DAR-ES-SALAAM INSTITUTE OF TECHNOLOGY

4 3 1
− −
 27  3  16  2  1  2
1. Evaluate (a)   +  −  (b) 32 x − 3x +1 − 3x + 3 = 0 .
 8  9  49 
(c) log5 x = 16 log x 5 (d) log 3 x − 6 log x 3 = −1 (e) 4 x − 6(2 x ) − 16 = 0 (f) log25  log58

2. (a) One factor of f ( x) = x 3 − 5 x 2 + 2 x + 8 is ( x + 1) . Find the other two factors

. (b) The expression 2 x 3 + ax 2 + bx + 6 is exactly divisible by (x+ 2) and on

division by (x+1) gives a remainder of -12. Calculate the values of ‘a’ and ‘b’

(c) Find the remainder when the expression x5 − x 2 + 1 is divided by x + 2 ,

(d) One of the factors of the cubic polynomial 2 x 3 + 3 x 2 + ax + b is x − 2 . When the
polynomial is divided by ( x − 1) , a remainder of -5 results. Find a and b , and hence
factorize the polynomial completely.
(e) If the expression ax 4 + bx3 − x 2 + 2 x + 3 is divided by x 2 − x − 2 the remainder is
3x + 5 . Find the values of a and b.
(f) Find the value of n when the expression x n − 5 x 2 − 4 is divided by x − 3 leaves the
remainder of 40.
1
(g) If 2log y x + 2log x y = 5 , show that log y x , is either or 2. Hence find all pairs of
2
values of x and y which satisfy simultaneously the equation above and the equation
xy = 27 .

27 n + 2 − 6  3 3n +3
3. (a) Simplify
3n  9 n+2

(b) Solve the equations log3 x + 3 log x 3 = 4

(c) Solve for x and y for the following system of simultaneous equations
3 x+ y = 9 


5 2 x −3 y
= 625


1
1. Matrix

A representation of items (information) stored or arranged in an order fashion is simply

called a Matrix. Mathematically, the concept of a matrix refers to a set of numbers, variables
or functions ordered in rows and columns.

Definition
A matrix is a rectangular array of numbers, letters or symbols. Matrices are always denoted
by upper cases and the entries inside the array are called elements that are uniquely
identified by their positions; such as i th row and j th column.

Consider the Matrix A below.

 a11 a12 a1n 
 
 a21 a22 a2 n 
A=
 
 
 am1 am 2 amn 
A is the matrix of m rows and n columns. The art of identifying a matrix in terms of its
number of rows and columns is known as Order (size or dimension) of a matrix i.e. matrix A
above has m × n (pronounced as m by n) order.

A matrix is said to be a column matrix (column vector) if it has only one column and several
row(s). While a matrix with only one row and several column(s) is called a row matrix (row
vector).
 1 
 
 5  is a column matrix, while ( a11 a12 a1n ) is a row matrix.
 −2 
 

2 Matrix Operation

(a) Matrix addition (subtraction): The operation is defined if two or more matrices have
the same order or dimension.
Example
1 0 1 0   2 1 1 0 
(i) Find the value of  2 3 4 5 +  2 − 7 0 5 
   
1 4 8 4   4 5 3 9 
 5 0 4  1 5 − 2 
(ii)  − 
1 − 3 2   7 2 5 
(b) Matrix Multiplication
(i) Scalar Multiplication: Multiply each element in a matrix by a scalar
 3 5   15 25 
For example 5  = 
1 4   5 20 

2
 (ii) Matrix Multiplication: Multiplication (product) of two matrices A and B is
defined if and only if the matrices are compatible, i.e., the number of columns
of first matrix should be equal to the number of rows of the second matrix.

 1 −3 1 2 
7
Example: 1. ( 2 − 3 4 )  5 0  2.  3 4   
 −2 4  5 68 
   
 3 1 2  4 − 1 2 
3.   
 5 1 7  3 1 3 
 3 5 1 0 
4. If f ( x) = 2 x − 4 I , find f ( A) given A =  , I = 
 1 4  0 1

(c) Transpose of a Matrix: The transpose AT of an n × m matrix A is the m × n matrix

3 1 2
obtained by interchanging the rows and columns of A. Example A =   , its
5 1 7
3 5 
transpose is A = 1 1 
T

2 7
 

3. Properties of Matrices
(i) Matrix addition is Commutative A + B = B + A
(ii) Matrix addition is Associative A + (B + C) = (A + B) + C
(iii) Matrix multiplication is not Commutative A  B  B  A .
1 0 
However, if I =   , then AI = IA
 0 1

4. Special Types of Matrices

(i) Square matrix: A square matrix is a matrix of dimension n. i.e., a matrix with
the same number of rows and columns (n x n matrix).
1 0 0
 
 0 1 0
(ii) Unit matrix/ Identity matrix: is a square matrix denoted as I = ,
 
 
1 0 1
with 1’s along the leading diagonal (the one that goes from top left-hand corner
to the bottom right-hand corner), and 0’s elsewhere.
1 0 0 
So the 3 x 3 unit matrix is I =  0 1 0 
 0 0 1
 

3
(iii) A diagonal matrix is a square matrix A whose elements above and below the
leading (principal) diagonal are all zero, that is aij = 0 for all i  j . [Hint:
Any identity matrix is a diagonal matrix]
3 0 0
For example, A = 0 9 0

0 0 7 

(iv) Symmetrical Matrix:

A real square matrix A is said to be symmetric if it is equal to its transpose,
i.e., AT =A,

 −3 1 5  −3 1 5
Example A = 1 0 − 2 , AT =  1 0 − 2

5 −2 4   5 − 2 4 

(v) Skew-Symmetric Matrix:

A real square matrix A is said to be skew-symmetric if it is equal to its
negative transpose, i.e., AT = -A
0 −4 1 0 4 −1
Example A = 4  0 
− 5 , 
A =  −4 0
T
5 

 −1 5 0  1 − 5 0

(vi) Triangular Matrix:

A real square matrix A, whose elements above the principal diagonal (or
below the principal diagonal) are all zeros is called triangular matrix

1 0 0

T1 =  −2 3 0  → Lower Triangular Matrix
 −1 5 2 
Example
1 6 − 1
T2 = 0 2 3  → Upper Triangular Matrix
0 0 − 4 

(vii) Orthogonal Matrix:

A square matrix A is said to be an orthogonal matrix if ATA = I i.e., AT = A−1

4
 3 −6 2 
1
Example A = 2 3 6 
7
6 2 − 3 

5. Trace of a matrix
If A is the square matrix, then the trace of A denoted by tr(A) is defined as the sum of
all entries in the main diagonal of A
3 − 6 2 
Example A =  2 3 6  , tr ( A) = 3 + 3 + (−3) = 3

6 2 − 3 

6. Determinant and Inverse of a Matrix

The determinant of a matrix is a scalar (number), obtained from the elements of the
square matrices by specified operations. It is denoted by det(A) or |A| for a square
matrix A
If det( A) = A = 0 , then matrix A is called a singular matrix (non-inversible matrix)
If det( A) = A  0 , then matrix A is called non- singular matrix (inversible matrix)
Inverse of matrix A, is denoted as A−1

Determinant and inverse of 2 x 2 Order of a Matrix

a b 
Given matrix A =   , det( A) = ad − bc
c d 
1  d −b
Inverse of A is A−1 =  
det( A)  −c a 

5 2
Example A =   , det( A) = (5  4) − (1 2) = 18
1 4 

1  4 − 2
Inverse of matrix A, A−1 =  
18  −1 5 

5
7. Determinant, Co-factor, and inverse of 3 x 3 Order of a Matrix
 a11 a12 a13 
Consider matrix A =  a21 a22 a23 
 a31 a32 a33 

Minor and Co-factor

The minor of the entry aij in a given matrix A, denoted as Mij is defined to be the
determinant of order (n − 1) formed by deleting the i th row and j th column of A matrix.
The scalar (number) Cij = (−1)i+jMij is called the Co-factor of the element aij of matrix A.
a21 a23
For example, minor of a12 of matrix A is M 12 =
a31 a33

a21 a23 a21 a23

Cofactor of the element a12 of matrix A C12 = (−1)1+ 2 =−
a31 a33 a31 a33

Determinant of matrix A: This can be obtained by summing the products of aij Cij in any of

the rows or columns. For instance, using first row expansion, the determinant of A is:
3
det( A) = a11C11 + a12C12 + a13C13 =  aij Cij
j =1

Co-factor matrix and adjoint matrix.

 a11 a12 a13 
Given matrix A =  a21 a22 a23  , C is a co-factor element of a then, the matrix of co-
 ij ij

 a31 a32 a33 

C11 C12 C13 

factor matrix A is given as C = C21 C22 C23 .

 C31 C32 C33 

Adjoint and Inverse of a 3x3 Matrix

The transpose of cofactor matrix A is called the adjoint matrix and is denoted as adj (A).
T
C11 C12 C13 
 
Thus, Adj ( A) = C = C21 C22 C23
T

 C31 C32 C33 

6
 Adj ( A)
Inverse of a 3x3 Matrix: A−1 =
det( A)

Exercise
 −1 − 1 2 

1. Show that the matrix A = 1 3 − 4  is singular

 1 2 − 3 

1 1 1
2. Find the co-factor and adjoint matrix of A, where A =  2 1 − 1
 
3 − 1 − 2

x y  x 6  4 x + y
3. (a) Find x, y, z and w if: 3 = + .
 z w   −1 2w   z + w 3 
(b) Use inverse matrix and Cramer’s methods to solve the system of linear equations:
1 2 3   x  6
 2 1 1   y  = 5 
    
3 1 − 2  z  1 

 −5 10 8

(c) Evaluate the determinant of matrix A = 4 − 7 − 6 
 
 −3 6 5

(d) What is an adjoint matrix?

4. (a) Solve the following linear equations by inverse matrix method
x + 2 y + 3z = 6 , 2x + y + z = 5 , 3x + y − 2 z = 1

2x − y = 3 
(c) Use Cramer’s method to solve the following system of linear equations .
x + 3 y = 5

(d) Evaluate:
1 4 3
6 2 5 .
1 7 0

4. (a) Construct a 3  2 matrix A = ( aij ) with the following entries:

7
 a11 = 1, a21 = 2, a31 = 3, a12 = 4, a22 = 5, a32 = 6 .

(b) Find the transpose of matrix A, AT if

1 0 1 0 
A =  2 3 4 5  .
 4 4 4 4

(c) Use Cramer’s rule to solve the following system of linear equations:
4 x − 2 y = 6

x + y = 3

(d) Use Inverse method to solve the following system of linear equations
4 x + 2 y + 3 z = 17
6 x + 5 y + 5 z = 31
x+ y+ z = 6

 x '   5 2  x 
5. (a) The transformation   =    maps the triangle A(3, 2), B(7, 2), C(3, 8) onto
 y '  1 4  y 
the triangle A ' B ' C ' . Find the coordinates of A ' , B ' , C ' and calculate the area of the
triangle A ' B ' C ' .
 x '   3 2  x 
(b) The linear transformation is defined by   =    . (i) Find the image of (1, 0)
 y '   5 4  y 
and (0, 1). (ii)Find the point whose image is (4, 6).

x −3 1 −1
(c) Solve the following equation − 7 x + 5 −1 = 0
−6 6 x−2

8
2. LINEAR SYSTEMS: GAUSSIAN ELIMINATION METHOD
An equation in n variables x1, x2, …, xn is said to a linear if it is in the form
a1 x1 + a2 x2 + ... + an xn = b , where a1, a2, …, an and b are constants. Two or more linear

equations that are to be solved simultaneously are said to form a system of linear equations
or more simply, a linear system. Thus, the linear equations of n variables is represented as:
a11 x1 + a12 x2 + + a1n xn = b1
a21 x1 + a22 x2 + + a2 n xn = b2
(1)

an1 x1 + an 2 x2 + + ann xn = bn
In matrix notation the system of linear equations is:
 a11 a 12 a1n  x1   b1 
    
 a21 a22 a2 n  x2   b2 
= ; (2)
    
    
 an1 an 2 ann  xn   bn 

or can be written as AX = B , where,

 a11 a 12 a1n 
 
a a22 a2 n 
A =  21 is the coefficient matrix of n x n order,
 
 
 an1 an 2 ann 

 x1 
 
x
X =  2  is the column matrix of n variables
 
 
 xn 

 b1 
 
b
B =  2  is the column matrix of the n right-side constants.
 
 
 bn 
The augmented matrix of the linear system is a matrix formed by appending the right-side
constants to the coefficient matrix as an extra column. From equation (2) the augmented
matrix of the linear system is:

9
  a11 a 12 a1n b1 
 
a a22 a2 n b2 
A | b =  21
 
 
 an1 an 2 ann bn 
(3)

To solve a linear system, we use a procedure based on elimination of variables. The

systematic version of this procedure is called Gaussian Jordan elimination after Carl
Fredrich Gauss (1777 -1855), one of the greatest mathematicians of all time. The basic ideas
behind elimination of variables that do not affect the solution set of the linear system are that:
1. Exchanging two equations
2. Multiplying an equation by a nonzero constant
3. Adding a multiple of one equation to another equation.
Note: A linear system is said to be consistent if it has solution, otherwise, it is called
inconsistent.

2.1 Elementary Row Operations on a given Augmented Matrix

The sequences of elementary row operations are used to reduce the augmented matrix in (3)
into ‘simple’ augmented matrix of the form:
1 0 0 p1 
 
0 1 0 p2 
I | p= ,
 
 
0 0 1 pn 

that corresponds to a ‘simple’ linear system such as x1=p1 , x2=p2 , …, xn= pn

Basic elementary row operations are such as:

1. Exchanging two rows (denoted as Ri  R j )
2. Multiply (divide) a row by a nonzero constant say ‘a’ (denoted as aRi → Ri )
3. Add a multiple of one row to another row (denoted as aRi + R j → R j )
(Note: The arrow → means ‘replace’ and  means ‘exchange’ rows.

Example-1
Use the Gaussian elimination method to solve the following linear system:
3x + 4y = 5
x + 9y = 17

10
Solution: The augmented matrix of the given linear system is

3 4 5
𝐴|𝑏 = ( | )
1 9 17

Augmented matrix Operation

3 4 5
Put a 1 in a11 ( | ) R1  R2
1 9 17

1 9 17
Put a 0 in a21 ( | )
3 4 5
−3R1 + R2 → R2
1 9 17
Put a 1 in a22 ( | )
0 −23 −46
1
− R2 → R2
23
1 9 17
Put a 0 in a12 ( | ) −9R2 + R1 → R1
0 1 2

1 0 −1
( | ),
0 1 2
which is the simple reduced augmented matrix, so x = -1 and y = 2

Example-2
Use the Gaussian elimination method to solve the following linear system:
2p + 4q = 30
2p + 2q +6r = 30
6p + 4q + 4r = 50

2 4 0 30
Solution: The augmented matrix of the given linear system is 𝐴|𝑏 = (2 2 6|30)
6 4 4 50

Augmented matrix Operation

2 4 0 30 1
Put a 1 in a11 (2 2 6 | 30) R1 → R1
6 4 4 50 2

1 2 0 15 −2R1 + R2 → R2
Put a 0 in a21 and a31 (2 2 6 | 30)
6 4 4 50 −6R1 + R3 → R3

11
 1 2 0 15 1
Put a 1 in a22 (0 −2 6| 0 ) − R2 → R2
0 −8 4 −40 2

1 2 0 15
Put a 0 in a32 (0 1 −3| 0 ) 8R2 + R3 → R3
0 −8 4 −40

1 2 0 15 1
Put a 1 in a33 (0 1 −3 | 0 ) − R3 → R3
0 0 −20 −40 20

1 2 0 15
Put a 0 in a23 (0 1 −3| 0 ) 3R3 + R2 → R2
0 0 1 2

1 2 0 15
Put a 0 in a12 (0 1 0| 6) −2R2 + R1 → R1
0 0 1 2

1 0 0 3
(0 1 0 | 6)
0 0 1 2
which is the simple reduced augmented matrix, so p = 3, q = 6 and r = 2

Exercise-1
1. Use the Gaussian elimination method to solve the following linear systems:
 1 2 −2  x   1  4x + 9y + 2z = 21 x+y+z=4
    
(a)  2 5 1  y  =  9  (b) -8x + 6y - 3z = 41 (c) 2x - 3y + 4z = 33
 1 3 4  z   9 
     3x + y - 5z = -73 3x - 2y - 2z = 2

x - y + 3z = 8 2x1 + x 2 + x 3 = 5 2x1 - x 2 + x 3 = 3
(d) 2x - y + 4z = 11 (e) 4x1 - 6x 2 = -2 (f) 4x1 +x 2 +3x 3 = 15
-x + 2y - 4z = -11 -2x1 + 7x 2 + 2x 3 = 9 x1 -x 2 + 5x 3 = 14

x1 +2 x 2 + x 3 - x 4 = 5
3x1 +6x 2 +4x 3 + 4x 4 = 16
(g)
4x1 +4x 2 +3x 3 + 4x 4 = 22
2x1 + x 3 +5x 4 = 15
2. A d.c. circuit comprises three closed loops. Applying Kirchhoff’s laws to the closed loops
gives the following equations for current flow in milliamperes:

12
 2l1 + 3l2 – 4l3 = 26 (1)
l1 – 5l2 – 3l3 = −87 (2)
−7l1 + 2l2 + 6l3 = 12 (3)
Use the Gaussian elimination method to solve for l1, l2 and l3.

3. In a mass-spring-damper system, the acceleration 𝑥̈ m/s2, velocity 𝑥̇ m/s and displacement

x m are related by the following simultaneous equations:
6.2 x + 7.9 x + 12.6 x = 18.0
7.5 x + 4.8 x + 4.8 x = 6.39
13.0 x + 3.5 x - 13.0 x = -17.4
By using Gaussian elimination, determine the acceleration, velocity and displacement for the
system, correct to 2 decimal places.

13
3. LINEAR PROGRAMMING: SIMPLEX METHOD
The simplex method is the method used to solve linear programming model containing two or
more decision variables. The method is an iterative procedure for solving linear programming
problems expressed in the standard form. Or

The simplex method is an approach to solving linear programming models by hand using slack
variables, tableaus, and pivot variables as a means to finding the optimal solution of an
optimization problem. Simplex tableau is used to perform row operations on the linear
programming model as well as for checking optimality.

The standard linear programming form is like:

Maximize/minimize the objective function F = c1 x1 + c2 x2 + ... + cn xn

Subject to constraints of the form of ≤ type:

a11 x1 + a12 x2 + ... + a1n xn  b1
a21 x1 + a22 x2 + ... + a2 n xn  b2
or
am1 x1 + am 2 x2 + ... + amn xn  bm
x1 , x2 ,..., xn  0

Subject to constraints of the form of ≥ type:

a11 x1 + a12 x2 + ... + a1n xn  b1

a21 x1 + a22 x2 + ... + a2 n xn  b2

am1 x1 + am 2 x2 + ... + amn xn  bm

x1 , x2 ,..., xn  0

Where, bj  0 for j = 1, 2,..., m.

That is, in order to qualify as a standard linear programming must have the following
features:

1. The objective function F is to be maximized or minimized.

2. The constraints of the ≤ type are for maximizing the objective function F, while the
constraints of ≥ type are for minimizing the objective function F.
3. All decision variables xi are nonnegative (xi ≥ 0).
4. The right-side constants bj in each constraint inequality are nonnegative ( b j  0 ).

14
3.1 Definitions of some Terminologies when solving Linear Programming by Simplex
Method
1) Basic variable: is the one that appears with a unit coefficient in that equation and zeros in
other equations.
2) Non-basic variables: are those variables which are not basic.
3) Pivot operation: is a sequence of elementary row operation that reduces a given system to
an equivalent system in which a specified variable has a unit coefficient in one equation
and zeros elsewhere.
4) Basic solution: the solution obtained by setting all non-basic variables to zero and solving
for basic variables.
5) Slack variable: is the variable added to inequality of ≤ type to make an equation;
e.g. x1 ≤ m changed into equation x1 +s1 =m. Slack variable represent the difference (takes
up the slack) between the left and right sides of an inequality.
6) Equational form: is equation obtained by writing objective function and constraints of
the linear programming model into equation form.
7) Initial Table: is the augmented matrix of the equational form.
8) Indicators: the numbers in the bottom row in the initial table obtained from the objective
function.
7) Pivot column: is the column in the simplex table that corresponds to the most negative
indicator. The column is containing a non-basic variable that is chosen to entering variable
a basic variable at a particular step of the simplex method.
9) Pivot row: is found by dividing each positive entry of the pivot column into the
corresponding right-side constant and picking the smallest ratio.
10) Terminal Table: is the last table for obtaining optimal solution: for maximum objective
function, all indicators must be non-negative, while minimization problem the indicators
negative.
11) Optimal solution: is the solution set that produces the optimum value (maximize or
minimize) the objective function.

3.2 Steps for Simplex Method for a Standard Maximum Linear Programming
(1) Set up the initial tableau for the given standard maximum problem. Find the most
negative indicator. This determines the pivot column.
(2) Divide each positive entry of the pivot column into the corresponding right-side constant.
The positive entry that produces the smallest ratio determines the pivot row. The entry in the
pivot row and pivot column is the pivot.
(3) Use elementary row operation to replace the pivot entry aij to be a 1 and then to change
other entries of the pivot column to zero.
(4) If possible, repeat steps (1) through (3) until all indicators are non-negative.

15
Example-1

Maximize the function F = 4 x1 + 5 x2

2 x1 + 3x2  9
Subject to x1 + x2  4
x1 , x2  0
Solution
First introduce slack variables s1 and s2 and write inequalities in equational form:
2 x1 + 3x2 + s1 =9
x1 + x2 + s2 = 4
x1 , x2 , s1 , s2  0
Equational of objective function F = 4 x1 + 5 x2 is written as −4 x1 − 5 x2 + F = 0
Step 1: Initial table
x1 x2 s1 s2 F
2 3 1 0 0 9
(⋯ 1 1 0 1 0 4
⋯ ⋯ ⋯ ⋯|⋯)
−4 −5 0 0 1 0

The basic variables associated with the initial tableau are s1 and s2 and the associated feasible
solution is x1 =0, x2 =0, s1 = 9, s2 =4

The most negative indicator is -5, therefore, column two of variable x2 is a pivot column. To
find a pivot row use minimum ratio rule:

Minimum ratio table

Row No Basic variable Ratio
1 s1 9/3 =3 (smallest ratio: R1 is the pivot row)
2 s2 4/1 =4

Thus, variable x2 has to enter the basic variable to replace s1 by pivoting entry a12 =3 to be 1
and zero entries in the pivot column using elementary row operations to obtain the second
tableau.

Second Tableau

x1 x2 s1 s2 F Operation
1
2/3 1 1/3 0 0 3 R1 → R1
3
( 1/3 0 −1/3 1 0| 1 ) − R1 + R2 → R2
⋯ ⋯ ⋯ ⋯ ⋯ ⋯
−2/3 0 5/3 0 1 15 5R1 + R3 → R3

16
The basic variables are now x2 and s2, and the associated feasible solution is x1 =0, x2 =3, s1
= 0, s2 =1, and the objective function F = 15. The presence of the negative indicator -2/3
suggests that the solution is still not optimal, because we can increase the value of F by
increasing x1 from its present value of zero. Therefore, column 1 is the pivot column.

Minimum ratio table

Row No Basic variable Ratio
1 x2 3/2/3 =9/2
2 s2 1/1/3 =3 (smallest ratio: R2 is the pivot row)

Thus, variable x1 has to enter the basic variable to replace s2 by pivoting entry a22 =1/3 to be
1 and zero entries in the pivot column using elementary row operations to obtain the third
tableau.

Third Tableau

x1 x2 s1 s2 F Operation
3R2 → R2
0 1 1 −2 0 1
1 0 −1 3 0 3 −2
(⋯ ⋯ ⋯ ⋯ ⋯ |⋯) R2 + R1 → R1
3
0 0 1 2 1 17
2
R2 + R3 → R3
3
The basic variables are now x1 and x2, and the associated feasible solution is x1 =3, x2 =1, s1
= 0, s2 =0, and the objective function F = 17. The corresponding corner point is (3, 1).

Example-2

Maximize the function F = 2 x1 + x2 + 3x3

x1 + 2 x2 + x3  12
2 x1 + x2 + x3  20
Subject to
x1 + x2 + 2 x3  20
x1 , x2 , x3  0
Solution
First introduce slack variables s1 and s2 and write inequalities in equational form:
x1 + 2 x2 + x3 + u1 = 12
2 x1 + x2 + x3 + u2 = 20
x1 + x2 + 2 x3 + u3 = 20
x1 , x2 , x3 , u1 , u2 , u3  0
Equational of objective function F = 2 x1 + x2 + 3x3 is written as −2 x1 − x2 − 3x3 + F = 0

17
 Initial table
x1 x2 x3 u1 u2 u3 F
1 2 1 1 0 0 0 12
2 1 1 0 1 0 0 20
1 1 2 0 0 1 0 |20
|
⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯
(−2 −1 −3 0 0 0 1 0)

The basic variables associated with the initial tableau are u1, u2 and u3 and the associated
feasible solution is x1 =0, x2 =0, x3=0, u1= 12, u2 =20, and u3 =20

The most negative indicator is -3, therefore, column three of variable x3 is a pivot column. To
find a pivot row use minimum ratio rule:

Minimum ratio table

Row No Basic variable Ratio
1 u1 12/1 =12
2 u2 20/1 = 20
3 u3 20/2 = 10 (smallest ratio: R3 is the pivot row)

Thus, variable x3 has to enter the basic variable to replace u3 by pivoting entry a33 =2 to be 1
and zero entries in the pivot column using elementary row operations to obtain the second
tableau.

Second Tableau

x1 x2 x3 u1 u2 u3 F Operations
1
1/2 3/2 0 1 0 −1/2 0 2 R3 → R3
3/2 1/2 0 0 1 −1/2 0 2
10
1/2 1/2 1 0 0 1/2 0
|10 − R3 + R1 → R1
|
⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ − R3 + R2 → R2
(−1/2 1/2 0 0 0 3/2 1 30)
3R3 + R4 → R4

The basic variables are now x3, u1, and u2 and the associated feasible solution is x1 =0, x2
=0, x3=10, u1= 2, u2 =10, and u3 =0, and the objective function F = 30. The presence of the
negative indicator -1/2 suggests that the solution is still not optimal, because we can increase
the value of F by increasing x1 from its present value of zero. Therefore, column 1 is the
pivot column. Minimum ratio table to find pivot row:

Minimum ratio table

Row No Basic variable Ratio
1 u1 2/1/2 =4 (smallest ratio: R1 is the pivot row)
2 u2 10/3/2 = 20/3
3 x3 10/1/2 = 20

18
Thus, variable x1 has to enter the basic variable to replace u1 by pivoting entry a11 =1/2 to be
1 and zero entries in the pivot column using elementary row operations to obtain the third
tableau.

Third Tableau
x1 x2 x3 u1 u2 u3 F Operations
2 R1 → R1
1 3 0 2 0 −1 0 4 3
− R1 + R2 → R2
0 −4 0 −3 1 1 0 4 2
0 −1 1 −1 0 1 0 |8 1
| − R1 + R3 → R3
⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ 2
(0 2 0 1 0 1 1 32)
1
R1 + R4 → R4
2

Since all indicators in the third tableau are non-negative, this table is terminal. The basic
variables are now x1, x3 and u2, and the associated feasible solution is x1 =4, x2 =0, x3=8,
u1= 0, u2 =4, and u3 =0, and the objective function F = 17. Thu the required maximum value
of F occurs at the corner point (4, 0, 8) and the optimal value of the objective function F =32.

Exercise
In problems 1 through 5, solve the given linear programming in two variables by the simplex
method. Sketch the feasibility regions and indicate the corner points and state the optimal
solution of maximization objective function F
x1 + x2  6
1. Maximize the function F = x1 − 2 x2 Subject to 2 x1 + 3x2  15
x1 , x2  0

4 x1 + 3x2  24
2. Maximize the function F = 5 x1 + x2 Subject to x1 + 3x2  15
x1 , x2  0

x1 + x2  8
x1 + 2 x2  11
3. Maximize the function F = 20 x1 + 30 x2 Subject to
x1 + 3 x2  12
x1 , x2  0

19
 3x1 + 4 x2  25
x1 + x2  7
4. Maximize the function F = x1 + 2 x2 Subject to
x1 + 2 x2  9
x1 , x2  0

4 x1 + 5 x2  37
x1  8
5. Maximize the function F = 11x1 − 4 x2 Subject to
x2  5
x1 , x2  0

In problems 6 through 10, solve the given linear programming in by the simplex method.

2 x1 + 3x2 + x3  80
6. Maximize the function F = x1 + x2 + 2 x3 Subject to x1 + x2 + x3  60
x1 , x2 , x3  0

x1 + 2 x2 + x3  20
5 x1 + x2 + 2 x3  50
7. Maximize the function F = 2 x1 + 3x2 + 4 x3 Subject to
− x1 + 3x2 + x3  30
x1 , x2 , x3  0

7 x1 + 12 x2 + 6 x3  50
18 x1 + 9 x2 + 4 x3  85
8. Maximize the function F = 5 x1 + 42 x2 + 26 x3 Subject to
−2 x1 + 14 x2 + x3  66
x1 , x2 , x3  0

4 x1 + 5 x3 + 6 x4  20
9. Maximize the function F = 4 x1 + 2 x2 + x3 − x4 Subject to x1 + x2 + 2 x3 + 4 x4  20
x1 , x2 , x3 , x4  0

− x1 + x2 + x3 + 3x4  20
3x1 − x2 + 3 x3 + 6 x4  30
10. Maximize the function F = 3x1 − x2 + 2 x3 − x4 Subject to
x1 + x2 + x3 − 3x4  15
x1 , x2 , x3 , x4  0

11. A manufacturer produces three types of electrical bulbs. The time required for molding,
trimming, and packaging is given in the Table below (times are given in hours per dozen of
bulbs).

20
 Process Type A Type B Type C Total time available in hours
Molding 1 2 3/2 12,000
Trimming 2/3 2/3 1 4,600
Packaging 1/2 1/3 1/2 2,400
Profit in $ 11 16 15 -
How many dozen of each type of bulbs should be produced to obtain a maximum profit?
[Answers: Type A: 600 dozen units, Type B: 5,100 dozen units and Type C: 800 dozen units]
12. A company makes three kinds of machines; I, II and III. There are three stages in the
production of each machine: S1, S2 and S3. The number of worker-hours expended on each
machine in each stage of production is given in the following table.

Machines Stages
S1 S2 S3
I 0.2 0.1 0.2
II 0.2 0.2 0.3
III 0.1 0 0.1
Suppose the company makes $10 on each type I machine, $15 on each type II, and $6 on
each type III. The company has available a total of 14 worker-hours per day for stage S1, 8
worker-hours per day for stage S2 and 20 worker-hours per day for stage S3. How many
machines of each type should be made to maximize total profit?

THE SIMPLEX METHOD FOR MINIMIZATION PROBLEM

So far, we have applied the simplex method only to linear programming problems in standard
form where the objective function was to be maximized, all inequalities are of the ≤ type, and
all right-side constants are nonnegative. In this section, we extend this procedure to linear
programming problems in which the objective function is to be minimized. A minimization
problem is in standard form if the objective function F is to be minimized; such as

Minimize the objective function F = c1 x1 + c2 x2 + ... + cn xn

Subject to constraints of the form of ≥ type:

a11 x1 + a12 x2 + ... + a1n xn  b1

a21 x1 + a22 x2 + ... + a2 n xn  b2
or
am1 x1 + am 2 x2 + ... + amn xn  bm
x1 , x2 ,..., xn  0
Subject to constraints of the form of ≥ and ≤ types:

21
a11 x1 + a12 x2 + ... + a1n xn  b1
a21 x1 + a22 x2 + ... + a2 n xn  b2

am1 x1 + am 2 x2 + ... + amn xn  bm

x1 , x2 ,..., xn  0
Where, bj  0 for j = 1, 2,..., m.

I: The Simplex Method for Minimization Problem to constraints of the form of ≥ and ≤
Types
Step 1: Change the objective function of minimize F = c1 x1 + c2 x2 + ... + cn xn to maximize
G = − F = −c1 x1 − c2 x2 − ... − cn xn
For instance, minimize F = 2 x1 + 5 x2 instead we maximize the function G = − F = −2 x1 − x2

Step 2: Convert the constraints of the ≥ type into ≤ type by multiplying by -1. For instance,
3x1 − 2 x2 + 7 x3  4 is equivalent to −3x1 + 2 x2 − 7 x3  −4
Step 3: Solving linear programming with one or more negative right-side constants. Two
phases are involved in eliminating the negative right-side constants.

Phase I: A Procedure for Finding a Corner Point

First goal is to transform the initial table into a form that corresponds to a basic feasible
solution. In practice, this is done by selecting a row in the initial table with a negative entry
in the right-most column and pivoting on the negative in that row. This procedure is
repeated as often as necessary to create a table in which the right-most column contains only
nonnegative entries.

Phase II: Solving Linear Programming After Completing Phase I

After completing Phase I procedure, apply the standard simplex method to find an optimal
solution.

Example 1:
x1 + x2  9
x1 + x2  3
Minimize F = 3x1 − 5 x2 Subject to
x1 + 2 x2  5
x1 , x2  0
Solution: The given problem is equivalent to:

22
 x1 + x2  9
− x1 − x2  −3
Maximize G = − F = −3x1 + 5 x2 Subject to
− x1 − 2 x2  −5
x1 , x2  0

Phase I: Initial table is

 x1 x2 s1 s2 s3 constant 
 1 1 1 0 0 9 
 
-1 -1 0 1 0 -3 
 
-1 -2 0 0 1 -5 
 
 
3 -5 0 0 0 0 
We want to end up with positive terms in the last (right-most) column. The pivot entry is
chosen arbitrarily, but a reasonable strategy is to pivot on a negative entry in a row with a
negative right-side entry, say the -1 in the (2, 2) position. The result is displayed in second
table:

Second Table
 x1 x2 s1 s2 s3 constant 
− R2 → R2  0 0 1 1 0 6 
 
− R2 + R1 → R1  1 1 0 −1 0 3 
 
2 R2 + R3 → R3  1 0 0 −2 1 1 
5R2 + R4 → R4  
 
 8 0 0 -5 0 15 
Since the second table has all positive term in the last column, the corresponding basic
feasible solution is x1 =0, x2 = 3, S1 =6, S2 =0, S3 =1.

Phase II: Now we have a table with a basic feasible solution, we proceed with a standard
simplex method. We find that column 4 is the pivot column (the column with the most
negative indicator) and row 1 is the pivot row, thus 1 is the pivot entry in (1, 4) position.
Third table is

Third Table

23
  x1 x2 s1 s2 s3 constant 
 0 0 1 1 0 6 
R1 + R2 → R2  
 1 1 1 0 0 9 
2 R1 + R3 → R3  
 1 0 2 0 1 13 
5R1 + R4 → R4  
 
 8 0 5 0 0 45 
Since the third table has no negative indicators, it is terminal and the corresponding basic
solution x1 =0, x2 = 9, S1 =0, S2 =6, S3 =13. The optimal value of Gmax =45, and it occurs at
the corner point (0, 9). The Fmin = -45.

Example 2:
3x1 + 2 x2 + 8 x3  144
Minimize F = x1 + 2 x2 + 4 x3 Subject to 3x1 + x2 + 6 x3  120
x1 , x2 , x3  0
Solution: The given problem is equivalent to:
−3x1 − 2 x2 − 8 x3  −144
Maximize G = − F = − x1 − 2 x2 − 4 x3 Subject to 3x1 + x2 + 6 x3  120
x1 , x2 , x3  0

Phase I: Initial table is

 x1 x2 x3 s1 s2 constant 
-3 -2 -8 1 0 -144 
 
 3 1 6 0 1 120 
 
 
 1 2 4 0 0 0 
Since there is a negative right-side constant in row 1, pivot arbitrarily on a negative entry in
that row with a negative right-side entry, say the -2 in the (1, 2) position. The Second Table
1  x1 x2 x3 s1 s2 constant 
− R1 → R1 3/2 1 
2  4 1/ 2 0 72 
− R1 + R2 → R2  3 / 2 0 2 1/ 2 1 48 
 
−2 R1 + R3 → R3  
 −2 0 − 4 1 0 − 144 
Since the second table has all positive term in the last column, the corresponding basic
feasible solution is x1 =0, x2 = 72, x3 =0, S1 =0, S2 =48.

Phase II: Now we have a table with a basic feasible solution, we proceed with a standard
simplex method. We find that column 3 is the pivot column (the column with the most

24
negative indicator). Forming ratios with the positive entries in the pivot column [72/4 =18,
48/2 =24], thus row 1 is the pivot row (smallest ratio =18), and 4 is the pivot entry in (1, 3)
position. Third table is displayed below:

Third table:

1  x1 x2 x3 s1 s2 constant 
R1 → R1 3/8 1/4 
4  1 − 1/ 8 0 18 
−2 R1 + R2 → R2  3 / 4 − 1/ 4 0 3 / 4 1 12 
 
4 R1 + R3 → R3  
 −1/ 2 1 0 1 0 − 72 
The third table is not the terminal since there is a negative indicator -1/2 in column 1.
Forming ratios with the positive entries in the pivot column [18/(3/8) =48, 12/(3/4) =16], thus
row 2 is the pivot row (smallest ratio =16), and 3/4 is the pivot entry in (2, 1) position. Fourth
table is displayed below:

Fourth table:

4  x1 x2 x3 s1 s2 constant 
R2 → R2
3  0 1/2 1 − 1/ 2 − 1/ 2 12 
3 
− R2 + R1 → R1  1 −2/3 0 1 4 / 3 16 
8  
1  
R2 + R3 → R3  0 2/3 0 1 2/3 − 64 
2
This fourth table is terminal. The corresponding basic feasible solution x1 =16, x2 = 0, x3 =12,
S1 =0, S2 =0. The optimal value of Gmax = -64, the required Fmin = 64 and it occurs at the
corner point (16, 0, 12).

II: The Simplex Method for Minimization Problem to constraints of both =, ≥ and ≤
Types
Equality constraint is written into two inequalities of ≤ and ≥ types. Then we apply Phase I
and Phase II to solve minimum linear programming.

Example 1:
5 x1 + 4 x2  40
3x1 + 2 x2  48
Minimize F = 3x1 − 7 x2 Subject to
4 x1 + 5 x2 = 80
x1 , x2 , x3  0

25
Solution: Writing the equality constraint 4 x1 + 5 x2 = 80 as 4 x1 + 5 x2  80 and 4 x1 + 5 x2  80 ,
the problem becomes
5 x1 + 4 x2  40
3 x1 + 2 x2  48
Minimize F = 3x1 − 7 x2 Subject to 4 x1 + 5 x2  80 equivalent to
4 x1 + 5 x2  80
x1 , x2  0

−5 x1 − 4 x2  −40
3 x1 + 2 x2  48
Maximize G = − F = −3x1 + 7 x2 Subject to 4 x1 + 5 x2  80
−4 x1 − 5 x2  −80
x1 , x2 , x3  0

Phase I: Initial table is

 x1 x2 s1 s2 s3 s4 constant 
 -5 -4 1 0 0 0 -40 
 
 3 2 0 1 0 0 48 
 
 4 5 0 0 1 0 80 
 -4 -5 0 0 0 1 -80 
 
 
3 
 -7 0 0 0 0 0 
Arbitrarily, we chose the number -4 in the (1,2) position as the pivot entry in that row with a
negative right-side entry, say the -2 in the (1, 2) position. The result is displayed in second
table:

Second Table

26
  x1 x2 s1 s2 s3 s4 constant 
5 1 
 1 - 0 0 0 10 
1  4 4 
− R1 → R1 1 1 
4  0 1 0 0 28 
−2 R1 + R2 → R2 2 2 
 9 5 
−5 R1 + R3 → R3 - 0 0 1 0 30 
5 R1 + R4 → R4  4 4 
9 5 
7 R1 + R5 → R5 4 0 - 0 0 1 -30 
4
 
 
 47 7 
 0 - 0 0 0 70 
 4 4 
Second table still has a negative right-side constant (-30) in the fourth row, so we remain in
Phase I and chose a second arbitrary pivot. This time we pick the entry -5/4 in the (4, 3)
position, and by pivoting on this entry we obtain the third table:

Third Table
 x1 x2 s1 s2 s3 s4 constant 
4
− R4 → R4  4 1 
5  1 0 0 0 - 16 
1  5 5 
R4 + R1 → R1 7 2 
4  0 0 1 0 16 
1  5 5 
− R4 + R2 → R2  0 0 0 0 1 1 0 
2  
5  -
9
0 1 0 0 -
4
24 
− R4 + R3 → R3  4 5 
4  
7  
R4 + R5 → R5  43 7 
4  0 0 0 0 - 112 
 5 5 
Since the third table has all nonnegative right-side constants, we are ready to enter Phase II.

Phase II: The only one indicator is the -7/5 in the s4 column (pivot column) and the
nonnegative ratios are 16/(2/5) =40, and 0/1 =0 (minimum ratio); thus, row 3 is the pivot row
and 1 is the pivot entry in the (3,6) position. The fourth table is displayed below:

27
 Fourth Table
 x1 x2 s1 s2 s3 s4 constant 
 4 1 
1  1 0 0 0 16 
R3 + R1 → R1 5 5 
5 7 
2
2  0 0 1 - 0 16 
− R3 + R2 → R2 5 5 
5
 0 0 0 0 1 1 0 
4  
R3 + R4 → R4  -9 0 4 
5 1 0 0 24
 4 5 
7
R3 + R5 → R5  
5  
 43 7 
 0 0 0 0 112 
 5 5 

Since all the indicators are nonnegative, the fourth table is terminal. The corresponding basic
feasible solution x1 =0, x2 = 16, S1 =24, S2 =16, S3 =0, S4 =0. The optimal value of Gmax =
112, the required Fmin = -112 and it occurs at the corner point (0, 16).

III: The Simplex Method for Minimization Problem to constraints of ≥ Type:

The Dual Formulation of a Linear Programming Problem

Just like a photograph can be analysed in both its positive and negative form, a linear
programming has both a primal formulation and a dual formulation, either of which can be
used to obtain an optimal solution. Primal formulation involving a minimization problem
with ≥ inequality constraints can be solved by analysing the dual formulation involving a
maximization problem with ≤ inequality constraints. Thus, minimization problem is referred
to primal formulation. Similarly, the objective function associated with this formulation is
called the primal the objective function and the inequality constraints are called primal
inequalities.

A. Dual Formulation of a Minimization Linear Programming Problem

Steps of changing primal formulation (minimization problem) to dual formulation. For
example;

Minimize F = 7 x1 + 20 x2
x1 + 2 x2  2
Subject to x1 + 5 x2  3
x1 , x2  0

28
Step I: Use the coefficients and right-side constants of the primal inequalities and the
coefficients of the primal objective function to form a matrix, say A.

1 2 2 
A = 1 5 3  ,
7 20 0 
where, row 1 and row 2 are the coefficients of primal inequalities and row 3 are coefficients
for primal objective function.

Step II: Form matrix B which is a transpose of matrix A (i.e., interchanging the rows and
columns of matrix A).

1 1 7 
B = A =  2 5
T
20
 2 3 0 

Step III: Reverse Step I to obtain a maximization problem with ≤ inequalities from matrix B.
This is the dual formulation. Specifically, the first two rows of matrix B are used for the dual
inequalities and the third for the dual objective function. Therefore,
Dual Formulation is:
Maximize G = 2 y1 + 3 y2
y1 + y2  7
Subject to 2 y1 + 5 y2  20
y1 , y2  0
Note: In order to avoid any confusion between the two formulations, we use different
decision variables. If primal variables are x1, x2, …, xn, then the dual variables become y1, y2,
…, ym

B. Solution of Minimization Linear Programming Problem by Duality

The Duality Theorem:
Suppose either a given primal minimization problem or the corresponding dual maximization
has an optimal solution. Then the other problem also has an optimal solution, and the
objective values of the two optimal solutions are equal.

Example 1: Use duality and the simplex method to solve following linear programming:
x1 + 2 x2  2
Minimize F = 7 x1 + 20 x2 Subject to x1 + 5 x2  3
x1 , x2  0
Solution:
The corresponding dual problem is:

29
 y1 + y2  7
Maximize G = 2 y1 + 3 y2 Subject to 2 y1 + 5 y2  20
y1 , y2  0

Initial Table for the Dual Problem:

 y1 y2 s1 s2 constant 
1 1 1 0 7 
 
2 5 0 1 20  . Column 2 and row 2 are pivot column and row,
 
 
 −2 − 3 0 0 0 
respectively. The pivot entry is 5 in (2, 2) position. Pivoting on this entry we obtain the
second table of the dual problem.

Second Table for the Dual Problem:

 y1 y2 s1 s2 constant 
3 1 
1  0 1 − 3 
R2 → R2 5 5 
5  2 
1
− R2 + R1 → R1  1 0 4  . This is not the terminal table.
5 5 
3R2 + R3 → R3  
 
− 4 0 0
3
12 
 5 5 

Third Table for the Dual Problem:

 y1 y2 s1 s2 constant 
5  5 1 
R1 → R1  1 0 − 5 
3  3 3 
2  2 1 
− R1 + R2 → R2 0 1 − 2  This is the terminal table.
5  3 3 
4  
R1 + R3 → R3  
5 0 0
4 1
16 
 3 3 
The terminal table has optimal solution Gmax = 16, which occurs at the corner point y1 =5 and
y2 =2. When we use geometric approach to solve primal problem, we find x1 =4/3, x2 = 1/3.
But the numbers 4/3 and 1/3 appear on the indicator line of the terminal table of the dual
problem, under the slack variables. These indicators are called terminal slack indicators of
the dual problem.

30
The use of Duality to Solve Minimization Problem with ≥ Constraints

Step 1: Obtain dual formulation of the given primal problem. This will be a maximization
problem with ≤ constraints.
Step 2: Introduce slack variables in the dual problem and solve by simplex method.
Step 3: Read the optimal solution to the primal (minimization) problem from the terminal
table of the dual. The required minimum occurs in the right-hand corner, and the
components of the corresponding corner point occur as terminal slack indicators.

Example: Suppose that in solving a particular minimization problem, we apply the simplex
method to the initial table of the dual form of the problem (maximization problem) to
reach its terminal table. The table below is the terminal simplex table of the dual
problems.
 y1 y2 y3 s1 s2 constant 
 2 5 
4 1 0 23 
 3 2 
 4 2 
 −2 0 1 − 11 
 3 3 
 
 
 5 0 0 3 7 200 
 
(a) What is the optimal solution of the dual (maximization) problem?
(b) What is the optimal solution of the primal (minimization) problem?

Solution: (a) The basic feasible solution that corresponds to the given table is y1 =0, y2 =23,
y3 =11, s1 =0, s2 =0. The maximum value of the dual objective function G, Gmax =200, and it
occurs at the corner point (0, 23, 11).
(b) The minimum of the primal objective function F is Fmin = 200, the terminal slack
indicators are at the point where x1 =3, x2 =7; that is the corner point (3, 7).

Exercise
1. Use simplex method to solve the minimum linear programming problems:
(a) Minimize F = 2 x1 − 7 x2 + 5 x3 (c) Minimize F = x1 + 5 x2 + 4 x3

4 x1 + 3 x2 + x3  125 5 x1 + 12 x2 + 8 x3  96
10 x1 + 5 x2 + 2 x3  295 5 x1 + 7 x2 + 5 x3  75
Subject to Subject to
x1 + x2 + 6 x3  90 −3x1 + 4 x2 + 4 x3 = 4
x1 , x2 , x3  0 x1 , x2 , x3  0

31
 (b) Minimize F = 3x1 + 2 x2
− x1 + x2  1
Subject to 3x1 + 2 x2  12
x1 , x2  0
2. State the dual problem of the given minimization problem. Hence, solve dual problem
using simplex method to determine the solution of the given primal problem:
(a) Minimize F = 5 x1 + 13x2 (d) Minimize F = 5 x1 − 13x2

x1 + 5 x2  15 x1 + 5 x2  15
2 x1 + 4 x2  24 2 x1 + 4 x2  24
Subject to Subject to
x1 + x2  7 x1 + x2  7
x1 , x2  0 x1 , x2  0

(b) Minimize F = 12 x1 + 26 x2 + 14 x3 (e) Minimize F = 4 x1 + 10 x2 + 9 x3

2 x1 + 4 x2 + x3  3 x1 + 3x2 + x3  6
Subject to x1 + 3x2 + 2 x3  4 Subject to x1 + x2 + 3x3  9
x1 , x2 , x3  0 x1 , x2 , x3  0

(c) Minimize F = 9 x1 + 8 x2 + 12 x3 (f) Minimize F = 2 x1 + 3x2 + 4 x3

3x1 + 4 x2 + 6 x3  2 x1 + 4 x2 + 2 x3  8
−2 x1 − 5 x2 + x3  3 x1 + 3x2 + 4 x3  7
Subject to Subject to
9 x1 + x2 + 2 x3  8 2 x1 + x2 + 3x3  6
x1 , x2 , x3  0 x1 , x2 , x3  0

3. Suppose that in solving a particular minimization problem, we apply the simplex method
to the initial table of the dual form of the problem (maximization problem) to reach its
terminal table. Tables below are the terminal simplex table of the dual problems. In each
case find the optimal solution of the dual problem and the optimal solution of the
corresponding primal (minimization) problem.

32
  y1 y2 y3 s1 s2 constant 
3 1 0 5 2 15 
 
(a)  − 1 0 1 −1 1 9  (b)
 
 
 2 0 0 1 3 173 
 y1 y2 y3 s1 s2 s3 constant 
1 -3 0 0 7 9 21 
 
0 1 0 1 −3 5 30 
 
0 4 1 0 8 1 15 
 
 
 0 2 0 0 4 3 110 

 y1 y2 s1 s2 s3s4 constant 
0 0 1 0 3 0 11 
 
1 0 0 0 − 1 1 15 
 
(c)  1 1 0 0 −2 0 7  (d)
0 0 0 1 5 0 3 
 
 
 2 7 0 124 
 0 0 0
 y1 y2 s1 s3 constant 
s2
 3 0 1 − 3 0 17 
 
 −1 0 0 4 1 21 
 
 2 1 0 7 0 9 
 
 
 4 0 0 9 0 83 

33