LINEAR

PROGRAMMING
IN MATRIX FORM
IEng-151
Advanced Math for Industrial Engineers
 Linear
Programming > - one of the scientific techniques
that is used to get an optimum
solution to the given business
problem by taking resource
scarcity and constraints into
account.

> - also known as Linear

Optimization
 Linear
Programming
! Optimization - is a general
term used to describe types of
problems and solution
techniques that are concerned
with the best (“optimal”)
allocation of limited resources
in projects.
OPTIMIZATION APPLICATIONS 4

1 Optimizing production plans,

2 Optimizing usage of transportation systems,

3 Optimal shipping,

4 Efficiency in running of power plants,

5 Designing optimal financial portfolios,
6 Minimization of environmental impact.
 Linear
Programming
> - methods for finding a maximum
(or a minimum) x = (x1..., xn) of a
linear objective function:

! 𝑧 = 𝑓 𝑥 = 𝑎1 𝑥1 + 𝑎2 𝑥2 + ⋯ + 𝑎𝑛 𝑥𝑛

…satisfying the constraints.

 Linear
Programming
! Objective Function – some
function (f), which is the focal
point or goal of the
optimization problem.
𝑧 = 𝑓 𝑥 = 𝑎1 𝑥1 + 𝑎2 𝑥2 + ⋯ + 𝑎𝑛 𝑥𝑛
 Linear
Programming ! Control Variables – one of the
several variables (x1, x2…xn) on
which function (f) depends
and whose values we can
control or choose. The choice
of values is not entirely free
but is subject to some
constraint.
 Linear
Programming
! Constraints - additional
restrictions arising from the
nature of the problem and the
variables. It can be in the form
of equations or inequalities.
 Linear
Programming
! Slack Variables - is a
nonnegative auxiliary variable
introduced for converting
inequalities to equations
because it “takes up the
slack” or difference between
the two sides of the inequality
 Linear
Feasibility Region - Set of all
Programming !
feasible solutions, meaning
solutions that satisfy all
constraints.

Basic Feasible Solution - a

feasible solution for which at
! least n - m of the variables
x1,…xn are zero.

**n number of variables and m

number of equations
 Linear
Programming
! A feasible solution is called an
optimal solution if, for it, the
objective function f becomes
maximum, compared with the
values of f at all feasible
solutions.
 Linear
Programming
! A feasible solution is called an
optimal solution if, for it, the
objective function f becomes
maximum, compared with the
values of f at all feasible
solutions.
Simplex
z GEORGE BERNARD DANTZIG (1914–
2005), American mathematician,
Method x1
who is one of the pioneers of linear
programming and inventor of the
simplex method.
.
He invented this famous method to
solve large-scale planning (logistics)
. problems.

. He also was a driving force in

establishing the field of linear
programming and became professor
xn of transportation sciences,
operations research, and computer
axn
science at Stanford University.
Linear Programming in Normal Form 14

Maximize
!
𝑓 = 𝑐1 𝑥1 + 𝑐2 𝑥2 + ⋯ 𝑐𝑛 𝑥𝑛

subject to the constraints

𝑎11 𝑥1 + ⋯ 𝑎1𝑛 𝑥𝑛 = 𝑏1
𝑎21 𝑥1 + ⋯ 𝑎2𝑛 𝑥𝑛 = 𝑏2
…………………………
𝑎𝑚1 𝑥1 + ⋯ 𝑎𝑚𝑛 𝑥𝑛 = 𝑏𝑚
𝑥𝑖 ≥ 0 𝑖 = 1, … , 𝑛
Linear Programming Sample Problem 15

1 Energy Savers, Inc., produces heaters of types S and L. The

wholesale price is per heater for S and for L. Two time
constraints result from the use of two machines and On one
needs 2 min for an S heater and 8 min for an L heater. On
one needs 5 min for an S heater and 2 min for an L heater.
Determine production figures and for S and L, respectively
(number of heaters produced per hour), so that the hourly
revenue is maximum.
Linear Programming Sample Problem 16

1 Solution. Production figures and must be nonnegative.

Hence the objective function (to be maximized) and the
four constraints are:
Linear Programming Sample Problem 17

1 Solution. Converting the first two inequalities to

equations by introducing two slack variables we
obtained the normal form of the problem:
Linear Programming Sample Problem 18

1 Solution. To find an optimal solution of it, we may

consider its augmented matrix. This matrix is called a
simplex tableau or simplex table (the initial simplex
table):
Linear Programming Sample Problem 19

1 Solution. Every simplex table gives a basic feasible

solution. It is obtained by setting the nonbasic variables
to zero. Thus T0 gives the basic feasible solution :
Linear Programming Sample Problem 20

1 Solution.
Step 1.
Operation 1 : Selection of the Column of the Pivot. Select as the
column of the pivot the first column with a negative entry in Row
1. In T0 this is Column 2 (because of the -40 ).

Operation 2 : Selection of the Row of the Pivot. Divide the right

sides [60 and 60 in T0] by the corresponding entries of the
column just selected (60/2=30, 60/5 = 12). Take as the pivot
equation the equation that gives the smallest quotient. Thus the
pivot is 5 because is smallest.
Linear Programming Sample Problem 21

1 Solution.
Step 1.
Operation 3 : Elimination by Row Operations. This gives
zeros above and below the pivot (as in Gauss–Jordan).
Linear Programming Sample Problem 22

1 Step 1.
We see that basic variables are now x1, x3 and nonbasic
variables are x2, x4.
Linear Programming Sample Problem 23

1 Step 1.
Setting the latter to zero, we obtain the basic feasible
solution given by T1).
Linear Programming Sample Problem 24

1 Step 2.
The basic feasible solution given by T1 is not yet optimal
because of the negative entry -72 in Row 1. Accordingly, we
perform the operations O1 to O3 again, choosing a pivot in the
column of -72.

Operation 1. Select Column 3 of in T1 as the column of the pivot

(because -72 <0).

Operation 2. We have 36/7.2 = 5 and 60/2 = 30. Select 7.2 as

the pivot (because 5 < 30).
Linear Programming Sample Problem 25

1 Step 2.
Operation 3. Elimination by row operations gives:
Linear Programming Sample Problem 26

1 Step 2.
We see that now x1,x2 are basic and x3, x4 nonbasic.
Setting the latter to zero, we obtain from the basic
feasible solution:
Linear Programming Sample Problem 27

Minimization. If we want to minimize (instead

of maximize), we take as the columns of the
pivots those whose entry in Row 1 is positive
(instead of negative).
Simplex Method: Difficulties 28

z DEGENERACY
x1 A degenerate feasible solution is a feasible
solution at which more than the usual number
.
n - m of variables are zero.
.
Here n is the number of variables (slack and
. others) and m the number of constraints (not
counting the conditions).
xn
 Simplex Method: Difficulties 29

• AB Steel, Inc., produces two kinds of iron by using three kinds of raw
material (scrap iron and two kinds of ore) as shown. Maximize the daily
profit.
 Simplex Method: Degeneracies 30

• Solution:
• Let x1 and x2 denote the amount (in tons) of iron and respectively, produced
per day. Then our problem is as follows. Maximize:
 Simplex Method: Degeneracies 31

• By introducing slack variables we obtain the normal form of the

constraints
 Simplex Method: Degeneracies 32

• We obtain from (1) and (2) the initial simplex table:

 Simplex Method: Degeneracies 33

• We obtain from (1) and (2) the initial simplex table T0. We see that are nonbasic x1, x2
variables and x3, x4, x5 are basic. With we have from (3) the basic feasible solution
 Simplex Method: Degeneracies 34

• Column Selection of Pivot. Column 2 (since -150 < 0).

• Row Selection of Pivot. 16/2 = 8, 8/1 = 8, 3.5/0 is not possible. Hence we
could choose Row 2 or Row 3. We choose Row 2. The pivot is 2.
 Simplex Method: Degeneracies 35

• Column Selection of Pivot. Column 3 (since -225<0).

• Row Selection of Pivot. 16/1=16, 0/0.5 = 0. Hence 0.5 must serve as the
pivot. Operation :
 Simplex Method: Degeneracies 36

• Column Selection of Pivot. Column 4 (since -150<0). Row Selection of

Pivot 16/2 = 8, 0/-0.5 = 0, 3.5/1 = 3.5 . We can take 1 as the pivot.
 Simplex Method: Degeneracies 37

Since Row 1 of has no negative entries, we have reached the maximum

daily profit This is obtained by using 4.5 tons of iron and 3.5 tons of iron.