Operations Research

Simplex, Duality and Sensitivity

&
Transportation Problem and Assignment Problem

Symbiosis International (Deemed

University)
• Linear Programming : Graphical Method and Simplex
method
• Graphical method ( coordinate system)
• Used when you are only concentrating on a 2 variable problem (x-y
coordinate)
• Objective function( 1) – constraints ( multiple) , non-negativity
• Steps of solving through graphical
• Special cases : unbounded solution, multiple optimal solution, infeasible,
redundancy
• Feasible region ? – convex – if
Convex nature of feasible region
Convex : Any two points which are
are taken within the region, the
line segment joining the the two
points lies within the region itself,
then it is called convex region
Simplex Method
• Why do we start with simplex ? To handle more than
two variables
• Steps of Simplex
• Type of objective function
• Nature of constraints

Special cases : unbounded, infeasible , multiple optimal solutions, degeneracy

Duality
Sensitivity
Session Objectives
By the end of this session, you will be able to:

1. Explain basics of simplex method/technique

2. Understand some special cases using simplex

3. Understand convex nature of feasible region

4. Understand transportation and assignment problem

What are we going to be
doing ?
Module No. Name of Topic
Module 1 Introduction to Operations Research (IA1)
Module 2 Linear Programming (IA1)
Module 3 Transportation Problem, Transhipment
Problem & Assignment Problem (IA 1)
Module 4 Decision Theory
Module 5 Queuing Theory
Module 6 Simulation Model

Module 7 CPM and PERT (IA1)

Module 8 Forecasting Techniques

Plotting : (for
inequation 1 )
0 9
4.5 0 A,B,C, and D are
extreme Points
Plotting : of the Feasible
0 11 Region
2.75 0

Plotting :
0 2
-2 0
 Q1 (same as Q1 of LPP-1)
Maximize
𝐬𝐮𝐛𝐣𝐞𝐜𝐭 𝐭𝐨
𝟐 𝒙𝟏 +𝟑 𝒙 𝟐 ≤ 60
𝟒 𝒙 𝟏+ 𝟑 𝒙𝟐 ≤ 96
0

We will learn how to start solving with simplex method

First step is to learn how to convert inequations into equations.
Remember for the converted problem:
(1)RHS should never be negative
(2)The unrestricted variables should be written as a difference of two variables
 Step 1 : Convert the inequalities to
equations
Maximize Maximize+0 +0

𝐬𝐮𝐛𝐣𝐞𝐜𝐭 𝐭𝐨 𝐬𝐮𝐛𝐣𝐞𝐜𝐭 𝐭𝐨
𝟐 𝒙𝟏 +𝟑 𝒙 𝟐 ≤ 60 𝟐 𝒙𝟏 +𝟑 𝒙 𝟐+ 𝒔𝟏=60 : slack
variable for
𝟒 𝒙 𝟏+ 𝟑 𝒙𝟐 ≤ 96 constraint 1
: slack
0 𝟒 𝒙 𝟏+ 𝟑 𝒙𝟐 + 𝒔 𝟐=96
variable for
constraint 2
Non-negativity
,0 for all
variables in
Remember: When the inequation is type then the amount of available of raw the system
materials can either be used at full level or less than the level.
The slack variable represents of underutilization of resources if it exists. If there is
no under utilization then the slack variable will come as 0 at the end of the solution
 Note that
only in the
first tables
Step 2 : Simplex initial Table it is the
RHS

𝒄 𝒋 𝟒𝟎𝟑𝟓 𝟎 𝟎
𝑪𝑩 B 𝒙 𝟏 𝒙𝟐 𝒔𝟏 𝒔𝟐 b Ratio=b/
corresponding
𝒔𝟏 𝟎0 ∗23
2 1 0 60 60
variable in entry
=30

𝒔𝟐 𝟎+¿
0 ∗4
4 3 0 1 96
column
2
96
4
= 24 Exiting
variable
-
0-0 𝟑𝟓 𝟎 𝟎

For a maximization problem if- ≤ 0 The minimum positive (≥ 0) ratio is

we have reached optimal, if not then considered to be the exiting variable
take the most positive value as entry
variable
Once the entry variable is decided
calculate the ratio as shown
 Note that
only in the
first tables
Step 2 : Simplex initial Table it is the
RHS

𝒄 𝒋 𝟒𝟎𝟑𝟓 𝟎 𝟎
𝑷𝒓𝒐𝒇𝒊𝒕 𝒙 𝟏 𝒙𝟐 𝒔𝟏 𝒔𝟐 b Ratio=b/
𝑪 Basic Var. corresponding
❑
𝒔 0 ∗23 2
1 0 60 60
𝟎𝑩 variable in entry
=30

𝒔𝟐
𝟏
𝟎 +¿
4 3 0 1 96
column
2
96
4
= 24 Exiting
variable
-
0-0 𝟑𝟓 𝟎 𝟎
 Note that
only in the
first tables
Step 2 : Simplex initial Table it is the
RHS

𝒄 𝒋 𝟒𝟎𝟑𝟓 𝟎 𝟎
𝑪𝑩 B 𝒙 𝟏 𝒙𝟐 𝒔𝟏 𝒔𝟐 b Ratio=b/
corresponding
𝒔𝟏 𝟎0 ∗23
2 1 0 60 60
variable in entry
=30

𝒔𝟐 𝟎+¿
0 ∗4
4 3 0 1 96
column
2
96
4
= 24 Exiting
variable
-
0-0 𝟑𝟓 𝟎 𝟎

For a maximization problem if- ≤ 0 The minimum positive (≥ 0) ratio is

we have reached optimal, if not then considered to be the exiting variable
take the most positive value as entry
variable
Once the entry variable is decided
calculate the ratio as shown
Step 3 : Getting the next table after finding out the
minimum ratio(table 1)
𝟒𝟎
𝒄 𝒋 𝟑𝟓 𝟎 𝟎
𝑪𝑩 B 𝒙 𝟏 𝒙𝟐 𝒔𝟏 𝒔𝟐 b Ratio=b/
corresponding
𝒔𝟏 𝟎 23 1 0 60 60
variable in entry
=30
column
2
96
𝒔 𝟐 𝟎 4 3 0 1 96 4
= 24 Exiting
variable
- -0 𝟑𝟓 𝟎 𝟎
Intersection of
entry column Entry
and exit row is
called the pivot
element
Table 2 (let’s see how we will fill each element)
𝒄𝒄𝟒𝟎
𝒋
𝟒𝟎𝟑𝟓
𝒋 𝟑𝟓 𝟎
𝟎 𝟎𝟎
𝑪𝑩 B 𝒙 𝟏 𝒙 𝟐 𝒔 𝟏 𝒔𝒔 𝟐 b
𝑪𝑩 𝒙𝟏 𝒙 𝒔 𝟐 b
𝒔𝟏 B 0 0
(
𝟐 1 𝟏 − 1 12
2
43 1
Replac
e
𝒔𝟏𝒙𝟏 0 40 0
42
3
4
101
4 2 4
4
12
− 96
exiting
𝒙𝟏
variabl 1 3 0 124
e by 40 4 4
Next how to fill in the non pivot row elements ; use the formula
entry
…
variabl To get updated Table 2 values , first Divide all new element = (
e elements in the pivot row by the pivot
2
𝒄 𝒋 𝟒𝟎 𝟑𝟓 𝟎 𝟎
The elements in the basis always give
the identity

𝑪𝑩 𝟒𝟎𝒙 𝟏
𝟑𝟓𝒙 𝟐𝒔 𝟏 𝒔 𝟐
𝒄B 𝒋 𝟎𝟎 b Ratio

𝟐 𝒔𝟏
Ratio=b/
𝑪𝒔𝑩 B0 𝒙 𝒙
3 𝒔 𝟐1 b corresponding
𝟏
3
2 −
1
𝟏
0
12 8 1
12
variable in entry
𝒔𝟏 0 0
2
1
2
2
−1 12 3
=8
column
Exiting
variable
𝒙 40 1 3 24224
𝒙𝟏 𝟏 40 14
3
4 0 14
0 4 2432 3
= 32

-
- 𝟓
𝟎-30=5
-40=0 𝟎
𝟎 − 𝟏𝟎
𝟎−𝟏𝟎=− 𝟏𝟎 4

Table Since all - does not match - ≤ 0 we have not reached

optimal, taking most positive value as entry variable
3
𝟒𝟎
𝒄 𝒋 𝟑𝟓 𝟎 𝟎
𝑪𝑩 B 𝒙 𝟏 𝒙 𝟐 𝒔𝟏 𝒔𝟐 b
𝒙𝟐 35 2
0 1 - 8
3
𝒙𝟏 40 1 0 -

- 𝟎𝟎 − 25 1 000
3
Since all - ≤ 0 we have reached optimal ; the values of variables are read from the column
b of the table . =18 and 8 and the profit = (40*18)+(35*8)=Rs. 1000
 • A firm produces three products A, B and C each of which passes through
three departments: Fabrication, Finishing and Packaging. Each unit of
product A requires 3,4 and 2 ; a unit of B requires 5,4 and 4 while each
unit of product C requires 2,4 and 5 hours respectively in the three
departments. Every day 60 hours are available in the fabrication
department, 72 hours available in the finishing departments and 100 hours
in the packaging department. The unit contribution of product A is Rs 5, of
product B is Rs. 10 and product C is Rs. 8.
• (a) Formulate the problem as a LPP.
• (b) Determine the number of units of each product that should be made
each day to maximize the total contribution.
(c) Also determine if any capacity would remain unutilized.
Formulation
A B C Available
hours
Where Definition of Variables:
Fabricati
Let number of units of product A to be produced 3 5 2 60
number of units of product B to be produced on
number of units of product C to be produced Finishin 4 4 4 72
g
Packagin 2 4 5 100
g
 Step 1 : Formulation and then Convert
the inequalities to equations
Maximize Maximize+0 +0 +0

𝐬𝐮𝐛𝐣𝐞𝐜𝐭 𝐭𝐨 𝐬𝐮𝐛𝐣𝐞𝐜𝐭 𝐭𝐨
𝟑 𝒙𝟏 +𝟓 𝒙 𝟐+𝟐 𝒙 𝟑 ≤60 𝟑 𝒙𝟏 +𝟓 𝒙 𝟐+𝟐 𝒙 𝟑 +𝒔𝟏 =60 : slack
variable for
𝟒 𝒙 𝟏+𝟒 𝒙𝟐 +𝟒 𝒙 𝟑 ≤72 constraint 1
𝟒 𝒙 𝟏+𝟒 𝒙𝟐 +𝟒 𝒙 𝟑 +𝒔𝟐 =72 : slack
𝟐 𝒙𝟏 +𝟒 𝒙𝟐 +𝟓 𝒙𝟑 ≤ 100 variable for
𝟐 𝒙𝟏 +𝟒 𝒙𝟐 +𝟓 𝒙𝟑 +𝒔 𝟑=100 constraint
: slack 2
0 variable for
,0 constraint 3
Non-negativity
for all
Remember: When the inequation is type then the amount of available of raw variables in
materials can either be used at full level or less than the level. the system
The slack variable represents of underutilization of resources if it exists. If
there is no under utilization then the slack variable will come as 0 at the end of
the solution
 𝟓𝟏𝟎𝟖𝟎 𝟎𝟎
𝑿𝑪
𝑩 𝑩
𝒙𝟏
𝒙𝟐𝒙𝟑
𝒔𝟏𝒔𝟐
𝒔𝟑𝒃 𝐑𝐚𝐭𝐢𝐨
𝟎
𝒔𝟏 𝟑𝟓 𝟐𝟏 𝟎𝟎 𝟔𝟎𝟏𝟐 Exiting
variable
𝒔𝟐
𝟎 𝟒𝟒 𝟒𝟎 𝟏𝟎 𝟕𝟐𝟏𝟖
𝟎
𝒔𝟑 𝟐𝟒 𝟓𝟎 𝟎𝟏 𝟏𝟎𝟎
𝟐𝟓
- 𝟓𝟏𝟎𝟖𝟎 𝟎𝟎
For a maximization problem if- ≤ 0 we have reached
optimal, if not then take the most positive value as entry
variable
𝟓 𝟏𝟎𝟖 𝟎 𝟎𝟎
𝑿𝑪
𝑩 𝑩𝒙𝟏 𝒙𝟐𝒙𝟑 𝒔𝟏 𝒔𝟐
𝒔𝒃𝟑 𝐑𝐚𝐭𝐢𝐨
𝒙𝟐 𝟏
𝟏𝟎𝟑 / 𝟓 𝟐/ 𝟓𝟏/ 𝟓𝟎𝟎 𝟏𝟐 𝟑𝟎
𝒔𝟐𝟎𝟖 / 𝟓
𝟎 𝟏𝟐 /𝟓 - 𝟏 𝟎 𝟐𝟒𝟏𝟎
𝒔𝟑
𝟎− 𝟐 /𝟓
𝟎 𝟏𝟕 /𝟓 - 𝟎𝟏 𝟓𝟐 𝟐𝟔𝟎 /𝟏𝟕
- −𝟏 𝟎 𝟒 −𝟐 𝟎𝟎
Table 3

𝟓 𝟏𝟎𝟖𝟎 𝟎 𝟎
𝒙𝟏 𝒙𝟐 𝒔𝟏
𝒙𝟑 𝒔𝟐 𝒔𝒃𝟑
𝒙𝟐
𝟏𝟎𝟏 / 𝟑𝟏 𝟎𝟏/ 𝟑− 𝟏 /𝟔
𝟎 𝟖
𝒙𝟑𝟎𝟐 / 𝟑𝟎 𝟏- 𝟓/𝟏𝟐 𝟎 𝟏𝟎
𝒔𝟑
𝟎− 𝟖 /𝟎𝟑 𝟎𝟏/ 𝟑−𝟏𝟕 /𝟏𝟐𝟏 𝟏𝟖
- − 𝟏/𝟎
𝟑 𝟎− 𝟐/−
𝟑𝟓/𝟎 𝟑
For a maximization problem since- ≤ 0 we have reached
optimal
The values of production variables are read from the column b of
the table .
=0 , 8 and 10

profit = (5*0)+(10*8)+(8*10)=Rs. 160

18 implies that there is an underutilization of 18 hours for the

packing process
Q3 Solution
 Charne’s Big M method / Big M
Q10 Solutionmethod – greater than equal ,
artificial variables

Minimization
objective
function :
Coefficient of
artificial variable
as –M In the
objective
function

Maximization
objective
function
Coefficient of
artificial variables
in +M in the
objective
function
Simplex
• LPP
• Objective Function – Maximize or Minimize
• Constraints – greater than equal to, less than equal to
and equal to
• Non-negativity
• Maximize , subject to less than equal to constraints
• Maximization
• Less than equal to
• Greater than equal --- -surplus add artificial
• Equality --- add articficial
• Maximization 3x1+ 4x2 • Maximization 3x1 + 4x2
• Subject to constraints • Subject to
• 5x1+5x2<= 15 • 5x1 + 5x2 + s1 =15
• 3x1 + 6x2 >=20 • 3x1 + 6x2 – s2 = 20
• 4x1+5x2=15 • 4x1 + 5x2 =15
• X1,x2 >=0
Special Cases : Simplex
• Unbounded solution : if at any stage while solving through the tables, an
entry variable exists but there is no leaving variable then it is a case of
unbounded solution. (leaving ratio- minimum positive)
• A minimum positive ratio will not exist if the ratio is negative or undefined.
• MCQ : If the leaving ratios are -4, 0, -8 for a simplex table what will you choose ? 0
• Leaving ratio as -4,-5, 7/0 choose the correct leaving ratio – none of the options
• Multiple optimal : If there are values of cj-zj of non-basic elements as 0 in
the optimal table then this is the case of multiple optimal solution.
• Degeneracy : if there is a tie in the case of leaving ratio, then the solution is
said to have degeneracy
• Infeasibility : there is an artificial variable that exists in the final table ( cj-zj
condition matched) then it is case of infeasibility
Q15 Solution
Q18 Solution
Q19 Solution
Q19 Solution Contd.
Q19 Solution Contd.
Q19 Solution Contd.
Multiple Optimal Solution
• You arrive at an optimal solution
• Once you arrive at the optimal --- there is an other option
which is equally good
• Maximize 120 x1 + 120 x2 = 120*3+120*6
• 120*6+120*3

• X1= 3 , x2= 6
• X1=6 X2= 3
Other topics LPP
• Duality – relationship between primal and dual

• Sensitivity
 Transportation Problem
Objective is to find
best cost / Minimum
cost for transporting
goods from multiple Step I: Initial Solution to To find out the initial
starting Points to Start with solution (Step I ) we have 3
multiple destinations methods

NWC LCM VAM

1
Module 3
Transportation and 2
Assignment MODI- 3 3
Step II: From Initial Solution
Problem Modified we move to optimality after
Distribution conditions are met
method By MODI method

Steps for MODI If

method do not Option 1 : Initial solution by We pick up option 3
change whichever NWC and then Optimality by i.e. VAM and then
initial solution you MODI Optimality as it converges
take , only the If the fastest to optimal
computation time Option 2 : Initial solution by solution (best solution)
matters LCM and then Optimality by Else option 1 and 2 could
MODI also be taken
What are we doing ?
Imagine this is your
notebook
Destinations

Origins

Remember to write your

values in the table spaced out
(write on every alternate line
Q1 a
A firm owns facilities at seven places. It has manufacturing
plants at places A,B and C with daily output of 500,300 and 200
units of an item respectively. It has warehouses at places P,Q,R
and S with daily requirements of 180,150, 350 and 320 units
respectively. Per unit shipping charges on different routes are:

Destinations
P Q R S
A 12 10 12 13
Origins
B 7 11 8 14
C 6 16 11 7
Steps for NWC
1. Before solving a transportation problem check TS=TD
(balanced TP)
2. Choose the north-west corner of the transportation matrix
3. Start allocation by first checking the supply and demand
combinations. The allocations are made on the condition
that origins cannot supply more than their capacity and
destination points cannot take in more than what they
have.
4. Make allocation till we have exhausted the demand and
supply.
5. Find total cost for this initial solution
So Let’s get are note books out and get to work
Solution of IBFS : By north-
west corner
Destinations
P Q R S Supply
A 12 180 10 150 12 170 13 500 320 170 0
Origins
B 7 11 8 180 14 120 300 120 0
C 6 16 11 7 200 200
Demand 180 150 350 320 1000
0 0 180 200
0

TS=TD
Solution of IBFS : By north-
west corner (Allocations)
Destinations
P Q R S
A 12 180 10 150 12 170 13
Origins
B 7 11 8 180 14 120
C 6 16 11 7 200

Total Cost =
(12*180)+(10*150)+(12*170)+(14*120)+(7*200)=10220
Steps for LCM
1. Before solving a transportation problem check TS=TD
(balanced TP)
2. Search the least cost cell in the matrix
3. If there is a tie in the least cost then choose the one with
which allows for maximum allocation. If again there is a tie
in the allocation also, then choose either cell to start
allocation.
4. Start with allocation from least cost cell.
5. Allocations are carried out by matching demand and supply
values. Allocation in a particular cell will not be more than
corresponding supply or demand.
6. Continue allocation till exhausting supply.
Q2 a IBFS : LCM
TD=TS=1000
P Q R S Sup.
A 12 10
10 12
12 13 500
500
500
(150) (350)
(350)
(50) (300) (300)
(300)
B 7 11 8 14 300
8 (0)
300
(0)
C 6 16 (300)
11 7 200
6 7 200
(20)
(0)
Req. (180)
180 150 350 (20)
320 100
0
180 (150) 350
350 320
320
(0) (0) (50)(50) (300)
(300)
(0) (0)
Q2 a IBFS : LCM Allocation
P Q R S Sup.
A 12 10 12 13 500

(150) (50) (300)

B 7 11 8 14 300

(300)
C 6 16 11 7 200

Req. (180)
180 150 350 (20)
320 100
0

Total Cost =
(10*150)+(12*50)+(13*300)+(8*300)+(6*180)+(7*20)=9620
Steps IBFS : VAM
1. Before solving a transportation problem check TS=TD
(balanced TP)

2. Calculate the penalties for each row and column separately.

Penalty is the difference between the least cost and next
least (best and next cost.
3. Look for the highest penalty value.
4. Start allocation in that row or column corresponding to
highest penalty value in the cell which has minimum cost.
5. If there is a tie in the penalty then choose that row/column
which has maximum allocation possible in its minimum cost
cell. If there still is a tie in the maximum allocation , then
choose the minimum cost . If still a tie then allocate in any with
the highest penalty .
6. Continue allocation till supply is exhausted.
Q 3 a IBFS : VAM
P Q R S Sup. I II III
A 12 10 12 13 500 2 2 2
(150) (230) (120)
B 7 11 8 14 300
300
300 1
1 3
(120)
(120)
C (180)
6 16 (120)
11 7 (0)
200
200 1 --- ---
(0)
Req. 180 150 350 (200)
320 100
0
180 150 350
350 320 320 TD=TS=1000
(0) (0) (230)
(230) (120)
(120)
(0) (0)
I 1 1 3 6
II 5 1 4 1
III --- 1 4 1
Q4 a
• A company has 3 plants at locations A,B and C  Assume that the penalty for failing
which produce the same product. It has to
supply this to the buyer located at D,E,F . The to meet the supply buyer
weekly plant capacities for A,B and C are 100, requirements is Rs. 4 , Rs. 3 and Rs.
800 and 150 units respectively, while the buyer 3 per unit in respect to D , E and F
requirements are 750, 200 and 500 units respectively.
respectively for D,E, and F. The unit shipping  Balance the problem using
costs(in Rs.) are given below : penalty
 Find the IBFS by VAM
Plant Buyer
 Test the optimality of the
D E F
solution by MODI method and
A 8 4 10 verify the solution
B 9 7 9  Give the transportation
schedule and the total
C 6 5 8
transportation cost. How much
is the penalty to be paid for the
unsatisfied demand?
Q4 Solution
D E F Sup.
A 8 4 10 100

B 9 7 9 800

C 6 5 8 150
D 4 3 3 400
Dema 750 200 500 1450 Is TD=TS? Notice TD>TS
nd

Case II : If TS >TD then

Balance by TD=1450
balance by adding a
adding a dummy TS=1050
demand point with
supply point Difference =400
demand equal to the
difference with cost 0
 Step I : Find initial solution by
VAM
D E F Sup. I II III
A 8 4 10 100
100 4 4
(100) (0) ---

B 9 7 9 800 2 2
(600) (100) (100) 2

C 6 5 8 150
150 1 1
(150) (0) 1

D 4 3 3 400
400
0 ---
(400) (0) ----

Deman 750
750 200200 500
500
d (600) (100) (100)
I 2 1 5
II 2 1 1
III 3 2 1
 Checking for Optimality after VAM
Check if m(no. of origins )+ n(no. of destinations)
-1 is equal to no. of allocation s
If this condition is met then you can go ahead for
MODI method
For allocated
D E F ui cells
A 8 4 10 cij =(ui+vj)
(100) [4] u1=-3
[2]
B 9 7 9
(600) (100) (100) u2=0 For
unallocated
C 6 5 8 cells
(150) u3=-3
[1] [2] Deltaij=
D 4 3 3 u4=-6 cij -(ui+vj)
[1] [2] (400)

vj v1=9 v2=7 v3=9

All delta values are positive , we have reachedTotal cost =

optimality (4*100)+(9*600)+(7*100)+(9*100)
+(6*150)+(3*400)= Rs. 9500
Q5 IBFS by VAM
1 2 3 4 Sup. I II III IV
7
A 20 30 50 17 7
(2) 3 13 33 33
7
(2) ---
(5) (2) (0)
10
B 70 35 40 60 10
(3) 5 5 20 20 20
(7) (3) (0)
18
18
C 40 12 60 25 18
(10)
(10) 13 13 35 --- ---
(8) (10) 0

Dema 55 88 7 7 1515 Is TD=TS?

nd (0) (0) (0) (13) 35

I 20 18 10 8
II -- 18 10 8
III --- --- 10 8
IV --- --- 10 43
--- --- 33 ---
Conditions to Check
Remember to check the m+n-1 condition
after the updated allocation
Have you calculated the total cost?

Total cost of IBFS =

(20*5)+(17*2)+(40*7)+
(60*3)+(12*8)+(25*10)= Rs. 940
 Optimality by MODI
after VAM 1 2 3 4 ui
A 20 30 50 17
(5) (2) u1=0
[26] [53]
B 70 35 + 40 60 -
u2=43
[7] [-12] (7) (3)
C 40 12 60 25
(8) - (10) u3=8
[12] [55] +

vj v1=20 v2=4 v3=-3 v4=17

All delta values are Start loop from cell Obtain theta which is For updated next
not positive therefore with most negative minimum of table , add theta to
optimality not delta. You are allowed allocations of cell allocation of cells
reached to turn only at the marked by - marked + in the loop
Thus construct loop allocated cells and subtract from
Let’s see how the You are supposed to allocation of cells
Theta min = min
loop is constructed come back at the marked as -
(8,3)=3
same cell from where
you started to close
the loop
Next Table after Updating
1 2 3 4 ui
A 20 30 50 17
(5) (2) u1=-31
[26] [41] Current Expenditure
B 70 35 40 60 u2=0 on Transport Rs. 1200
[19] (3) (7) [12] Total Saving (by using
C 40 12 60 25 transportation
(5) (13) u3=-23 solution)
[12] [43] =Rs.(1200-904)
vj v1=51 v2=35 v3=40 v4=48 =Rs. 296

Since deltas are all Total cost =

greater than 0 we have
reached unique (20*5)+(17*2)+(35*3)+(40*7)+
optimal solution
(12*5)+(12*5)= Rs. 904
Q6 DIY Find the optimal
solution
D1 D2 D3 D4 Availability
P1 19 30 50 12 7
P2 70 30 40 60 10
Plant
P3 40 10 60 20 8
Req. 5 8 7 15
Q6 DIY IBFS Is TD=TS? TD=35
TS=25

by VAM Add a dummy origin

D1 D2 D3 D4 Avail.
P1 19 30 50 12 7
P2 70 30 40 60 10
P3 40 10 60 20 8
Req. 5 8 7 15

D1 D2 D3 D4 Avail.
P1 19 30 50 12 7
P2 70 30 40 60 10
P3 40 10 60 20 8
P4 0 0 0 0 10
Req. 5 8 7 15 35 TD=TS
 Q6 DIY IBFS by VAM
D1 D2 D3 D4 Avail. I II III IV
P1 19 30 50 12 5 750 7 7 7 7
2
P2 70 30 40 60 10 10 30 10
8 2 02 10
P3 40 10 60 20 8 80 10 10 20
20 --
P4 03 0 07 0 10
30 0 0 0 0
Req. 520 80 70 15
70 35
I 19 10 40 12
II 19 10 -- 12
III 19 -- -- 12
IV 19 -- -- 12
 Q6 DIY Start Checking for
Optimality
Is m+n-1 = no.
D1 D2 D3 D4 Avail. of allocations?
P1 19 30 50 12 5 7
2
P2 70 30 40 60 10
8 2 Yes?
P3 40 10 60 20 8
8
P4 03 0 07 0 10 Then move
Req. 5 8 7 15 35 ahead to
obtain
optimal
solution by
MODI
Q6 DIY Start Checking for Optimality
All delta values are not positive therefore
optimality not reached
Thus construct loop
Let’s see how the loop is constructed

D1 D2 D3 D4 ui Start loop from cell with most negative delta.

P1 19 30 50 12 + You are allowed to turn only at the allocated
u1=0 cells
- 2 [48] [31] 5 You are supposed to come back at the same
P2 70 30 40 + 60 - cell from where you started to close the loop

[3] 8 [-13] 2 u2=48

P3 40 10 60 20 Obtain theta which is
u3=8 minimum of
[13] [20] [33] 8
P4 0 0 0 0 allocations of cell
marked by -

+3 7 u4=-19
[37] - [7] Theta min = min
vj (2,2,7)=2
v1=19 v2=-18 v3=19 v4=12

For updated next table , add theta to

allocation of cells marked + in the loop and
subtract from allocation of cells marked as
-
 Q6 DIY Start Checking for Optimality
Is m+n-1 = no. of D1 D2 D3 D4 ui
allocations? P1 19 30 50 12
u1=12
No … This is called [7] [28] [38] 7
a case of P2 70 30 40 60
degenerate [30] 8 x 2 [20] u2=40
solution P3 40 10 60 20

In simple words the [20] 8 u3=20

P4 0 0[0] [40]
0 0
solution goes into a x
loop 5 5 u4=0
[10] E
You cannot solve vj v1=0 v2=-10 v3=0 v4=0
directly, add an
allocation called
epsilon (small Independent Now
quantity) cell … as you check
cannot make m+n-1
This epsilon is An independent cell a loop from condition
added to a is one from which this cell … so We have
minimum cost you cannot form a allocate got 7 cells
independent cell loop (unallocated) epsilon allocated
Q6 DIY Start Checking for
Optimality
Since we have all
delta values
positive therefore,
we have reached
optimality

First calculate the

total cost
Total Optimal cost =
(12*7)+(30*8)+(40*2)+
(20*8)+(0*5)+(0*E)= Rs. 564

Moreover, tell me You will notice that some Will explain how to handle
are all the delta or one delta cell value is 0 multiple optimal solutions
values >0 … this can be identified as in the next slide
a case of multiple solution
case
 Q6 DIY For getting alternate optimal
Since one of the delta values D1 D2 D3 D4 ui
is 0, thus we have the case of P1 19 30 50 12
multiple optimal solution u1=12
[7] [28] [38] 7
This means we can find an P2 70 - 30 40 60
+
alternate optimal solution [30] 8 2 [20] u2=40
P3 40 10 60 20
To get the alternate optimal + -
[20] 8 u3=20
solution , we start making a P4 0 0[0] [40]
0 0
loop from the cell which has
delta value as 0 5 5 u4=0
[10] - E +

After forming loop , again vj v1=0 v2=-10 v3=0 v4=0

mark alternately +, -

As earlier find minimum Total Alternate Optimal cost =

theta value (12*7)+(30*3)+(40*7)+
(10*5)+(20*3)= Rs. 564
theta= min(8,5,8)=5
Note the optimal cost will be
+5 to cells marked[+] and the same , only the allocations
subtract -5 to cells marked [-] will change
MCQs
• What is use of artificial variables in simplex ?
• If you have 3 constraints in a LPP , what is the order of the identity matrix? ( 3
x3 )
• In LPP – graphical method : what is the nature of the feasible solution? –
convex in nature
• In LPP how do you identify in simplex degeneracy – if there are ties in the
leaving ratio
• In Transportation in the north-west corner rule which of these is not
considered – minimum cost is not considered
• In transportation how does dummy row or dummy column help? - for
satisfying the RIM conditions
• Is assignment problem a special case of Transportation
• Transportation is a special case of LPP
• Dual has the same number of constraints as there are number of variables in
the primal.
• Profit maximization in case of assignment problem – we convert it to
opportunity loss problem and then go ahead
Special Cases : TP
• Degeneracy
• Multiple optimal
• Unbalanced
• Prohibited routes
Session Outcomes
In this session you learned about:

1. Basics of simplex method/technique

2. Some special cases using simplex

3. Convex nature of feasible region

4. Transportation and assignment problem

Thank You