R3  R1  R1.  
R2 / 3  R2 0 0 1 0 0 0 1 0
We observe that r  e  n  3 , according to Theorem 1.2, the linear system of

algebraic equations of this example has only the trivial solution x  y  z  0 .
Gaussian-Jordan elimination;
3x  9 y  z  0
2x  6 y  7z  0
x  3 y  6 z  0.
The corresponding augmented matrix under three elementary row operations
becomes
 3 9 1 0 R1  R3 1 3  6 0
 2 6 7 0  2 R1  R 2  R 2, 0 0 19 0
   
1 3  6 0  3R1  R3  R3. 0 0 19 0
three other elementary row operations give the reduced echelon form:
 R 2  R3  R3, 1 3 0 0
R 2 / 19  R 2 ., 0 0 1 0 
 
6 R 2  R1  R1. 0 0 0 0

algebraic equations of this example has infinitely many sets of solutions.
The variables x, z which correspond to the leading ones are leading variables, and y
is the free variable. x, z are expressed in terms of y:
x  3 y  0, x  3 y
y is free
z  0.
the linear system of algebraic equations of this example has infinitely many sets of
solutions of the form:
{x, y , z}  {3 y , y , 0}
as an example, for y  1 one has { x, y , z}  {3, 1, 0} .
EXAMPLE 1.10 Find all solutions of the below system of algebraic equations by the use of Gaussian-
Jordan elimination;
2 x  2 y  4 z  2w  0
x  y  4 z  3w  0
2 x  7 y  19 z  3w  0.
We observe that m  n , the number of equations is less than the number of
equations. The number of leading variables is always equal to the number of the
nonzero rows, and they cannot be larger than the number of equations n. Hence
r  e  m  3  n  4 , and according to Theorem 1.2, the linear system of
To find the type of solutions let us consider the corresponding augmented matrix:
2 2 4 2 0
1  1  4 3 0 
 
2 7 19  3 0
Under three elementary row operations becomes
R1 / 2  R1 1 1 2 1 0
 R1  R 2  R 2, 0  2  6 2 0 
 
 2 R1  R3  R3. 0 5 15  5 0
three other elementary row operations give the reduced echelon form:
 R 2  R3  R3,
1 0  1 2 0
 R2 / 2  R2 , 0 1 3  1 0 
 R 2  R1  R1,  
0 0 0 0 0
 5 R 2  R 3  R 3.
The variables x, y which correspond to the leading ones are leading variables, and
z, w are free variables x, y are expressed in terms of z, w :
x  z  2w  0, x  z  2w
y  3z  w  0, y  3z  w
z, w are free.
the linear system of algebraic equations of this example has infinitely many sets of
solutions of the form:
{x, y, z , w}  {z  2 w,  3 z  w, z , w}
as an example, for z  1, w  2 one has { x, y , z , w}  {3, 1, 1, 2} .
EXAMPLE 1.11 Gaussian Elimination

Solve
by Gaussian elimination and back-substitution.
Solution :
We convert the augmented matrix
to the row-echelon form
The system corresponding to this matrix is
Solving for the leading variables yields
Substituting the bottom equation into those above yields
and substituting the second equation into the top yields x=1, y=2, z=3 .
EXERCISES 1.2
1. Solve each of the following systems by Gauss–Jordan elimination.
(a)
(b)
(c)
(d)
(a)
(b)
(c)
(d)
(a)
(b)
(c)
(d)
x1 + 4x2 − x4 + 7x6 − 9x7 = 3
2x1 + 8x2 − x3 + 3x4 + 9x5 − 13x6 + 7x7 = 9
2x3 − 3x4 − 4x5 + 12x6 − 8x7 = 1
−x1 − 4x2 + 2x3 + 4x4 + 8x5 − 31x6 + 37x7 = 4
(e)
x1 − x2 − 2x3 + x4 + 11x5 = 13
x1 − x2 + x3 + x4 + 5x5 = 16
2x1 − 2x2 + x4 + 10x5 = 21
2x1 − 2x2 − x3 + 3x4 + 20x5 = 38
2x1 − 2x2 + x3 + x4 + 8x5 = 22
5. Without using pencil and paper, determine which of the following

homogeneous systems have nontrivial solutions.
(a)
(b)
2x1 − 3x2 + x3 − 6x4 = 0
4x1 + x2 + 2x3 + 9x4 = 0
3x1 + x2 + x3 + 8x4 = 0
(c)
(d)
(e)
6.Solve the following homogeneous systems of linear equations by any method.

(a)
(b)
(c)
6. Solve the following homogeneous systems of linear equations by any

method.
(a)
(b)
−11x1 + 2x2 − 14x3 = 0
23x1 − 6x2 + 33x3 = 0
14x1 − 2x2 + 17x3 = 0
(c)
x1 − x2 + 2x3 = 0
2x1 + x2 + x3 = 0
x1 + x2 = 0
(d)
2x1 + x2 + 7x3 − 7x4 = 0
−3x1 + 4x2 − 5x3 − 6x4 = 0
x1 + x2 + 4x3 − 5x4 = 0
(e)
(c)
8. Solve the following systems by any method.
(a)
(b)
9. Solve the following systems, where a, b, and c are constants.
(a)
(b)
10. For which values of a will the following system have no solutions? Exactly
one solution? Infinitely many solutions?
11. Reduce
to reduced row-echelon form.

12. Find two different row-echelon forms of
13. Solve the following system of nonlinear equations for the unknown angles α,
β, and γ, where , , and
.
13. Show that the following nonlinear system has 18 solutions if ,

, and .
14. For which value(s) of λ does the system of equations have nontrivial
solutions?
15. Solve the system for , , and in the two cases ,

.
16. Solve the following system for x, y, and z.

1.3 Determinants of Matrices and Cramer’s Rule
In the previous subsection we defined matrices as arrays of numbers:
 a11 a12 ... a1n 

a a 22 ... a 2 n 
A  [aij ]   21 (1.17)
 ... ... ... ... 
 
a m1 am2 ... a mn 
In the above array there are m rows and n columns, and it is called an m n matrix.
DEFINITION 1.4 The columns

 a1k 
Vk   ... , k  1,2,...., n
a mk 
of the m n matrix in (1.11) are called the column vectors of the matrix. A matrix
can be written in terms of its column vectors as follows:
A  [aij ]  [V1 ,V2 ,...,Vn ] .
DEFINITION 1.5 The determinant of a real 2  2 matrix is the real number
a a 12 
det A  det  11  a11 a 22  a12 a 21 .
a 22 
(1.18)
a 21
EXAMPLE 1.12 Find the determinant of the given 2  2 matrix.

 2 5  2 5
A  , det A  det    2.6  5.1  12  5  7 .
 1 6   1 6 
DEFINITION 1.6 Minors and cofactors. The ijth minor M ij of a real n  n square matrix, is the
determinant of the ( n  1)  ( n  1) matrix obtained from the matrix A , deleting it’s

kth row and j th column.
The ijth cofactor Aij of a real n  n square matrix is
Ai , j  (1) i  j M ij .
Aij is also called the cofactor of the matrix element aij .
DEFINITION 1.7 The determinant of a real n  n square matrix is the real number
 a11 a12 ... a1n 

a a 22 ... a 2 n 
det A  det  21  a k1 Ak1  a k 2 Ak 2  ...  a kn Akn . (1.19)
 ... ... ... ... 
 
a n1 an2 ... a nn 
The expansion (1.19) in the above definition, to find the determinant, is called the
cofactor expansion of the determinant along the kth row. It is also possible to
expand the determinant using the elements of any column.
det A  a1 j A1 j  a 2 j A2 j  ...  a nj Anj
is the expansion of the determinant along the j th column.
EXAMPLE 1.13 Find the determinant of the given 3 3 matrix by the use of cofactor expansion of
the determinant along the first row.
2 3 4
A   2  3 1
 det A  a11A11  a12 A12  a13 A13
 3 2 7
where
- 3 1
M 11  det   -3.7 - 1.2  -23, A11  (1)11 M 11  23,
2 7 
- 2 1
M 12  det    -2.7 - 1.3  -17, A12  (1)1 2 M 12  17,
3 7
- 2 - 3
M 13  det    -2.2 - (-3).3  5, A13  (1)13 M 13  5.
3 2
1.3 Determinants of Matrices and Cramer’s Rule 27
Therefore one has

det A  2.( 23)  3.17  4.5  25 .
EXAMPLE 1.14 Find the determinant of the given 4  4 matrix by the use of cofactor expansion of
the determinant along the first column.
1 2 3 4
0 5 6 7 
A det A  a11 A11  a 21 A 21  a31 A 31  a 41 A 41
0 0 8 9
 
2 4 6 9
where a 21  a31  0 . Therefore we need only to find A11 and A41 .
5 6 7 
M 11  det 0 8 9   5(72 - 54) - 6(0 - 36)  7(0 - 32)  82,
4 6 9 
A11  (1)11 M 11  82,
2 3 4
M 41  det 5 6 7  0(21 - 24) - 8(14 - 20)  9(12 - 15)  21,
0 8 9
A12  (1) 41 M 41  21
Therefore one has
det A  1.82  2.( 21)  40.
In the above A11 is found through the cofactor expansion of the related determinant
along the first row cofactor expansion, and A41 is found through the cofactor
expansion of the related determinant along the first column.
1.3.1 Elementary row and column operations and determinants
In the above calculations we experienced that if a row or column contain zeros,

computations of determinants need less computational efforts. We have an
experience in the echelon forms of the augmented matrices, which are achieved
through the elementary row operations, and which lead us matrices with more zeros,
and therefore to equivalent systems of algebraic equations whose solutions are

straightforward.
In the following theorem we’ll state the effects of elementary row (column)
operations on the determinant of a matrix.
THEOREM 1.3 The effect of elementary row (column) operations on the determinant of a matrix are
as follows:
1. Interchanging of two rows (columns), change the sign of the determinant,
2. Multiplying any of the rows (columns) by a number, causes the multiplication of
the determinant by the same number,
3. Multiplication of any of the rows (columns) by a number, and addition on
another row (column) does not change the determinant.
EXAMPLE 1.15 Find the determinant of the given 4  4 matrix, first simplifying it by the use of the
elementary row (column) operations.
1 2 1 1
2 1 3 3   2 R1  R 2  R 2
det A  det 
0 1 2 3  R1  R 4  R 4
 
 1 4 2 4 
1 2 1  1
0 R3  R 2
3 1 5 
 det  3 R 2  R3  R3
0 1 2 3 
   6R2  R4  R4
0 6 1 3 
1 2 1 1 
0 R3  R 4
1 2 3 
  det  2 R 4  R3  R3
0 0  5 14 
 
0 0 11  15
1 2 1  1
0 1  2 3 
 det  5 R3  R 4  R 4
0 0 1 13 
 
0 0  5 14 
1 2 1  1
0 1  2 3 
 det   79.
0 0 1 13 
 
0 0 0 79 
det A  a11 A11  a 21 A 21  a31 A 31  a 41 A 41
where a 21  a 31  a 41  0 , hence we need only to find A11 .
1 - 2 3 
M 11  det 0 1 13   1.1.79  79, A11  (1)11 M 11  79.
0 0 79
Therefore one has
det A  1.79  79.
In the above A11 is found through the cofactor expansion of the related determinant
along the first column.
DEFINITION 1.7 Singular matrices. An n  n square matrix A is called to be singular if det A  0 .

It is clear that matrices with rows (columns) wit all zero entries are singular. Also it
can be shown that, a matrix A is singular if it has linearly dependent sets of rows
(columns).
1.3.2 Cramer’s rule1

Linear system of algebraic equations with nonsingular coefficient matrices can be
solved by the rule given in the following theorem which is due to Cramer.
1
Gabriel Cramer (1704–1752) was a Swiss mathematician. Although Cramer does not rank with the great
mathematicians of his time, his contributions as a disseminator of mathematical ideas have earned him a well-deserved
place in the history of mathematics. Cramer traveled extensively and met many of the leading mathematicians of his
day.
Cramer's most widely known work, Introduction à l'analyse des lignes courbes algébriques (1750), was a study and
classification of algebraic curves; Cramer's rule appeared in the appendix. Although the rule bears his name, variations
of the idea were formulated earlier by various mathematicians. However, Cramer's superior notation helped clarify and
popularize the technique.
Overwork combined with a fall from a carriage led to his death at the age of 48. Cramer was apparently a good-natured
and pleasant person with broad interests. He wrote on philosophy of law and government and the history of
mathematics.
He served in public office, participated in artillery and fortifications activities for the government, instructed workers
on techniques of cathedral repair, and undertook excavations of cathedral archives. Cramer received numerous honors
for his activities.
THEOREM 1.4 Let the coefficient matrix of a linear system of n algebraic equations for n unknowns
{x1 , x 2 ,...., x n }
a11 x1  a12 x 2  ...  a1n x n  b1
a 21 x1  a 22 x 2  ...  a 2 n x n  b2
(1.20)
........................
a n1 x1  a n 2 x 2  ...  a nn x n  bn
and the vector b of the right hand sides be as follows:
 a11 a12 ... a1n 

a  b1 
... a 2 n 
 [V1 , V2 ,..., Vn ] , b   ...  .
a 22
A  [aij ]   21
 ... ... ... ...   
  bn 
a n1 an2 ... a nn 
Let us define a matrix corresponding to each variable as follows:
Axi  [V1 , V2 ,..., Vi 1 , b, Vi 1 ,.....Vn ] .
If the coefficient matrix A is not singular, then the unique solution of (1.15) is
obtained by the formulae:
det Axi
xi  , i  1,2,..., n .
det A
This theorem can be proven after the theorems related to the algebraic operations of
matrices are proven.
EXAMPLE 1.16 Find the solution of the following linear system of algebraic equations by the use of
the Cramer’s rule.
6x  7 y  3
8 x  9 y  4.
In the above
6 7   3 3 7  6 3 
A  , b   4 , Ax  4 9  , A y  8 4,
8 9      
and hence
det A  2, det Ax  1, det Ay  0,
1 1 0
x  , y  0.
2 2 2
5x  2 y  2 z  1
x  5 y  3z  2.
5x  3 y  5z  2
In the above
5 2  2 1
A  1 5  3, b   2 ,
 
5  3 5   2 
1 2  2 5 1  2 5 2 1 
Ax   2 5  3, Ay  1  2  3, Az  1 5  2
    
 2  3 5  5 2 5  5  3 2 
and hence
det A  96, det Ax  32, det Ay  64, det Az  32,
32 1  64 2  32 1
x  , y  , z  .
96 3 96 3 96 3
2 x  5 y  2 z  3w  3
3 x  6 y  5 z  2 w  2.
4 x  5 y  14 z  14 w  11
5 x  10 y  8 z  4 w  4.
In the above
2 5 2  3 3
3 6 5 2   
A , b   2  ,
4 5 14 14  11
   
5 10 8 4  4
and hence
det A  5, det Ax  330, det Ay  135, det Az  30, det Aw  20,
330  135  30  20
x  66, y  27, z  6, w  4.
5 5 5 5
x  2y  z  w  u  2
x  2 y  4 z  w  5u  3.
3 x  4 y  5 z  w  2u  7
x  3 y  4 z  5w  u  4
2 x  5 y  8 z  4 w  u  7.
In the above
1 2 1  1  1  2
1 2 4 1 5    3
  
A  3 4 5  1  2  , b  7  ,
   
1 3 4 5  1  4
2 5 8 4 1  7
and hence
det A  126, det Ax  231, det Ay  0,
det Az  42, det Aw  21, det Au  0,
 231 11 0  42 1
x  , y  0, z   ,
 126 6  126  126 3
 21 1 0
w  , u  0.
 126 6  126
EXAMPLE 1.20 Use Cramer's rule to solve

Solution
Therefore,
Exercises 1.3
1. Let
(a) Find all the minors of A.

(b) Find all the cofactors A.
2. Let
Find , , , , , , , .
3. Evaluate the determinant of the matrix in Exercise 1 by a cofactor expansion

along
(a) the first row

(b) the first column
(c) the second row
(d) the second column
(e) the third row
(f) the third column
4. For the matrix in Exercise 1, find

(a) Adj(A)
(b) Inverse of A
In Exercises 5–10 evaluate by a cofactor expansion along a row or column of

your choice.
5. 6.
7. 8.
9. 10.
In Exercises 11–14 find inverse of A.
11. 12.
13. 14.
15.
In Exercises 16–21 solve by Cramer's rule, where it applies.
16. 17.
18. 19.
20. 21.
22. Show that the matrix
is invertible for all values of ; then find the inverse of A.

23. Use Cramer's rule to solve for y, without solving for x, z, and w .
24. Let be the system in Exercise 23.

(a) Solve by Cramer's rule.
(b) Solve by Gauss–Jordan elimination.
(c) Which method involves fewer computations?
25. Prove that if and all the entries in A are integers, then all the entries in are
integers.
26. Let be a system of linear equations in unknowns with integer

coefficients and integer constants. Prove that if, the solution has integer
entries.
27. Prove that if A is an invertible lower triangular matrix, then is lower

triangular.
28. Derive the last cofactor expansion listed in Formula 4.
29. Prove: The equation of the line through the distinct points and can be
written as
30. Prove: , , and are collinear points if and only if
31. (a)
If is an “upper triangular” block matrix, where and are square matrices, then
. Use this result to evaluate for
(b) Verify your answer in part (a) by using a cofactor expansion to evaluate .
32. Prove that if A is upper triangular and is the matrix that results when the
ith row and th column of A are deleted,
then is upper triangular if .
33. What is the maximum number of zeros that a matrix can have without
having a zero determinant? Explain your reasoning.
34. Let A be a matrix of the form

How many different values can you obtain for by substituting numerical
values (not
necessarily all the same) for the *'s? Explain your reasoning.
35. Indicate whether the statement is always true or sometimes false. Justify
your answer by giving
a logical argument or a counterexample.
(a) is a diagonal matrix for every square matrix .
(b) In theory, Cramer's rule can be used to solve any system of linear
equations, although
the amount of computation may be enormous.
(c) If A is invertible, then must also be invertible.
(d) If A has a row of zeros, then so does .
1.4 Matrices
In this section we begin our study of matrix theory by giving some of the
fundamental definitions. We shall see how to define arithmetic operations of
addition, subtraction, and multiplication in the set of matrices.
In Section 1.2 we used rectangular arrays of numbers, called augmented

matrices, to abbreviate systems of linear equations. However, matrices as
rectangular arrays of real numbers arise in many contexts.
EXAMPLE 1.20 For example, the following rectangular array with three rows and seven columns
might describe the number planes that take off from Istanbul Ataturk Airport to four
continents during a certain week:
M TU W TH F S SU
AF 45 33 78 90 55 63 17
AM 72 45 78 90 57 97 45
AU 12 17 36 33 23 24 12
EU 102 89 110 145 167 123 112
The following rectangular array of numbers with four rows and seven columns,
called a “matrix” can summarize the information above:
 45 33 78 90 55 63 17 
 72 45 78 90 57 97 45 

 12 17 36 33 23 24 12 
 
102 89 110 145 167 123 112
DEFINITION 1.8 A matrix is a rectangular array of numbers. The numbers in the array are called the
entries in the matrix.
EXAMPLE 1.21 Some examples are,

2
 2 5 1 6 9 3
 2 0 0 7 
5  
A , B  4, C  7  , D  3  1 0
0 1 1 6 3  
  4
5 7 9 0 3
7 
The size of a matrix is described in terms of the number of rows (horizontal lines)
and columns (vertical lines) it contains. For example, the matrix A in Example 1.21
has four rows and five columns, so its size is 4 by 5 (written4 5). In a size
description, the first number always denotes the number of rows, and the second
denotes the number of columns. The remaining matrices in Example 1.21 have sizes
1 1, 5 1 , and 1 3 respectively. A matrix with only one column is called a
column matrix (or a column vector), and a matrix with only one row is called a row
matrix (or a row vector). Thus, in Example 1.21 the matrix C is a column matrix, the
matrix D is a row matrix, and the matrix B is both a row matrix and a column
matrix.
We shall use capital letters to denote matrices and lowercase letters to denote
numerical quantities; thus we might write
When discussing matrices, it is common to refer to numerical quantities as scalars.

Unless stated otherwise, scalars will be real numbers; complex scalars will be
considered later on.
The entry that occurs in row i and column j of a matrix A will be denoted by .
Thus a general 3 4 matrix might be written as
and a general matrix as
(1.21)
When compactness of notation is desired, the preceding matrix can be written as

1.4 Matrices 41
the first notation being used when it is important in the discussion to know the size,
and the second being used when the size need not be emphasized. Usually, we shall
match the letter denoting a matrix with the letter denoting its entries; thus, for a
matrix B we would generally use for the entry in row i and column j, and for a
matrix C we would use the notation .
The entry in row i and column j of a matrix A is also commonly denoted by the
symbol . Thus, for matrix (1.21) above, we have
and for the matrix

2 1 7
3 4 2
we have 2, 1, 7, 3, 4, 2.
Row and column matrices are of special importance, and it is common practice to
denote them by boldface lowercase letters rather than capital letters. For such
matrices, double subscripting of the entries is unnecessary. Thus a general 1
row matrix a and a general 1 column matrix b would be written as
A matrix A with n rows and n columns is called a square matrix of order n, and the
shaded entries , , ,… in (1.22) are said to be on the main diagonal of
A.
(1.22)
1.4.1 Operations on Matrices
So far, we have used matrices to abbreviate the work in solving systems of linear
equations. For other applications, however, it is desirable to develop an “arithmetic
of matrices” in which matrices can be added, subtracted, and multiplied in a useful

way. The remainder of this section will be devoted to developing this arithmetic.
DEFINITION 1.8 Two matrices are defined to be equal if they have the same size and their
corresponding entries are equal.
, , , , .
EXAMPLE 1.21 Consider the matrices
If , x=2 then A=B, but for all other values of x the matrices A and B are not equal,
since not all of their corresponding entries are equal.
DEFINITION 1.9 If A and B are matrices of the same size, then the sum is the matrix obtained by
adding the entries of B to the corresponding entries of A. Matrices of different sizes
cannot be added.
In matrix notation,
.
2 7 1 5
,
3 4 6 0
Then
2 1 7 5 3 2
.
3 6 4 0 9 4
DEFINITION 1.10 The matrix 0 with all entries zero is called the zero matrix.
Addition by a zero matrix does not alter the matrix.

0 .
1.4 Matrices 43

2 7 0 0
,
3 4 0 0
Then
2 0 7 0 2 7
.
3 0 4 0 3 4
DEFINITION 1.11 The additional inverse of a matrix is .

0 .
EXAMPLE 1.24 Consider the matrix

2 1 7
3 4 2
Then
2 1 7
3 4 2
and
2 2 1 1 7 7 0 0 0
.
3 3 4 4 2 2 0 0 0
THEOREM 1.5 Matrix addition is commutative
and associative
Proof , and in real numbers, addition is commutative , and

hence
Similarly the associativity of the matrix addition is a result of the associativity of the
addition in real numbers.
Consider the matrices

2 1 7 3 2 4 0 7 1
, ,
3 4 2 6 0 1 2 3 4
Then
2 3 1 2 7 4 5 3 3
3 6 4 0 2 1 3 4 3
3 2 2 1 4 7
6 3 0 4 1 2
and
2 3 0 1 2 7 7 4 1 5 10 2
3 6 2 4 0 3 2 1 4 1 7 7
2 3 0 1 2 7 7 4 1
=
3 6 2 4 0 3 2 1 4
DEFINITION 1.10 If A is any matrix and c is any scalar, then the product is the matrix obtained by
multiplying each entry of the matrix A by c. The matrix is said to be a scalar
multiple of A.
In matrix notation,
.
EXAMPLE 1.25 The scalar multiple of ,

2 7
3 4
For scalar c=6;
6 2 6 7 12 42
6 .
6 3 6 4 618 24
It is common practice to denote 1 by – and 1 by – .
If , ,…, are matrices of the same size and , ,…, are scalars, then an
expression of the form
is called a linear combination of , ,…, with coefficients , ,…, .

1.4 Matrices 45
EXAMPLE 1.26 Consider the matrices A, B, and C;

2 3 4 1 2 0 6 3 9
, , .
1 2 1 1 4 1 12 15 0
Then
2 3 4 1 2 0 1 6 3 9
2 5
1 2 1 1 4 1 3 12 15 0
4 5 2 6 10 1 8 0 3 1 3 11
2 5 4 4 20 5 2 5 0 11 11 3
is the linear combination of A, B, and C with scalar coefficients 2, −5, and 1/3.
THEOREM 1.6 Multiplication of matrices by scalars is associative and distributive:
Proof Proof of the theorem relies on the associativity of the multiplication in real numbers,
,
and distribution of the multiplication over addition:
.
Thus far we have defined multiplication of a matrix by a scalar but not the
multiplication of two matrices. Since matrices are added by adding corresponding
entries, it would seem natural to define multiplication of matrices by multiplying
corresponding entries. However, it turns out that such a definition would not be very
useful for most problems.
Experience has led mathematicians to the following more useful definition of matrix
multiplication.
DEFINITION 1.11 Row-column multiplication If A is an matrix and B is an matrix,

then the product is the matrix where
∑ .
That is, to find the entry in row i and column j of ,

1) single out row i from the matrix A and column j from the matrix B.
2) multiply the corresponding entries from the row and column together, and
then add up the resulting products.

6 12
2 3 4
, 3 15 .
1 2 1
9 0
Since A is a 2 3matrix and B is a 3 2 matrix, the product is a 2 2 matrix. To
determine, for example, the entry in row 2 and column 1 of , we single out row
2 from A and column 1 from B. Then, as illustrated below, we multiply
corresponding entries together and add up these products.
1 6 2 3 1 9 21
The entry in row 2 and column 2 of is computed as follows:
1 12 2 15 1 0 42
The computations for the first row entries are

2 6 3 3 4 9 57
2 12 3 15 4 0 69.
Therefore
57 69
.
21 42
The definition of matrix multiplication requires that the number of columns of the
first factor A be the same as the number of rows of the second factor B in order to
form the product . If this condition is not satisfied, the product is undefined. A
convenient way to determine whether a product of two matrices is defined is to write
down the size of the first factor and, to the right of it, write down the size of the
second factor. If, as in (1.23), the inside numbers are the same, then the product is
defined. The outside numbers then give the size of the product.
(1.23)
1.4 Matrices 47
EXAMPLE 1.28 Determining Whether a Product Is Defined

Suppose that A, B, and C are matrices with the following sizes:
Then by (1.23), AB is defined and is a 7 3 matrix; BC is defined and is a 4 3

matrix; CA and is defined and is a 7 4 matrix.
The products BA, AC, and CB are all undefined.
Partitioned Matrices
A matrix can be subdivided or partitioned into smaller matrices by inserting
horizontal and vertical rules between selected rows and columns. For example, the
following are three possible partitions of a general 3 4 matrix A—the first is a
partition of A into four submatrices , , , ; the second is a partition of A
into its row matrices , , ; and the third is a partition of A into its column
matrices , , , :
Matrix Multiplication by Columns and by Rows

Sometimes it may be desirable to find a particular row or column of a matrix
product without computing the entire product. The following results, whose proofs
are left as exercises, are useful for that purpose:
(1.24)
(1.25)
EXAMPLE 1.29 If A and B are the matrices

6 12
2 3 4
, 3 15 ,
1 2 1
9 0
then from (1.24) the second column of matrix AB can be obtained by the
computation
12
2 43 2 12 3 15 4 0 69
15
1 12 1 12 2 15 1 0 42
0
and from (1.25) the first row matrix of can be obtained by the computation
6 12
2 3 4 3 15
9 0
2 6 3 3 4 9 2 12 3 15 4 0 57 69 .
If , ,…, denote the row matrices of A and , , … , denote the column

matrices of B, then it follows from Formulas (1.24) and (1.25) that
(1.26)
(1.27)
Matrix Products as Linear Combinations

Row and column matrices provide an alternative way of thinking about matrix
multiplication. For example, suppose that
Then
(1.28)
1.4 Matrices 49
In words, (1.28) tells us that the product Ax of a matrix A with a column matrix x is
a linear combination of the column matrices of A with the coefficients coming from
the matrix x. In the exercises we ask the reader to show that the product of a 1
matrix y with an matrix A is a linear combination of the row matrices of A
with scalar coefficients coming from y.
EXAMPLE 1.30 The matrix product
can be written as the linear combination of column matrices
The matrix product
can be written as the linear combination of row matrices
It follows from (1.26) and (1.27) that the jth column matrix of a product AB is a
linear combination of the column matrices of A with the coefficients coming from
the jth column of B.
EXAMPLE 1.31 We can easily show that
The column matrices of AB can be expressed as linear combinations of the column

matrices of A as follows:
Matrix Form of a Linear System

Matrix multiplication has an important application to systems of linear equations.
Consider any system of m linear equations in n unknowns.
Since two matrices are equal if and only if their corresponding entries are equal, we
can replace the m equations in this system by the single matrix equation
The matrix on the left side of this equation can be written as a product to give
If we designate these matrices by A, x, and b, respectively, then the original system

of m equations in n unknowns has been replaced by the single matrix equation
.
The matrix A in this equation is called the coefficient matrix of the system. The
augmented matrix for the system is obtained by adjoining b to A as the last column;
thus the augmented matrix is
.
1.4 Matrices 51
Matrices Defining Functions

The equation with A and b given defines a linear system to be solved for x.
But we could also write this equation as , where A and x are given. In this
case, we want to compute y. If A is , then this is a function that associates with
every 1column vector x an 1 column vector y, and we may view A as
defining a rule that shows how a given x is mapped into a corresponding y. This idea
is discussed in more detail forthcoming chapters.
EXAMPLE 1.32 Consider the following matrices.
The product is
so the effect of multiplying A by a column vector is to change the sign of the second
entry of the column vector. If we view the column vector x as locating a point in the
plane, then the effect of A is to reflect the point about the x-axis.
Figure 1.1
For the matrix
the product is
so the effect of multiplying B by a column vector is to interchange the first and

second entries of the column vector, also changing the sign of the first entry. The
effect of B is to rotate the line segment from the origin to the point through a right
angle
Figure 1.2
Transpose of a Matrix
DEFINITION 1.12 If A is any matrix, then the transpose of A, denoted by , is defined to be the
matrix that results from interchanging the rows and columns of A; that is, the first
column of is the first row of A, the second column of is the second row of A,
and so forth.
EXAMPLE 1.32 The following are some examples of matrices and their transposes.
Observe that not only are the columns of the rows of A, but the rows of are
the columns of A. Thus the entry in row I and column j of is the entry in row j and
column i of A; that is,
(1.29)
Note the reversal of the subscripts.
In the special case where A is a square matrix, the transpose of A can be obtained by
interchanging entries that are symmetrically positioned about the main diagonal. In
12 it is shown that can also be obtained by “reflecting” A about its main diagonal.
1.4 Matrices 53
(1.30)
1.4.2 Properties of Matrix Multiplication
AB and BA need not be equal. Equality can fail to hold for three reasons: It can
happen that the product AB is defined but BA is undefined. For example, this is the
case if A is a matrix and B is a matrix. Also, it can happen that AB and BA are both
defined but have different sizes. It is also possible to have even if both AB
and BA are defined and have the same size.

1 0 1 2
,
4 5 4 0
Then one has
1 2 7 10
, , .
24 8 4 0
THEOREM 1.7 When matrix multiplications are defined, the following equalities are hold:
1) (associative)
2) (left distributive)
3) (right distributive)
4) (associativity of multiplication by a scalar)
EXAMPLE 2 Associativity of Matrix Multiplication

As an illustration of the associative law for matrix multiplication, consider
Then
Thus
and
so , as guaranteed by Theorem 1.4.1c.
EXAMPLE 3 The Cancellation Law Does Not Hold

Consider the matrices
You should verify that
Thus, although , it is incorrect to cancel the A from both sides of the equation and
write . Also, ,
yet and . Thus, the cancellation law is not valid for matrix multiplication, and it is
possible for a product of matrices
to be zero without either factor being zero.
In spite of the above example, there are a number of familiar properties of the real
number 0 that do carry over to zero matrices.
Some of the more important ones are summarized in the next theorem. The proofs
are left as exercises.
Identity Matrix
DEFINITION 1.13 The square matrix with all ones on the main diagonal and all zero off
diagonals is called identity matrix .
For example
1 0 0
0 1 0.
0 0 1
THEOREM 1.8 1) For any 1 row matrix b one has:

. (1.31)
2) For any 1 column matrix c one has:
. (1.32)
Proof 1) Let the row matrix … then by the Definition 1.11

, ∑ , .
2) Let column matrix c be
1.4 Matrices 55
then by the Definition 1.11

, ∑ , .
THEOREM 1.9 For any matrix A one has:

.
Proof From (1.31), and (1.27) one deduces that
While (1.32), and (1.26) yields

… … … .
EXAMPLE 1.34 The following example shows the effect of the unit matrix in multiplication.
1 0 0 8 3 4 8 3 4 1 0 0 8 3 4
0 1 0 5 1 9 5 1 9 0 1 0 5 1 9
0 0 1 4 5 6 4 5 6 0 0 1 4 5 6
DEFINITION 1.14 For an matrix A, the square matrix with the property
is called the multiplicative inverse of A, and A is said to be invertible. If no such

matrix can be found, then A is said to be singular.
EXAMPLE 1.35 Let

1 0 2
0 1 1
1 1 2
Then
1 2 2 1 0 0
1 0 1 , 0 1 0.
1 1 1 0 0 1
EXAMPLE 1.36 The matrix

1 4 0
2 5 0
3 6 0
is singular. To see why, let
be any matrix. The third column of AB is

0
0
0
Thus
1 0 0
0 1 0
0 0 1
whose third column is
0
0 .
1
Properties of Inverses
It is reasonable to ask whether an invertible matrix can have more than one inverse.
The next theorem shows that the answer is no—an invertible matrix has exactly one
inverse. Proof is easy and left to the reader.
THEOREM 1.10 If B and C are both inverses of the matrix A, then B=C.
The following theorem gives conditions under which a matrix is invertible and
provides a simple formula for the inverse.
1.4 Matrices 57
THEOREM 1.11 The matrix A is invertible if det 0 , in which case the inverse is given by the
formula
Where Aij is the ijth cofactor of the real n  n square matrix A given by the
Definition 1.6.
Proof Recall that . By the Definition 1.11 of the matrix multiplication

1,
, ∑ .
0,
THEOREM 1.12 If A and B are invertible matrices of the same size, then AB is invertible and
.
Proof
.
Although we will not prove it, this result can be extended to include three or more
factors; that is, a product of any number of invertible matrices is invertible, and the
inverse of the product is the product of the inverses in the reverse order:
… … .
Applying the formula in Theorem 1.12, we obtain
Also,
Therefore as guaranteed by Theorem 1.12.

Powers of a Matrix
Next, we shall define powers of a square matrix and discuss their properties.
DEFINITION 1.14 If A is a square matrix, then we define the nonnegative integer powers of A to be
, … , 0.
Moreover, if A is invertible, then we define the negative integer powers to be

… , 0.
Because this definition parallels that for real numbers, the usual laws of exponents
hold.
THEOREM 1.12 If A is an invertible matrix, then:

(a) is invertible and .
(b) is invertible and for 0,1,2, … .
(c) For any nonzero scalar k, the matrix kA is invertible and ..
Proof is left to the reader.
EXAMPLE 1.37 Let
Then
Polynomial Expressions Involving Matrices

If A is a square matrix, say , and if
(1.33)
1.4 Matrices 59
is any polynomial, then we define
where I is the identity matrix. In words, is the matrix that results when
A is substituted for x in (1.33) and is 1 replaced by I.
EXAMPLE 1.37 Let
Then
Properties of the Transpose

The next theorem lists the main properties of the transpose operation.
THEOREM 1.12 If the sizes of the matrices are such that the stated operations can be performed, then
(a) ,
(b)
(c) , where k is any scalar
(d) .
If we keep in mind that transposing a matrix interchanges its rows and columns,
parts (a), (b), and (c) should be self-evident.
Invertibility of a Transpose
The following theorem establishes a relationship between the inverse of an
invertible matrix and the inverse of its transpose.
THEOREM 1.12 If A is an invertible matrix, then is also invertible and

. (1.34)
Proof .
.
which completes the proof.
Applying Theorem 1.12 yields

Chapter 1

Transféré par

Informations du document

Description originale:

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Chapter 1

Transféré par

Droits d'auteur :

Formats disponibles

1.1 Historical Remarks on Algebraic Equations

Muhammad Ibn Musa al-Khwarizmi to Whom we Owe the Word Algebra

1.2.1 The method of Gauss1-Jordan2 elimination

1.2.2 The method of echelon forms

 a11 a12 ... a1n 

The coefficients of a linear system of equations

 a11 a12 a13 

 a11 a12 a13 b1 

DEFINITION 1.2 An augmented matrix in row echelon form if

If we continue to achieve the reduced row echelon form, we have

and finally the operation  3R 2 / 2  R1  R1 leads to the reduced row echelon

We observe that r  m  n  3 , according to Theorem 1.1, the linear system of

a) For a  2 the echelon form is

We observe that r  e  n  3 , according to Theorem 1.1, the linear system of

We observe that r  e  2  n  3 , according to Theorem 1.1, the linear system of

Applying the three elementary row operations below we get

1.2.3 Homogeneous systems

equal to the number of the nonzero rows, e  r . Depending on the number n of

We observe that r  e  n  3 , according to Theorem 1.2, the linear system of

We observe that r  e  2  n  3 , according to Theorem 1.2, the linear system of

EXAMPLE 1.11 Gaussian Elimination

by Gaussian elimination and back-substitution.

to the row-echelon form

The system corresponding to this matrix is

Solving for the leading variables yields

Substituting the bottom equation into those above yields

1. Solve each of the following systems by Gauss–Jordan elimination.

3. Solve each of the following systems by Gauss–Jordan elimination.

4. Solve each of the following systems by Gauss–Jordan elimination.

2x1 − 2x2 + x3 + x4 + 8x5 = 22

5. Without using pencil and paper, determine which of the following

6.Solve the following homogeneous systems of linear equations by any method.

6. Solve the following homogeneous systems of linear equations by any

8. Solve the following systems by any method.

9. Solve the following systems, where a, b, and c are constants.

to reduced row-echelon form.

12. Find two different row-echelon forms of

13. Show that the following nonlinear system has 18 solutions if ,

15. Solve the system for , , and in the two cases ,

16. Solve the following system for x, y, and z.

In the previous subsection we defined matrices as arrays of numbers:

 a11 a12 ... a1n 

DEFINITION 1.4 The columns

DEFINITION 1.5 The determinant of a real 2  2 matrix is the real number

EXAMPLE 1.12 Find the determinant of the given 2  2 matrix.

determinant of the ( n  1)  ( n  1) matrix obtained from the matrix A , deleting it’s

The ijth cofactor Aij of a real n  n square matrix is

Aij is also called the cofactor of the matrix element aij .

 a11 a12 ... a1n 

Therefore one has

where a 21  a31  0 . Therefore we need only to find A11 and A41 .

1.3.1 Elementary row and column operations and determinants

In the above calculations we experienced that if a row or column contain zeros,

and therefore to equivalent systems of algebraic equations whose solutions are

det A  a11 A11  a 21 A 21  a31 A 31  a 41 A 41

where a 21  a 31  a 41  0 , hence we need only to find A11 .

DEFINITION 1.7 Singular matrices. An n  n square matrix A is called to be singular if det A  0 .

1.3.2 Cramer’s rule1

 a11 a12 ... a1n 

EXAMPLE 1.20 Use Cramer's rule to solve

(a) Find all the minors of A.

3. Evaluate the determinant of the matrix in Exercise 1 by a cofactor expansion

(a) the first row