Welcome

to
Understanding and Optimization of
Transportation Problem

Dr. Ramesh Kumar Chaturvedi

BBAU, Lucknow
 What is Transportation Problem?
• A special case of Linear Programming.
• In 1941 by FL Hitchcock, then in 1947 by TC Koopmans
in a study titled “Optimum utilization of Transportation
system”
• RIM conditions
• Destination demand and Supply
• Minimization of transportation cost
• Unit cost of transportation is known and fixed
• Cost is proportional to number of units transported
• Important terms: Initial Basic Feasible solution,
Optimal solution, Degeneracy, Balanced problem
 A Transportation Problem

Project A B C Availability

1 2 7 4 50

2 3 3 1 80

3 5 4 7 70

4 1 6 2 140

Demand 70 90 180
 Conditions to be fulfilled for IBFS
• Number of allocation should be m+n-1
• All allocations must be independent
 Methods
• NWS
• LCM
• VAM
 A Transportation Problem
solution by NWC
Project A B C Availability

1 2 (50) 7 4 50

2 3 (20) 3(60) 1 80

3 5 4(30) 7(40) 70

4 1 6 2(140) 140

Demand 70 90 180 340

Total cost=1020
 A Transportation Problem
solution by LCM
Project A B C Availability

1 2 7(20) 4(30) 50

2 3 3 1(80) 80

3 5 4(70) 7 70

4 1(70) 6 2(70) 140

Demand 70 90 180 340

Total cost=830
 Steps in VAM
• Construct Transportation Table
• For each row calculate the difference between two lowest
cost entries.
• For each column calculate the difference between two lowest
cost entries.
• Select the row or column that has largest difference
• In row or column identified in above step select the cell that
has lowest cost entry.
• Assign the maximum possible number of units to the cell
selected in step 5.
• Reapply the steps 2 to 6 iteratively till total demand is met by
supply
 A Transportation Problem
solution by VAM
Project A B C Availability

1 2 7(20) 4(30) 50

2 3 3 1(80) 80

3 5 4(70) 7 70

4 1(70) 6 2(70) 140

Demand 70 90 180 340

Total cost=830
 Variations of Transportation problem
• Degeneracy in transportation problem when
No. of allocations is less than m+n-1
• Unbalanced problem
• Converting a maximization problem into
minimization problem
 Dealing Degeneracy
A B c d E F
1 9 12 9 6 9 10 5
2 7 3 7 7 5 5 6
3 6 5 9 11 3 11 2
4 6 8 11 2 2 10 9
4 4 6 2 4 2 22

Solve the problem with VAM

 Dealing Degeneracy
A B c d E F
1 9 12 9(5) 6 9 10 5
2 7 3(4) 7 7 5 5(2) 6
3 6 (1) 5 9(1) 11 3 11 2
4 6(3) 8 11 2(2) 2(4) 10 9
4 4 6 2 4 2 22

Is Number of allocations is less than m+n-1?

 Dealing Degeneracy
Solve Degeneracy by assigning € in unoccupied
cell with smallest cost, yet closed loop is not
formed.
A B c d E F
1 9 12 9(5) 6 9 10 5
2 7 3(4) 7 7 5 (€) 5(2) 6
3 6 (1) 5 9(1) 11 3 11 2
4 6(3) 8 11 2(2) 2(4) 10 9
4 4 6 2 4 2 22
 Problem with degeneracy

Project A B C Availability

1 2 7 4 50

2 3 8 1 80

3 6 4 7 70

4 1 6 2 120

Demand 50 90 180
 Solution of Problem with
degeneracy (By NWC)
Project A B C Availability

1 2(50) 7 4 50

2 3(€) 8(80) 1 80

3 6 4(10) 7(60) 70

4 1 6 2(120) 120

Demand 50 90 180

Note: Degeneracy may occur in subsequent iterations while optimizing

the IBFS
 Unbalanced Problem
Total Demand ≠ Total Supply

No feasible solutions can be obtained for such

problems
1. In case total demand is more than supply than
dummy source is added to fulfil the Balance
requirement.
2. In case total supply is more than total demand
than dummy destination is added to fulfil the
Balance requirement.
 An Unbalanced Transportation
Problem
Is this Problem Unbalanced?
Project A B C Availability

1 2 7 4 50

2 3 8 1 80

3 6 4 7 70

4 1 6 2 140

Demand 90 90 180
Using Dummy for an Unbalanced
Transportation Problem
Project A B C Availability

1 2 (50) 7 4 50

2 3 3 1 80

3 5 4 7 70

4 1 6 2 140

Dummy 0 0 0 20

Demand 90 90 180
 Converting a maximization problem
into minimization problem

1. Subtract all the costs in transportation table

in from highest cost.

2. The Table so obtained is Now transformed

into minimization problem and can be solve
with methods discussed earlier.
 Is the solution (IBFS) best?
• Decision making
• Concept of Bounded rationality in decision
making
 Types of decision situation
1. Hobson's Choice
2. Certainty
3. Uncertainty
4. Risk
 Decision Making

Simon’s Model of decision Making

Is it possible to execute Simons
model practically?

Concept of Bounded Rationality

and
Optimality
Optimality in Transportation
Problem

Stepping Stone Method

MODI Method
 Optimality in Transportation
Problem - Stepping Stone Method
•This method involves evaluation of each unoccupied cell for
opportunity cost.

•If opportunity cost of any unoccupied cell is positive (if more

than one unoccupied cell has positive opportunity cost, then
the cell with maximum positive OC), then this cell is bought in
the solution.
Note: Since Each unoccupied cell is to be evaluated for OC then on with
maximum OC is selected and then re allocation is to be done, this method
becomes cumbersome, hence not much in use. (M-1)(N-1) unoccupied cells
are there that has to be evaluated. i.e. mn – (m+n-1) cells
 MODI Method
Modified Distribution
•MODI method is more apt in optimizing
Transportation problem.

•Opportunity cost for unoccupied cells can be

calculated arithmetically.

•Only one cell is evaluated by forming the close

loop path to improve the solution
 Steps in MODI Method
1. Setup IBFS by NWS/LCM/VAM
2. Introduce dual variables corresponding to supply and demand
constraints. If there are m supply and n demand points then m+n
dual variables will be used. Let for row ui (i=1,2,…m) and for
column vj be (j=1,2,…n) be dual variables corresponding to supply
and demand constraints. Cij Is the cost in a cell in transportation
table.
3. Variables ui and vj are such that, Opportunity cost(OC) = ui + vj -
cij .
4. Now, ui + vj = cij for all occupied cells because opportunity cost for
occupied cells are Zero.
5. Since the number of dual variables is m+n and number of
occupied cells is m+n-1, one variable is assumed arbitrarily, say
u1=0
6. After assigning u1=0; Calculate the remaining ui and vj values from
the relationship ui + vj = cij
 Steps cont…
7. Calculate the opportunity cost for all unoccupied cells from the
relationship ui + vj - cij .
8. If opportunity cost of all unoccupied cells is zero or negative than
an optimal solution is achieved.
9. In case cells have positive opportunity cost, select the cell with
maximum OC. Starting from the selected cell and moving only
along horizontal or vertical lines trace a closed path back to the
selected cell such a way all corners of this path are occupied cells.
Beginning with this cell assign positive (+) and negative (-) signs
alternatively to all corners (the occupied cells) of the closed path.
10. v Determine the smallest quantity in a negative position on the
closed path. Add this quantity to all corners with positive sign and
subtract it from all corners with negative sign. This gives an
improved solution.
11. Check this solution for optimality by going through with step 7.
 Solving for optimality an IBFS
obtained through NWC
v1 v2 v3 v4

Project A B C Availability

u1 1 2 (50) 7 4 50

u2 2 3 (20) 3(60) 1 80

u3 3 5 4(30) 7(40) 70

u4 4 1 6 2(140) 140

Demand 70 90 180 340

Transportation cost is Rs. 1020.

Is it Optimal?
 Solving for Optimality
• From all occupied cells calculate the values of u1
and v1 from ui + vj = cij ; assuming u1=0, we get
as follows:
For cell 1A; 2 = u1 + v1; Since u1 = 0, v1=2
For cell 2A; 3= u2 +v1; since v1=2, u2=1
For cell 2B; 3= u2 +v2; since u2=1, v2=2
For cell 3B; 4= u3 +v2; since v2=2, u3=2
For cell 3c; 7= u3 +v3; since u3=2, v3=5
For cell 4c; 2= u4 +v3; since v3=5, u4=-3
 Solving for Optimality cont…
• Let us now calculate OC from all unoccupied cells
as follows:

For cell 1B; u1 + v2 – C12 = 0+2-7 = -5

For cell 1C; u1 + v3 – C13 = 0+5-4 = 1
For cell 2C; u2 + v3 – C23 = 1+5-1 = 5
For cell 3A; u3 + v1 – C31 = 2+2-5 = -1
For cell 4A; u4 + v1 – C41 = -3+2-1 = -2
For cell 4B; u4 + v2 – C42 = -3+2-6 = -7
 Solving for optimality Solution 1

v1 v2 v3 v4

Project A B C Availability

u1 1 2 (50) 7 (-5) 4 (+1) 50

u2 2 3 (20) 3(60) 1(+5) 80

u3 3 5 (-1) 4(30) 7(40) 70

u4 4 1(-2) 6 (-7) 2(140) 140

Demand 70 90 180 340

Cell 2C has +5 OC, This solution is not optimal. Hence a

closed path is traced from this cell to get improved solution
 Solving for optimality
v1 v2 v3 v4

Project A B C Availability

u1 1 2 (50) 7 (-5) 4 (+1) 50

u2 2 3 (20) 3(60) _ 1(+5) + 80

u3 3 5 (-1) 4(30) + 7(40) _ 70

u4 4 1(-2) 6 (-7) 2(140) 140

Demand 70 90 180 340

40 at cell 3c is the smallest quantity at –ve

corner, this will be added and subtracted to +ve
and –ve corners to get new solution.
 Improved solution 2
v1 v2 v3 v4

Project A B C Availability

u1 1 2 (50) 7 4 50

u2 2 3 (20) 3(20) 1(40) 80

u3 3 5 4(70) 7 70

u4 4 1 6 2(140) 140

Demand 70 90 180 340

Now check this solution for

optimality
 Solving for Optimality cont…
• Let us now calculate OC from all unoccupied cells as
follows:
• Recalculate the values of u1 and V1 from relationship
ui + vj = cij ; assuming u1=0, we get v1=2, v2=2 and v3=0
similarly u2=1, u3=2, u4=2.
• Opportunity cost as follows:
• For cell 1B = -5
• For cell 1C; = -4
• For cell 3C; = -5
• For cell 3A; = -1
• For cell 4A; = +3
• For cell 4B; = -2
 Calculating opportunity cost for
Improved solution 2 we get…..
v1 v2 v3 v4

Project A B C Availability

u1 1 2 (50) 7 (-5) 4 (-4) 50

u2 2 3 (20) 3(20) 1(40) 80

u3 3 5 (-1) 4(70) 7(-5) 70

u4 4 1 (+3) 6 (-2) 2(140) 140

Demand 70 90 180 340

Cell 4A is still have +3 as OC, hence this needs to be

improved, by tracing a closed path from this cell and
re-allocating the supplies to get nest improved
solution.
 Calculating opportunity cost for
Improved solution 2 we get…..
v1 v2 v3 v4

Project A B C Availability

u1 1 2 (50) 7 (-5) 4 (-4) 50

u2 2 3 (20) _ 3(20) 1(40) + 80

u3 3 5 (-1) 4(70) 7(-5) 70

u4 4 1 (+3) + 6 (-2) 2(140) _ 140

Demand 70 90 180 340

20 at 2A cell is the smallest quantity at –ve

corner, this will be added and subtracted to +ve
and –ve corners to get new solution.
 Improved solution 3
v1 v2 v3 v4

Project A B C Availability

u1 1 2 (50) 7 4 50

u2 2 3 3(20) 1(60) + 80

u3 3 5 4(70) 7 70

u4 4 1 (20) 6 2(120) 140

Demand 70 90 180 340

Check this for optimality

 Solving for Optimality cont…
• Let us now calculate OC from all unoccupied cells as
follows:
• Recalculate the values of u1 and V1 from relationship
ui + vj = cij ; assuming u1=0, we get v1=2, v2=5 and v3=3
similarly u2=-2, u3=-1, u4=-1.
• Opportunity cost as follows:
• For cell 1B = -2
• For cell 1C; = -1
• For cell 2A; = -3
• For cell 3A; = -4
• For cell 3C; = -5
• For cell 4B; = -2
 Optimality test for Improved
solution 3
v1 v2 v3 v4

Project A B C Availability

u1 1 2 (50) 7 (-2) 4 (-1) 50

u2 2 3 (-3) 3(20) 1(60) 80

u3 3 5 (-4) 4(70) 7(-5) 70

u4 4 1 (20) 6 (-2) 2(120) 140

Demand 70 90 180 340

All opportunity costs are Negative, hence this is

optimal solution with transportation cost of RS. 760.
 Thank you
Keep your surrounding Clean
and Avoid Pollution