(GC302) Decision Science

Transportation Models

Assureshift.in

Dr. Sachin Vyavhare

Associate Professor
[email protected]
 Learning Objectives
• Concept of Transportation Models
• Formulation of Transportation Model
• Terminology of Transportation Model
Examples :
• Initial Basic Solution a) NWCM ;b) LCM c) VAM
• Optimal Solution MODI Method

[email protected]
 Concept of Transportation Model

It is a typical Operations Research techniques aims to find a least

cost route for the transportation of goods from the sources (company’s
plant) to the destination (warehouse/distribution
[email protected]
centre.
Mathematical Formulation of
Transportation Model
Destination (Warehouse)

Sources W1 W2 ………… Wn Supply

(Factory)

F1 C11 C12 ………… C1n S1

F2 C21 C22 ………… C2n S2
………… ………… ………… ………… ………… …………
Fm Cm1 Cm2 ………… Cmn Sm
Demand d1 d2 ………… dn N=Total
Supply/
Demand

[email protected]
Let Xij , represents the number of units to be transported from ith source
(i=1,2,…,m) to the jth destination ( j=1,2,…,n); then the total cost of
transportation is give by ion
Objective function:

[email protected]
 Terminology of Transportation Model

• Balanced Transportation Problem

• Unbalanced Transportation Problem

• Transportation Table:
It represents the data about the supply at
source , demand at destination and per unit
cost from each source to each destination

[email protected]
Terminology of Transportation Model

• Dummy Source or destination:

• Adding an extra row / column to the
transportation table with “0” cost per unit . It
is used for balancing supply and demand
• Initial Feasible solution: NWCM/LCM/VAM
• The solution that satisfy the demand and
supply condition (m+n-1)
• Optimum Solution: MODI (Modified
distribution Method)
• The solution with minimum total cost
[email protected]
 General procedure for solving a Transportation Problem
Step 1: Define the objective function to be minimized with the constraints
imposed on the problem.
Step 2: Set up the transportation table with m rows representing the sources
and n columns epresenting the destinations along with their supply and
demand figures as well as the unit cost figures in the cells. Add dummy row or
column, if necessary, to balance the problem.
Step 3: Develop the Initial Feasible Solution to the problem using either
NWCM, LCM or VAM.
Step 4: Examine the feasibility of the initial solution. The solution is feasible if
it has allocations at (m+n-1) number of hells at independent positions i.e. it
has (m+n-1) independent occupied cells.
Step 5: Test the solution, if found feasible, for optimality by finding out the
opportunity costs for all the unoccupied or empty cells. Non-Negative (i.e.
zero or positive) opportunity cost for all empty cells indicates that the
solution is optimal.
Step 6: If the solution is not optimal, we improve it by using the Stepping
Stone Method or the MODI method.
Step 7: Repeat steps 4, 5 and 6 until an optimal solution is obtained.
[email protected]
 Methods for Finding the Initial
Feasible Solution
(A)North West Corner Method (NWCM)
(B)Least Cost Method (LCM) or Matrix
Minima Method
(C)Vogel's Approximation Method (VAM)

[email protected]
 (A) North West Corner Method (NWCM)
Step 1: Select the North West corner cell in the transportation
table and allocate as many units as possible to it after checking
the supply (in row) and the demand (in column) position for that
cell.
Step 2: Reduce the supply and demand figures for the
corresponding row and column accordingly.
Step 3: Cover the row or column where the supply or demand
gets fully exhausted ( i.e. becomes 0) to get a reduced
transportation table.
Step 4: Go to step 1 and repeat the procedure until total supply
is fully allocated to the cells so as to fulfill the total demand.
Note: Though it is easy, this method requires many further
improvements to obtain the optimum solution.

[email protected]
Case I- Transportation Problem
(Initial Basic Solution)
a) North-West Corner Method

[email protected]
[email protected]
 N

[email protected]
Since the number of positive allocation (occupied cell)=m+n-1
i.e. 3+4-1 = 6 Allocation
The Initial Basic Solution by NWCM Method is as follows

Source (Factory) Destination Qty*Cost/unit

(Warehouse)
F1 W1 5*19
F1 W2 2*30
F2 W2 6*30
F2 W3 3*40
F3 W3 4*70
F3 W4 14*20
Total =Rs.1015
Transportation
Cost

[email protected]
(B) Least Cost Method (LCM) or Matrix Minima
Method
Step 1: (i) Select a cell with minimum unit transportation cost
from the table.
(ii) If there are more than one cells with minimum unit cost i.e.
there is a tie, then select that cell among them where more
number of units can be allocated (after considering their row
supply and column demands.) If there is a tie again, then select a
cell randomly from them.
Step 2: Allocate maximum possible number of units to it.
Reduce the corresponding supply and demand figures
accordingly to get a reduced transportation table as in case of
NWCM.
Step 3: Repeat steps 1 and 2 until entire supply is exhausted to
fulfill the entire demand.
[email protected]
Case I- Transportation Problem
(Initial Basic Solution)
b) Least Cost Method

[email protected]
Since the number of positive allocation (occupied cell)=m+n-1
i.e. 3+4-1 = 6 Allocation
The Initial Basic Solution by LCM Method is as follows

Source (Factory) Destination Qty*Cost/unit

(Warehouse)

F1 W4 7*10
F2 W1 2*70
F2 W3 7*40
F3 W1 3*40
F3 W2 8*8
F3 W4 7*20
Total Transportation =Rs.814
Cost

[email protected]
 (C) Vogel's Approximation Method (VAM)
Step 1: Compute the penalty (ie. the difference between the two
smallest unit cost figures of the cells) for each row and column.
Step 2: Identify the row or column with highest penalty and choose
the cell with smallest unit cost in it.
(If there is a tie for highest penalties, select the row or column
containing the minimum cost cell, among them. If there is a tie
again, then select that cell where maximum allocation is possible
or we can select it randomly).
Step 3: Allocate maximum possible units to the selected cell and
reduce its row supply and column demand accordingly. Obtain the
reduced table then as done previously.
Step 4: Recompute the penalties for the reduced table. [If a row
(column) has been covered the row (column) penalties remain
unchanged]. Repeat the above procedure until the entire demand
and supply gets [email protected]
 Case I- Transportation Problem
(Initial Basic Solution)
c) Vogel’s Approximation Method
Row Penalties

Row
Penalties

[email protected]
 Case I- Transportation Problem
(Initial Basic Solution)
c) Vogel’s Approximation Method

[email protected]
Since the number of positive allocation (occupied cell)=m+n-1
i.e. 3+4-1 = 6 Allocation
The Initial Basic Solution by LCM Method is as follows

Source (Factory) Destination Qty*Cost/unit

(Warehouse)

S1 D1 5*19
S1 D4 2*10
S2 D3 7*40
S2 D4 2*60
S3 D2 8*8
S3 D4 10*20
Total Transportation =Rs.779
Cost

[email protected]
Q.2) Find the initial solution for the following
transportation problem using (i) NWCM (ii) LCM
and (iii) VAM. The supply, demand and unit cost
figures are given. (May-June-2024)

W1 W2 W3 W4 Supply
P1 190 300 500 100 70
P2 700 300 400 600 90
P3 400 100 400 200 180
Demand 50 80 70 140

[email protected]
 Method for Finding the Optimum Solution
 Once the initial solution is obtained, we verify it for
feasibility by finding the number of occupied cells.
 If the number of occupied cells (at independent positions)
is equal to (number of rows + number of columns -1 i.e. m
+ n -1) then the solution is feasible, otherwise it is called as
a case of degeneracy (described further in detail).
 Thus, If the solution is feasible we improve it further using
the Stepping Stone Method or Modified Distribution
Method to obtain the optimum solution. As the Modified
Distribution Method is more convenient to use, only this
method has been described here.

[email protected]
 A) Modified Distribution Method (MODI) or UV method:
Step 1: Determine the Initial Basic Feasible Solution and verify that it is feasible i.e. it has
(m+n-1) independent occupied cells.
Step 2: Determine the row numbers (denoted as ui i.e. u1, u2, u3, etc.) and the column
numbers (denoted as vj i.e. v1, v2,v3, etc.) by using the formula: Cij = ui + vj for each
occupied cell (Cij is the corresponding per unit transportation cost). Choose here one of
the values of ui or vj as 0 arbitrarily (prefer that row or column which contains maximum
number of occupied cells), and then use the formula for finding out the remaining value
one by one.
Step 3: Now consider the unoccupied cells and find the opportunity cost (improvement
index) for each of them by using the formula, Opportunity cost dij = Cij- (ui + vj) for these
cells.
Step 4: Check the sign of all the opportunity costs. If all of them are non-negative (i.e.
zero or positive) then it implies that the solution obtained is optimum. If at least one of
the opportunity costs is negative, then it implies that the solution is not optimum and
there is a scope for improvement. Then go to step 5.
Step 5: Select the unoccupied cell with highest negative opportunity cost. (If there are
more than one such cells then select the cell where more units can be shifted, as
described in further steps).

[email protected]
Step 6: For this cell, trace a closed path using most direct route through at least 3
occupied cells and then back to the unoccupied cell. Use only vertical and horizontal lines
and take turns only at the occupied cells.
Step 7: Assign (+ve) and (-ve) signs alternately to each of the corner cells along the closed
path, starting with a (+ve) sign for the selected unoccupied cell.
Step 8: Find the maximum number of units to be shifted to this unoccupied cell. (These
are equal to the least of the number of units in the cells with -ve signs). Add this number
of units to the cells with (+ve) sign and subtract it from the cells with (-ve) sign along the
closed path.
Step 9: Go to Step 1 and repeat the procedure until an optimum solution is obtained.
Note:
(i) The initial feasible solution must have at least one allocation in each of its rows and
columns.
(ii) The allocations must be made in independent positions (i.e. the positions of cells
Where a closed path cannot be traced as explained above) and must be (m + n - 1) in
number.
(iii) The Opportunity cost (dij) for an unoccupied cell indicates the change in the total cost
if one Unit is shifted to these cells. Hence a (-ve) opportunity cost indicates saving in the
cost and hence we select the unoccupied cell with highest (-ve) opportunity cost.
[email protected]
 Case II- Transportation Problem
(Minimization)
Solve the transportation problem and
Obtain Optimal solution

[email protected]
PART 1-Initial Basic Solution by VAM

[email protected]
[email protected]
Factory Warehouse Qty * cost /unit=TC

F1 W1 5*19=95
F1 W4 2*10=20
F2 W3 7*40=280
F2 W4 2*60=120
F3 W2 8*8=64
F3 W4 10*20=200
Total Transportation cost Rs. 779

[email protected]
PART 2-Optimal Solution by MODI
(Modified Distribution Method)

# Step 1- Determine the [email protected]

row numbers(ui-u1,u2…) and column number
(vj-v1,v2….) by using the formula cij=ui+vj for all occupied cells
# step 2-For all occupied cell
Take v4=0 because there are THREE allocation in column (W4)

F1W1c11=u1+v1 19=10+v1 v1=9

F1W4c14=u1+v4 10=u1+0 u1=10

F2W3 c23=u2+v3 40=60+V3 v3= -20

F2W4 c24=u2+v4 60=u2+0 u2=60

F3W2 c32=u3+v2 8=20+v2 v2= -12

F3W4 c34=u3+v4 20=u3+0

[email protected]
u3=20
# step 3 For all unoccupied cell
Find opportunity cost as dij = cij - (ui + vj)

F1W2d12 = c12 - (u1 + v2) =30-(8-12) = +32

F1W3d13 = c13 - (u1 + v3) =50-(-20+10) = +60

F2W1 d21 = c21 - (u2 + v1) =70-(9+60) = +1

F2W2 d22 = c22 - (u2 + v2) =30-(-12+60) = -18

F3W1 d31 = c31 - (u3 + v1) =40-(20+9) = +11

F3W3 d33 = [email protected]

- (u3 + v3) =70-(-20+20)= +70
# step 4: Since all the dij values are NOT Non-
Negative. There is scope for improvement of
optimal solution.

[email protected]
# step 5- Trace a close loop from cell F2W2 using horizontal and vertical lines and
taking turn ONLY at occupied Cell. Assign (+) sign and (-) sign alternatively to the
corner cells along path . Note that starting with (+Sign) for occupied cell at P2W2.

+ --

-- +

[email protected]
# step 6- Find the minimum no. of units to be shifted to P2W2 , the cell with (-) Sign
along the closed path is P3W2 AND P2W4 which contain 2 and 8 units respectively .
The least quantity is 2 units to be shifted to P2W2 cell in closed path. ADD 2units at
(+) sign and DEDUCT 2 from (-sign).

[email protected]
The improved solution as follows

[email protected]
Take u1=0 because there are Two allocation in row (F1)

F1W1c11=u1+v1 19=0+v1 v1=19

F1W4c14=u1+v4 10=0+v4 v4=10

F2W2 c22=u2+v2 30=u2-2 u2= 32

F2W3 c23=u2+v3 40=u2+32 u2=8

F3W2 c32=u3+v2 8=10+v2 v2= -2

F3W4 c34=u3+v4 20=u3+10

[email protected]
u3=10
For all unoccupied cell
Find opportunity cost as dij = cij - (ui + vj)

F1W2d12 = c12 - (u1 + v2) =30-(0-2) = +32

F1W3d13 = c13 - (u1 + v3) =50-(0+32) = +18

F2W1 d21 = c21 - (u2 + v1) =70-(19+8) = +43

F2W4 d24 = c24 - (u2 + v4) =60-(10+8) = +42

F3W1 d31 = c31 - (u3 + v1) =40-(19+10) = +11

F3W3 d33 = [email protected]

- (u3 + v3) =70-(32+10)= +28
All dij (Opportunity cost) values are non-
negative . Hence there is no further scope for
reducing transportation cost. The optimum
solution as follows:
Factory Warehouse Qty * cost /unit=TC

F1 W1 5*19=95
F1 W4 2*10=20
F2 W2 2*30=60
F2 W3 7*40=280
F3 W2 6*8=48
F3 W4 12*20=240
Total Transportation cost Rs. 743

[email protected]
[email protected]
 Destination (Warehouse)
Source W1 W2 W3 W4 Supply
(Factory)

F1

F2

F3

Demand

[email protected]
I II III IV V

[email protected]
Special Cases in TP

[email protected]
 (A) Unbalanced Problem
If the total supply from all the sources is not equal to the total demand at all
the destinations, it is an unbalanced transportation problem. It is balanced by
introducing a dummy row or column as follows: .

(i) If total demand is greater than total supply, add a dummy row or source
with its supply equal to the difference between total demand and total
supply. Consider the per unit transportation cost for all its cells as zero. The
quantity supplied from this dummy source to various destinations in the
optimum solution indicates the unfulfilled demand at these destinations.

(ii) If total supply is greater than total demand, add a dummy column or
destination with its demand equal to the difference between total supply and
total demand. Consider the per unit cost for all its cells as zero. The quantity
allotted to this dummy destination indicates the unused capacity at the
respective sources.

Solve further the problem as an usual balanced transportation problem.

[email protected]
 B) Maximization Problem:
Transportation problem method can be used for a maximization (say
profit maximization) problem, by converting it into an equivalent
minimization problem as follows:
(i) Write down the profit table (or profit matrix) with given supply,
demand and per unit profit figures.
(ii) Add a dummy source or destination to it (with '0’ unit profit for its
cells) if necessary
(iii) Locate the largest per unit profit figure in the table and subtract
all profit figures (including itself) from it to get an equivalent relative
loss matrix.
(iv) Solve it further as a normal transportation problem to get the
optimum solution table.
(v) Rewrite the optimum table by replacing the per unit cost figures by
the original profit figures and then find the maximum profit for the
optimum plan.
[email protected]
Example 8: Solve the following T.P. to maximise
profit and give the criteria for optimality.

Destination
Origins
1 2 3 4 Supply
A 40 25 22 33 100
B 44 35 30 30 30
C 38 38 28 33 70
Demand 40 20 60 30

[email protected]
 Step:1: Add dummy

Destination
Origins
1 2 3 4 Dummy Supply

A 40 25 22 33 0 100

B 44 35 30 30 0 30

C 38 38 28 33 0 70

Demand 40 20 60 30 50 200

[email protected]
Step:2 : Obtain Relative Loss Matrix

Origin Destination
s
1 2 3 4 Dummy Supply

A 4 19 22 11 44 100

B 0 9 14 14 44 30

C 6 6 16 11 44 70

Demand 40 20 60 30 50 200

[email protected]
(iii) Now, we can solve further by using (VAM+MODI) method to
get the optimum plans as follows:

A-> 1 (20), A -> 4 (30), A->5 (50), B->1 (20), B->3 (10), C->2 (20),
and C->3 (50).

However, for calculating the total profit we must consider the

original profit figures for these cells. Thus, we get,

Total profit = 20 (40) + 3O (33) + 0 + 20 (44) + 10 (30) + 20 (38) +

50 (28) = 5130
Again here 50 units at A will be actually unutilized (being
supplied to dummy destination '5')

[email protected]
 (C) Case of Degeneracy
This is said to occur when the number of independent occupied cells (at any
solution stage) is less than (m + n- 1). This is resolved as follows:
(i) For degeneracy at the initial solution: Assign an artificial quantity ε
(epsilon) to one or more of the unoccupied cells, (depending upon the
difference in the number of occupied cells and m+n-1) at the independent
positions with lowest unit costs. ε remains until the final solution is obtained
or the degeneracy is removed, whichever occurs first. (Note that ε is an
infinitesimally small positive quantity, so that it does not affect the demand
and supply constraints and is just used to occupy a cell temporarily so as to
enable further operations).

(ii) Degeneracy during further stages of solution: This occurs when the
reallocation or shifting of units to an unoccupied cell (in MODI method)
vacates two or more occupied cells simultaneously. This is resolved by
allotting ε to one or more of the recently vacated cells along the closed path,
having lowest unit transportation costs. If there is a tie, assign ε arbitrarily.
[email protected]
Example : Solve the following Transportation
Problem to minimize the total cost.

Origins Destinations Availability

D1 D2 D3
O1 40 70 90 300
O2 12 80 30 400
O3 60 90 45 200
Requirement 300 300 300

[email protected]
Solution:
Total supply=Total demand=900
the problem is balanced
Step1: Now, IBFS by VAM,

[email protected]
Step2: Thus, the initial solution has 4 independent occupied cells which is
less than (m + n- 1 = 3 + 3- 1 = 5) by 1. Thus there is a degeneracy. Hence we
allocate quantity ε to an independent unoccupied cell with smallest cost i.e.
to cell O2D1
(Here we can not draw a closed for this cell, hence it is independent, unlike
cell O3D3).
Thus, we have the initial solution with the degeneracy removed as follows.

Now, m+n-1=5, its feasible.

[email protected]
Step3: Apply MODI method for improvement

For occupied cell calculate Cij= ui+vj

Assume u2=0
C11= u1+v1
C21= u2+v1
C22= u2+v2
C23=u2+v3
C32= u3+v2

[email protected]