LINEAR

PROGRAMMING
Chapter 4
Outcomes
Formulate the objective function (focus for LP)
Perform the graphical method and find solution as well as formulate
complex LP problems.
Solve using simplex method and perform sensitivity analysis
Formulation of the Objective
Function
The formulation of objective function is one of the crucial steps in the
application of optimization to a practical problem.
You must be able to translate a verbal statement or concept of the desired
objective into mathematical terms.
In the chemical industries, the objective function often is expressed in
units of currency because the goal is to minimize costs or maximize profits
subject to a variety of constraints.
Other examples include:
Maximization of the yield of a component in a reactor
Minimization of the use of utilities in a heat exchanger network
Minimization of the volume of a packed column
Minimizing the differences between a model and some data
When formulating the mathematical statement of the objective, functions
that are more complex or nonlinear are more difficult to solve in
optimization use optimization software instead.
Introduction about LP
The goal of linear programming (LP) is to determine the values of decision variables
that maximize or minimize a linear objective function, where the decision variables
are subjected to linear constraints.
Is a special case of general constrained optimization problem.
In general setting, we want to find a point (a value of decision variable) that
minimizes/maximizes the objective function, simultaneously satisfies the constraints.
This point is a feasible point.
In LP, the objective function is linear, and the set of feasible points is determined by
a set of linear equations and/or equalities.
The methods in LP provide a way of choosing the best feasible point among the
many possible feasible points.
Is one of the most widely used optimization techniques and perhaps the most
effective.
Historically, Leonid Kantorovich in 1939 had developed the method as a way to plan
expenditures and returns, as to reduce costs to the army and increase losses
incurred by the enemy (WWII).
Forms

The
basic form of linear programming (LP) is;
min/max f(x)
subject to:
h(x) = 0
g(x) 0
Where x is assumed as a set of continuous variables. The functions of f(x), g(x)
and h(x) are must all linear.
Also, can be written as;
min/max f =
subject to:
= , i = 1,2,,m, j = 1,2,,n
In a standard (matrix) form;
min/max cTx
subject to:
Ax = b and/or Ax b and/or Ax b
x0
Where c Rn, b Rm and A Rmxn.
Let us start with few examples
We have done few examples on how to
formulate optimization problems involving
LP. Now let us learn on how to solve them..

Graphical method
Simplex method
 Consider the following
Giapettos Incorporated manufactures wooden soldiers and trains. The data for these
manufacturing are given below:
Each soldier built:
Sold for $27 and uses $10 worth of raw materials.
Increase Giapettos variable labor/overhead costs by $14.
Requires 2 hours of finishing labor.
Requires 1 hour of carpentry labor.
Each train built:
Sold for $21 and uses $9 worth of raw materials.
Increases Giapettos variable labor/overhead costs by $10.
Requires 1 hour of finishing labor.
Requires 1 hour of carpentry labor.
Each week Giapetto can obtain:
All needed raw materials.
Only 100 finishing hours.
Only 80 carpentry hours.
Also:
Demand for the trains is unlimited.
At most 40 soldiers are bought each week.
Giapetto wants to maximize weekly profit (revenues expenses). Develop a mathematical
model for Giapettos situation that can be used to maximize the weekly profit and SOLVE it by
using graphical method.
Cont..
For the Giapetto problem model, combining the sign restrictions x1 0 and x2
0 with the objective function and constraints yields the following optimization
model:
Max z = 3x1 + 2x2 (objective function)
Subject to (s.t.):
2 x1 + x2 100 (finishing constraint)
x1 + x2 80 (carpentry constraint)
x1 40 (constraint on demand for soldiers)
x1 0 (sign restriction)
x2 0 (sign restriction)
3.2 Graphical Solution to a 2-Variable LP
Finding the Feasible Solution
Since the Giapetto LP has two variables, it may be solved
graphically. The feasible region is the set of all points satisfying the
constraints:
Giapetto Constraints
2 x1 + x2 100 (finishing constraint)
x1 + x2 80 (carpentry constraint)
x1 40 (demand constraint)
x1 0 (sign restriction)
x2 0 (sign restriction)

A graph of the constraints and feasible region is shown on the next

slide.
 3.2 Graphical Solution to a 2-Variable LP
X2
From figure, we B
see that the set
finishing constraint Feasible Region
of points
satisfying the D
Giapetto LP is
bounded by the demand constraint

five sided
polygon DGFEH. G

Any point on or
in the interior of z = 100
this polygon carpentry constraint

(the shade
area) is in the
feasible region. F
z = 180
z = 60
E A C
H
10 20 40 50 60 80 X1
 3.2 Graphical Solution to a 2-Variable LP

Having identified the feasible region

for the Giapetto LP, a search can
begin for the optimal solution which
will be the point in the feasible
region with the largest z-value.
 3.2 Graphical Solution to a 2-Variable LP
X2
To find the optimal B

solution, graph a finishing constraint Feasible Region

line on which the
points have the D
same z-value. In a
demand constraint
max problem,
such a line is
G
called an
isoprofit line
z = 100
while in a min
carpentry constraint
problem, this is
called the isocost
line. The figure
shows the isoprofit F
z = 180
lines for z = 60, z z = 60
= 100, and z = E A C
H
180 10 20 40 50 60 80 X1
 3.2 Graphical Solution to a 2-Variable LP
X2
The last isoprofit B

intersecting finishing constraint Feasible Region

(touching) the
feasible region D
indicates the
demand constraint
optimal solution
for the LP. For
G
the Giapetto
problem, this
z = 100
occurs at point G
carpentry constraint
(x1 = 20, x2 = 60,
z = 180).
F
z = 180
z = 60
E A C
H
10 20 40 50 60 80 X1
 3.1 - What Is a Linear Programming Problem?
For a maximization problem, an optimal solution to an LP is a point
in the feasible region with the largest objective function value.
Similarly, for a minimization problem, an optimal solution is a point
in the feasible region with the smallest objective function value.

Most LPs have only one optimal solution. However, some LPs
have no optimal solution, and some LPs have an infinite number
of solutions. Section 3.2 shows the optimal solution to the
Giapetto LP is x1 = 20 and x2 = 60. This solution yields an
objective function value of:

z = 3x1 + 2x2 = 3(20) + 2(60) =

$180
When we say x1 = 20 and x2 = 60 is the optimal solution, we are
saying that no point in the feasible region has an objective
function value (profit) exceeding 180.
Simplex Method and Conversion
LP into Standard Form
Before the simplex algorithm can be used to solve an LP, the LP must be
converted into a problem where all the constraints are equations and all
variables are nonnegative. An LP in this form is said to be in standard form.
Consider the following example.
Leather Limited manufactures two types of leather belts: the deluxe
model and the regular model. Each type requires 1 square yard of
leather. A regular belt requires 1 hour of skilled labor and a deluxe belt
requires 2 hours of skilled labor. Each week, 40 square yards of leather
and 60 hours of skilled labor are available. Each regular belt contributes
$3 profit and each deluxe belt $4. Write an LP to maximize profit.
The decision variables are:
x1 = number of deluxe belts produced weekly
x2 = number of regular belts produced weekly
Cont..
the appropriate LP is:
max z = 4x1 + 3x2
s.t. : x1 + x2 40 (leather constraint)
2x1 +x2 60 (labor constraint)
x 1, x 2 0
To convert the leather and labor () constraints to equalities, we
define for each constraint a slack variable si (si = slack variable for the
i th constraint). A slack variable is the amount of the resource unused
in the i th constraint.
 The LP not in standard form is:

max z = 4x1 + 3x2

s.t. : x1 + x2 40 (leather constraint)
2x1 + x2 60 (labor constraint)
x1, x2 0

The same LP in standard form is:

max z = 4x1 + 3x2

s.t. x1 + x2 + s1 = 40
2x1 + x2 + s2 = 60
x1, x2, s1, s2 0

In summary, if a constraint i of an LP is a constraint, convert it to an equality constraint

by adding a slack variable si to the i th constraint and adding the sign restriction s i 0.
 A constraint can be converted to an equality constraint. Consider the
formulation below:

min z = 50 x1 + 20x2 + 30X2 + 80 x4

s.t. 400x1 + 200x2 + 150 x3 + 500x4 500

3x1 + 2x2 6
2x1 + 2x2 + 4x3 + 4x4 10
2x1 + 4x2 + x3 + 5x4 8
x1, x2, x3, x4 0

To convert the i th constraint to an equality constraint, define an

excess variable (sometimes called a surplus variable) e i (ei will
always be the excess variable for the i th constraint. We define ei
to be the amount by which i th constraint is over satisfied.
Transforming the LP on the previous slide to standard form yields:

min z = 50 x1 + 20x2 + 30X2 + 80 x4

s.t.
400x1 + 200x2 + 150 x3 + 500x4 e1 = 500
3x1 + 2x2 - e2 =6
2x1 + 2x2 + 4x3 + 4x4 - e3 = 10
2x1 + 4x2 + x3 + 5x4 - e4 = 8
xi, ei > 0 (i = 1,2,3,4)

In summary, if the i th constraint of an LP is a constraint, it

can be converted to an equality constraint by subtracting the
excess variable ei from the i th constraint and adding the sign
restriction ei 0.
If an LP has both and constraints, apply the previous procedures
to the individual constraints. Consider the example below.

Nonstandard Form Standard Form

max z = 20x1 + 15x2 max z = 20x1 + 15x2

s.t. x1 100 s.t. x1 + s1 = 100

x2 100 x2 + s2 = 100

50x1 + 35x2 6000 50x1 + 35x2 + s3 = 6000

20x1 + 15x2 2000 20x1 + 15x2 - e4 = 2000

x1, x2 > 0 x1, x2, s1, s2, s3, e4 > 0

Simplex Method
Suppose an LP with m constraints and n variables has been converted
into standard form. The form of such an LP is:

max ( or min) z = c1x1 + c2x2 + +cnxn

s.t. a11x1 + a12x2 + + a1nxn =b1

a21x1 + a22x2 + + a2nxn =b2
. . .
. . .
am1x1 + am2x2 + + amnxn =bm
xi 0 ( i = 1,2, , n)
Cont..
a11 a12 .... a1n x1 b1

a21 a22 .... a2n x2 b2
If we define: A x b
.... .... .... .... .... ....
a a .... a x b
m1 m2 mn n m

The constraints may be written as a system of equations Ax = b.

Consider a system Ax = b of m linear equations in n variables

(where n m).

A basic solution to Ax = b is obtained by setting (n m) variables

equal to 0 and solving for the remaining m variables. This
assumes that setting the (n m) variables equal to 0 yields a
unique value for the remaining m variables, or equivalently, the
columns for the remaining m variables are linearly independent.
 To find a basic solution to Ax = b, we choose a set of (n m) variables
(the nonbasic variables, or NBV) and set each of these variables equal
to 0. Then we solve for the values of the n (n m) = m variables (the
basic variables, or BV) that satisfy Ax = b.

Different choices of nonbasic variables will lead

to different basic solutions. Consider the basic x1 + x2 =3
solutions to the system of 2 equations shown to - x2 + x3 = -1
the right:

The number of non-basic variables = 3 2 =

1. Setting, for example, NBV = {x3} (as shown x1 + x2 =3
to the right, then BV = {x1, x2}. We can obtain
- x2 + 0 = -1
the values for these basic variables by setting
x3 = 0. Solving we find x1 = 2, x2 = 1. Thus, x1
= 2, x2 = 1, and x3 = 0 is a basic solution.
If NBV = {x1} and BV = {x2, x3} are chosen instead, the basic solution
becomes x1 = 0, x2 = 3, and x3 = 2.

If NBV = {x2} and BV = {x1, x3} are chosen instead, the basic solution
becomes x1 = 3, x2 = 0, and x3 = -1.

Some sets of m variables do not yield a x1 + 2x2 + x3 = 1

basic solution. Consider the linear
system shown to the right: 2x1 + 4x2 + x3 = 3

If NBV = {x3} and BV = {x1, x2} the x1 + 2x2 = 1

corresponding basic solution would be: 2x1 + 4x2 = 3

Since this system has no solution, there is no basic solution

corresponding to BV = {x1, x2}.
 Any basic solution in which all variables are nonnegative is called a
basic feasible solution ( or bfs). For the basic solutions on the
previous slides, x1 = 2, x2 = 1, x3 = 0 and x1 = 0, x2 = 3, x3 = 2 are
basic feasible solutions, but the basic solution x1 = 3, x2 = 0, x3 = -1
fails to be a bfs (because x3 < 0).

The following two theorems explain why the concept of a basic

feasible solution is of great importance in linear programming:

Theorem 1 The feasible region for any linear programming problem

is a convex set. Also, if an LP has an optimal solution, there must be
an extreme point of the feasible region that is optimal.

Theorem 2 For any LP, there is a unique extreme point of the LPs
feasible region corresponding to each basic feasible solution. Also,
there is at least one bfs corresponding to each extreme point in the
feasible region.
 X2
The relationship between D
extreme points and basic
feasible solutions outlined in
Theorem 2, is seen in the
Leather Limited problem. The Labor Constraint
LP (with slack variables) was: B

Feasible Region
max z = 4x1 + 3x2
s.t. x1 + x2 + s1 = 40
Leather Constraint
2x1 + x2 + s2 = 60 E
x1, x2, s1, s2 0

Both inequalities are satisfied in

the shaded area. The extreme C A
points are of the feasible region F 10 20 30 40 50 X1

are B, C, E, and F.
 Basic Nonbasic Basic Feasible Corresponds to Corner
Variables Variables Solution Point

x1, x2 s1 , s 2 s1 = s2 = 0, x1 = x2 = 20 E

x1, s1 x2 , s 2 x2 = s2 = 0, x1 = 30, s1 = 10 C

x1, s2 x2 , s 1 x2 = s1 = 0, x1 = 40, s2 = - 20 Not a bfs since s2 < 0

x2, s1 x1 , s 2 x1 = s2 = 0, s1 = - 20 x2 = 60 Not a bfs since s1 < 0

x2, s2 x1 , s 1 x1 = s1 = 0, x2 = 40, s2 = 20 B

s1, s2 x1 , x 2 x1 = x2 = 0, s1 = 40, s2 = 60 F

The table above shows the correspondence between the basic

feasible solutions to the LP and the extreme points of the feasible
region. The basic feasible solutions to the standard form of the LP
correspond in a natural fashion to the LPs extreme points.
Adjacent Basic Feasible Solutions

For any LP with m constraints, two basic feasible solutions are

said to be adjacent if their sets of basic variables have (m 1)
basic variables in common.

For example in the Leather Limited LP on the previous slide, the bfs
corresponding to point E is adjacent to the bfs corresponding to
point C. These points share (m 1 = 2 - 1 = 1) one basic
variable, x1. Points E (BV = {x1,x2}) and F (BV = {s1,s2}) are not
adjacent since they share no basic variables.

Intuitively, two basic feasible solutions are adjacent if they both lie on
the same edge of the boundary of the feasible region.
General description of the simplex algorithm solving an
LP in a maximization problem:

Step 1 Find a bfs to the LP. We will call this bfs the initial bfs.
In general, the most recent bfs will be called the current
bfs, so at the beginning of the problem, the initial bfs is the
current bfs.

Step 2 Determine if the current bfs is an optimal solution to

the LP. If it is not, find an adjacent bfs that has a larger z-
value.

Step 3 Return to Step 2, using the new bfs as the current bfs.
 The Simplex Algorithm Procedure for maximization LPs

Step 1 Convert the LP to standard form

Step 2 Obtain a bfs (if possible) from the standard form
Step 3 Determine whether the current bfs is optimal
Step 4 If the current bfs is not optimal, determine which nonbasic variable
should be come a basic variable and which basic variable should
become a nonbasic variable to find a bfs with a better objective function
value.
Step 5 Use EROs to find a new bfs with a better objective function value.
Go back to Step 3.

In performing the simplex algorithm, z c1x1 c2x2 - - cnxn = 0

write the objective function in the form:

We call this format the row 0 version of the objective function (row 0 for short).
 The Simplex Algorithm (max LPs)
Consider the simplex algorithm applied to the maximization below:

The Dakota Furniture company manufactures desk, tables, and chairs.

The manufacturer of each type of furniture requires lumber and two
types of skilled labor: finishing and carpentry. The amount of each
resource needed to make each type of furniture is given in the table
below.

Resource Desk Table Chair

Lumber 8 board ft 6 board ft 1 board ft

Finishing hours 4 hours 2 hours 1.5 hours

Carpentry hours 2 hours 1.5 hours 0.5 hours

 The Simplex Algorithm (max LPs)

At present, 48 board feet of lumber, 20 finishing

hours, 8 carpentry hours are available. A desk sells
for $60, a table for $30, and a chair for $ 20.

Dakota believes that demand for desks and chairs

is unlimited, but at most 5 tables can be sold.

Since the available resources have already been

purchased, Dakota wants to maximize total
revenue.
 The Simplex Algorithm (max LPs)
Define:
x1 = number of desks produced
x2 = number of tables produced
x3 = number of chairs produced.

The LP is: max z = 60x1 + 30x2 + 20x3

s.t. 8x1 + 6x2 + x3 48 (lumber constraint)
4x1 + 2x2 + 1.5x3 20 (finishing constraint)
2x1 + 1.5x2 + 0.5x3 8 (carpentry constraint)
x2 5 (table demand constraint)
x1, x2, x3 0
 The Simplex Algorithm (max LPs)
Step 1 - Convert the LP to Standard Form

Canonical Form 0 Basic

Variable

Row 0 z 60x1 30x2 20x3 =0 z=0

Row 1 8x1 + 6x2 + x3 + s1 = 48 s1 = 48

Row 2 4x1 + 2x2 + 1.5x3 + s2 = 20 s2 = 20

Row 3 2x1 + 1.5x2 + 0.5x3 + s3 =8 s3 = 6

Row 4 x2 + s4 =5 s4 = 5

If we set x1 = x2 = x3 = 0, we can solve for the values s1, s2, s3, s4.
Thus, BV = {s1, s2, s3, s4} and NBV = {x1, x2, x3 }. Since each constraint is
then in canonical form (BVs have a coefficient = 1 in one row and zeros in all
other rows) with a nonnegative rhs, a bfs can be obtained by inspection.
 The Simplex Algorithm (max LPs)
Step 2 Obtain a Basic Feasible Solution

To perform the simplex algorithm, we need a basic (although not

necessarily nonnegative) variable for row 0. Since z appears in
row 0 with a coefficient of 1, and z does not appear in any other
row, we use z as the basic variable. With this convention, the
basic feasible solution for our initial canonical form has:
BV = {z, s1, s2, s3, s4} and NBV = {x1, x2, x3 }.

For this initial bfs, z = 0, s1= 48, s2= 20, s3 = 8, s4 =5, x1= x2 = x3 = 0.
As this example indicates, a slack variable can be used as a basic
variable if the rhs of the constraint is nonnegative.
 The Simplex Algorithm (max LPs)
Step 3 Determine if the Current BFS is Optimal
Once we have obtained a bfs, we need to determine whether it is
optimal. To do this, we try to determine if there is any way z can
be increased by increasing some nonbasic variable from its
current value of zero while holding all other nonbasic variables at
their current values of zero. Solving for z in row 0 yields:
Z = 60x1 + 30x2 + 20x3

For each nonbasic variable, we can use the equation above to determine if
increasing a nonbasic variable (while holding all other nonbasic variables to
zero) will increase z. Increasing any of the nonbasic variables will cause
an increase in z. However increasing x1 causes the greatest rate of
increase in z. If x1 increases from its current value of zero, it will have to
become a basic variable. For this reason, x1 is called the entering
variable. Observe x1 has the most negative coefficient in row 0.
 The Simplex Algorithm (max LPs)
Step 4 - We choose the entering variable (in a max problem) to the
nonbasic variable with the most negative coefficient in row 0 (ties
broken arbitrarily). We desire to make x1 as large as possible but as
we do, the current basic variables (s1, s2, s3, s4) will change value.
Thus, increasing x1 may cause a basic variable to become negative.
RATIO
From row 1 we see that s1 = 48 8x1. Since s1 0, x1 48 / 8 = 6
From row 2, we see that s2 = 20 4x1. Since s2 0, x1 20 / 4 = 5
From row 3, we see that s3 = 8 2x1. Since s3 0, x1 8 / 2 =4
From row 4, we see that s4 = 5. For any x1, s4 will always be 0

This means to keep all the basic variables nonnegative, the largest we
can make x1 is min {6, 5, 4} = 4.
 The Simplex Algorithm (max LPs)
The Ratio Test

When entering a variable into the basis, compute the ratio:

rhs of row / coefficient of entering variable in row

For every constraint in which the entering variable has a positive

coefficient. The constraint with the smallest ratio is called the
winner of the ratio test. The smallest ration is the largest value of
the entering variable that will keep all the current basic variables
nonnegative.

Make the entering variable x1 a basic variable in row 3 since this row
(constraint) was the winner of the ratio test (8/2 =4).
 The Simplex Algorithm (max LPs)
To make x1 a basic variable in row 3, we use elementary
row operations (eros) to make x1 have a coefficient
of 1 in row 3 and a coefficient of 0 in all other rows.
This procedure is called pivoting on row 3; and row
3 is called the pivot row. The final result is that x1
replaces s3 as the basic variable for row 3. The term
in the pivot row that involves the entering basic
variable is called the pivot term.

Step 5 - The Gauss-Jordan method using eros and

simplex tableaus shown on the next slide makes x1 a
basic variable.
The Simplex Algorithm (max LPs)
 The Simplex Algorithm (max LPs)
The result is:
Canonical Form 1 Basic
Variable
Row 0 z + 15x2 - 5x3 + 30s3 = 240 z = 240

Row 1 - x3 + s1 - 4s3 = 16 s1 = 16
Row 2 - x2 + 0.5 x3 + s2 - 2 s3 =4 s2 = 4
Row 3 x1 + 0.75x2 + 0.25x3 + 0.5s3 =4 x1 = 4
Row 4 x2 + s4 = 5 s4 = 5

In canonical form 1, BV = {z, s1, s2, x1, s4} and NBV = {s3, x2, x3 }. yielding
the bfs z = 240, s1 = 16, s2 = 4, x1 = 4, s4 = 5, s3 = x2 = x3 = 0.

The procedure from going to one bfs to a better adjacent bfs is called
an iteration (or sometime pivot) of the simplex algorithm.
 The Simplex Algorithm (max LPs)
Returning to Step 3, we attempt to determine if the current bfs is optimal.

Rearranging row 0 from Canonical Form 1, and solving for a yields:

z = 240 15x2 +5x3 -30s3

The current bfs is NOT optimal because increasing x3 to 1 (while

holding the other nonbasic variable to zero) will increase the value
of z. Making either x2 or s3 basic will caused the value of z to
decrease.
Step 4 - Recall the rule for determining the entering variable is the
row 0 coefficient with the greatest negative value. Since x3 is the
only variable with a negative coefficient, x3 should be entered into
the basis.
 The Simplex Algorithm (max LPs)
Begin iteration (pivot) 2
Performing the ratio test using x3 as the entering variable yields the
following results (holding other NBVs to zero):

From row 1, s1 0 for all values of x3 since s1 = 16 + x3

From row 2, s2 0 if x3 > 4 / 0.5 = 8
From row 3, x1 0 if x3 >4 / 0.25 = 16
From row 4, s4 0 for all values of x3 since s4 = 5

This means to keep all the basic variables nonnegative, the largest
we can make x1 is min {8,16} = 8. So, row 2 becomes the pivot row.

Step 5 - Now use eros, to make x3 a basic variable in row 2.

The Simplex Algorithm (max LPs)
 The Simplex Algorithm (max LPs)
The result is:

Canonical Form 2 Basic

Variable
Row 0 z + 5x2 + 10s2 + 10s3 = 280 z = 280

Row 1 - 2x2 + s1 + 2s2 - 8s3 = 24 s1 = 24

Row 2 - 2x2 + x3 + 2s2 - 4s3 =8 x3 = 8

Row 3 x1 + 1.25x2 - 0.5 s2 + 1.5s3 =2 x1 = 2

Row 4 x2 + s4 =5 s4 = 5

In Canonical Form 2, BV = {z, s1, x3, x1, s4} and NBV = {s3, s2, x2 },
yielding the bfs z = 280, s1 = 24, x3 = 8, x1 = 2, s4 = 5, s2 = s3 = x2 = 0.
 The Simplex Algorithm (max LPs)

Solving for z in row 0 yields: Z = 280 - 5x2 - 10s2 -10s3

We can see that increasing x2, s2, or s3 (while holding the other
NBVs to zero) will not cause the value of z to decrease. The
solution at the end of iteration 2 is therefore optimal. The
following rule can be applied to determine whether a canonical
forms bfs is optimal:

A canonical form is optimal (for a max problem) if each nonbasic

variable has a nonnegative coefficient in the canonical forms row 0.
 The Simplex Algorithm (min LPs)
Two different ways the simplex method min z = 2x1 3x2
can be used to solve minimization
problems. Consider the LP shown s.t. x1 + x2 4
to the right: x1 x2 6
x1, x2 0

Method 1

The optimal solution is the point (x1,x2) max -z = -2x1 + 3x2

that makes z = 2x1 3x2 the smallest.
s.t. x1 + x2 4
Equivalently, this point makes max -z
= - 2x1 + 3x2 the largest. This means x1 x2 6
we can find the optimal solution to the x1, x2 0
LP by solving the LP shown to the
right.
 The Simplex Algorithm (min LPs)
In solving this modified LP, use z as the basic variable in row 0.
After adding slack variables s1 and s2 to the constraints we
obtain the initial tableau.
 The Simplex Algorithm (min LPs)
The optimal solution (to the max problem) is -z = 12, x2 = 4, s2 = 10, x1
= s1 = 0. Then the optimal solution to the min problem is z = -12,
x2=4, s2 = 10, x1 = s2 = 0. The min LP objective function confirms
this: z = 2x1-3x2 =2(0) 3 (4) = -12.

In summary, multiply the objective function for the min

problem by -1 and solve the problem as a maximization
problem with the objective function z. The optimal
solution to the max problem will give you the optimal
solution for to the min problem. Remember that (optimal
z-value for the min problem) = - (optimal z-value for the
max problem).
 The Simplex Algorithm (min LPs)
Method 2

A simple modification of the simplex algorithm can be

used to solve min problems directly. Modify Step 3 of
the simplex algorithm as follows: If all nonbasic
variables (NBV) in row 0 have nonpositive coefficients,
the current bfs is optimal. If any nonbasic variable has a
positive coefficient, choose the variable with the most
positive coefficient in row 0 as the entering variable.
This modification of the simplex algorithm works
because increasing a nonbasic variable (NBV) with a
positive coefficient in row 0 will decrease z.