4th Edition

N. D. Vohra

Quantitative
Techniques in
Management

© 2010
Chapter 5

Specially Structured
Linear Programmes I:
Transportation and
Transhipment Problems
Contents
1. Statement of the Transportation
Problem
2. Solution to the Problem
a) Simplex Method
b) Transportation Method
3. Transportation Method
a) Initial Feasible Solution
North-West Corner Rule
Least Cost Method
Vogel’s Approximation Method
b) Testing the Optimality
Stepping Stone Method
MODI Method
c) Obtaining an Improved Solution
Contents
(…continued)
4. Special Topics
Unbalanced Problems
Prohibited Routes
Unique vs. Multiple Optimal
Solutions
Degeneracy
Maximization Problems
5. Dual of the Transportation Problem
6. Sensitivity Analysis
7. Transshipment Problem
A Typical Transportation
Problem
Inputs:
Sources with availability
Destinations with requirements
Unit cost of transportation from
various sources to destinations

Objective:
To determine schedule of
transportation to minimise total
transportation cost
Transportation Problem

Simplex Tableau: Optimal Solution

Destinations
From
To Supply
P Q R S

A 12 10 12 13 500
Source

B 7 11 8 14 300
C 6 16 11 7 200
Demand 180 150 350 320 1,000
Transportation Problem

Number of sources = 3
Number of destinations = 4
There are 12 routes available,
represented by 12 cells
Aggregate demand = Aggregate
supply = 1000 units
This is a balanced transportation
problem
For a problem, if Aggregate
demand ≠ Aggregate supply, the
problem is called unbalanced
transportation problem
LP Formulation of
Transportation Problem
Min Z = 12 x11 + 10 x12 + 12 x13 + 13 x14 +
7 x21 + 11 x22 + 8 x23 + 14 x24 + 6 x31 + 16x32
+ 11 x33 + 7 x34
Subject to
x11 + x12 + x13 + x14 = 500
x21 + x22 + x23 + x24 = 300
x31 + x32 + x33 + x34 = 200
x11 + x21 + x31 = 180
x12 + x22 + x32 = 150
x13 + x23 + x33 = 350
x14 + x24 + x34 = 320
xij ≥ 0 i = 1, 2,3 and j = 1,2,3,4

There are 12 decision variables and 7

constraints. One of the constraints is
Transportation Method
Step 1
Balance the problem if AD and AS
are unequal; place an M in the cost
cell if some route is prohibited;
and convert into equivalent min
problem if it is a max problem
Step 2
Obtain initial solution
Methods:
North-West Corner Rule:
considers only demand and
availability
Least Cost Method: considers unit
cost, demand and availability
Vogel’s Approximation Method:
considers cost differences, demand
and availability
Initial Solution: NWC Method
TC=Rs. 10,220
Destinations
From
To Supply
P Q R S
180 150 170
A 12 10 12 13 500
Source

180 120
B 7 11 8 14 300
200
C 6 16 11 7 200
Demand 180 150 350 320 1,000

Allocation Sequence
1. Allocate 180 units on route AP
2. Move to cell AQ and allocate 150 units
3. Move to cell AR and allocate 170 units
4. Move to cell BR and allocate 180 units
5. Move to cell BS and allocate 120 units
6. Finally, allocate 200 units to cell CS
Initial Solution: Least Cost
Method TC=Rs. 9,620
Destinations
From
To Supply
P Q R S
150 50 300
A 12 10 12 13 500
Source

300
B 7 11 8 14 300
180 20
C 6 16 11 7 200
Demand 180 150 350 320 1,000

Allocation Sequence
1. Allocate 180 units on route AP
2. Select least cost cell CP, allocate 180 units
3. Delete column P, adjust supply at C to 20
4. Select cell CS (least cost among the undeleted
cells) and allocate 20 units. Delete the row C
and adjust demand at S as 300
5. Select cell BR, assign 300 units to it. Delete the
row B and adjust demand at R to 50
6. Finally, Allocate 150 units to AQ, 50 to AR and
300 to AS
Initial Solution: VAM or
Penalty Method TC=Rs. 9,440
Destinations
From
To Supply
P Q R S
150 230 120
A 12 10 12 13 500
Source

180 120
B 7 11 8 14 300
200
C 6 16 11 7 200
Demand 180 150 350 320 1,000

VAM involves obtaining and using cost

differences

For every row and every column, obtain

the difference between two least cost
values

Select largest of the cost differences and

choose the least cost cell in that
row/column
Initial Solution: VAM or
Penalty Method (…continued)
Allocate the quantity equal to lower of
the supply and demand values
Delete the row or column that is
satisfied
Repeat the steps until all
supply/demand requirements are met
with
Allocations are:
Cost
Step Cell Qty. Adjustment
Diff.
Delete row C
1 6 CS 200
Revise col S 120
Delete col P
2 5 BP 180
Revise row B 120
Delete row B
3 4 BR 120
Revise col R 230
4 Allocate 150 units to AQ, 230 to AR and 120 to AS
Transportation Method (…
continued)
Step 3
Test of optimality: MODI Method
First of all, check if the number of
occupied cells = m + n – 1, where
m = number of columns
n = number of rows
If yes, proceed to assign ui values to the
rows and vj values to the columns, and
calculate ∆ij values for unoccupied cells
Assign arbitrarily a value to u or v. You
can assign any one ui or vj value. Let u1 =
0
Since u1 = 0, select an occupied cell in
row 1, find vj value such that ui + vj = cij.
With u1 = 0, for AQ, v2 =10; for AR, v3
=12 and for AS, v4 =13
MODI Method (…
continued)
Using v3 =12 choose BD and u2 = -4,
since 12 – 4 = 8 (cost)
With u2 = -4, choose BP, assign v1 =11
Finally, with v4 = 13, and the occupied
cell CS, assign u3 = -6, as 13 – 6 = 7
Having assigned ui and vj values,
consider all unoccupied cells one by
one, add up corresponding row and
column (ui and vj ) values and subtract
there from the cost element, cij. Now, Δij
= ui + vj – cij
What do Δij values indicate?
If all ∆ij values are ≤ 0, it is optimal solution
If all ∆ij values are negative, it is unique
optimal solution
If some ∆ij value/s = 0 while others are
negative, there are multiple optimal
Testing Optimality: MODI
Method
P Q R S SS ui
150 230 120
A 12 10 12 13 500 0
180 120
B 7 11 8 14 300 -4
120
C 6 16 11 7 200 -6

DD 180 150 350 320 1,000

vj 11 10 12 13

Δ11 = -1; Δ22 = -5; Δ24 = -5; Δ31 = -1; Δ32

= -12; and Δ33 = -5

Since all values are negative, this

solution is optimal and unique
Transportation Method (…
continued)
Step 4
If some Δij > 0, the solution is not
optimal. To improve a non-optimal
solution:
a) Begin with cell having largest ∆ij, draw a
closed loop (or path):
Move alternately between rows and
columns
Stop only at occupied cells
Start with a + sign in the cell of origin, place
– and + signs alternately on cells on the
path
b) Consider cells with ‘–’ sign, choose the
least quantity in them; add it to each
cell with ‘+’ sign and subtract it from
each cell with ‘–’ sign
c) Obtain revised solution and go back to
step 3
Closed Loop
A closed loop:
Ends in the same cell where it
begins
Drawn through only occupied cells
except the starting cell
May be drawn clock-wise or anti
clock-wise
Always involves an even number of
cells – involving equal number of
cells with ‘–’ and ‘+’ signs
Involves only horizontal and
vertical movements
Does never involve diagonal
movement
Can take any shape
Drawing of a Loop
Initial Solution - LCM
P Q R S SS ui
150 50 300
A 12 10 12 13 500 0
0 + -
300
B 7 11 8 14 300 -4
1 -5 - -5
180 20
C 6 16 11 7 200 -6
- +
-12 -5
1,00
DD 180 150 350 320
0

vj 12 10 12 13

Loop: BP – BR – AR – AS – CS – CP – BP

Shall involve transferring 180 units

Drawing of a Loop
(…continued)
Occupied Cells

1 2 3 4 5

1 -  Start of
Loop

2 + -

4 + -

5 - +

The loop can take any shape

The loop can also cross over itself
Degeneracy

Present when number of occupied

cells is smaller than m + n – 1

Prevents testing optimality of the

solution

Is removed by placing an
infinitesimally small value ε in each
of the required number of
independent cells (An independent
cell is one beginning from which a
closed loop cannot be drawn)
Degenerate Solution

1 2 3 4 5 SS
30 10
A 6 4 9 1 0 40

20 20
B 20 6 11 3 0 40

50
C 7 1 0 14 0 50

60 30
D 7 1 12 6 0 90

DD 90 30 50 30 20 220

Not independent Independent

cell cell
Dual of a Transportation
Problem
A transportation problem can be
expressed as an LPP:
Decision variables: quantities
shipped on various routes
Constraints: demand and availability
values
Objective function co-efficients: cost
elements

As such, its dual can be written

The ui and vj values in the optimal

solution indicate optimal values of
the dual variables
Dual of the Transportation
Problem
Maximise Z = 500u1 + 300u2 + 200u3
+180v1 + 150v2 + 350v3 + 320v4

Subject to
u1 + v1 ≤ 12 u2 + v1 ≤ 8
u1 + v2 ≤ 10 u2 + v2 ≤ 14
u1 + v3 ≤ 12 u3 + v3 ≤ 6
u1 + v1 ≤ 13 u3 + v1≤ 16
u2 + v2 ≤ 7 u3 + v2 ≤ 11
u2 + v3 ≤ 11 u3 + v3 ≤ 7

u1, u2, u3, v1, v2, v3,v4 : unrestricted in

sign
Sensitivity Analysis

Deals with:
Changes in cost element of a non-
basic variable (an unoccupied cell)

Changes in cost element of a basic

variable (an occupied cell)

Changes in a particular source

supply and a particular destination
demand by a given amount, k

Effects of setting upper/lower

limits on the values of xij variables
Production Planning
Problem as TP
Key Data about an item:
Can be produced in regular time
(R) or overtime (O)
Unit cost of production: R: Rs 6 and
O: Rs 10
Production capacity: R: 60 units
and O: 20 units
Carrying cost: Rs 3/unit/month
Demand in Months 1, 2, 3 and 4:
50, 70, 85, and 75 units
respectively
(continued…)
Production Planning
Problem as TP
Destination
Origin Supply
1 2 3 4 D
R 6 9 12 15 0 60

O 10 13 16 19 0 20

R M 6 9 12 0 60

O M 10 13 16 0 20

R M M 6 9 0 60

O M M 10 13 0 20

R M M M 6 0 60

O M M M 10 0 20

Demand 50 70 85 75 40 320
Transhipment Problem

Allows for shipment from one

source to another and from one
destination to another

An m-origin n-destination
transportation problem becomes a
m+n origin and m+n destination
transhipment problem

With minor modifications, problem

is solved using transportation
method

Transhipment allows lowering total

transportation cost
Multiple Choice Questions
A transportation problem is
balanced when

1. Total availability (TA) and Total

Demand (TD) are equal, and
number of sources is equal to
number of destinations.

2. TA and TD are equal, irrespective

of the number of sources and
destinations.

3. Number of sources matches with

number of destinations.

4. None of the routes is prohibited.

Multiple Choice Questions
A solution to a transportation
problem is degenerate when:

1. the number of unoccupied cells is

less than r + c – 1 (r: rows, c:
columns).

2. The number of unoccupied cells is

equal to r + c – 1.

3. The number of occupied cells is at

least r + c – 1.

4. The number of occupied cells is

less than r + c – 1.
Multiple Choice Questions
Mark the wrong statement:

1. An unbalanced transportation
problem can be converted into a
balanced transportation problem
through the addition of an
appropriate slack variable.
2. In North-West Corner Rule, first
allocation is always made by
beginning in the upper-left hand
corner of the tableau.
3. The North-West Corner Rule
provides a systematic but inefficient
method of finding initial solution to
a transportation problem.
4. It is necessary to make number of
sources and destinations equal
before applying N-W Corner Rule.
Multiple Choice Questions
The cost elements of a row in a
problem are: [32 18 24 18]. In
applying VAM, what cost
difference will be considered for
this row?

1. 6

2. 0

3. 8

4. 14
Multiple Choice Questions
Mark the wrong statement:

1. The Least Cost Method always gives

the least cost solution to a
transportation problem.
2. In the Penalty Method, a cost
difference indicates the penalty of
not using the least-cost route.
3. In case there is a tie in the maximum
penalty, either of them can be
selected for making allocation.
4. In comparison with the North-West
Corner Rule, both, the Least-Cost
Method and Vogel’s Approximation
Method, are better because they
provide “near optimal” initial
solution.
Multiple Choice Questions

To remove degeneracy, an  is
placed in an unoccupied cell which

1. has the least cost value.

2. has the largest cost value.

3. is an independent cell, beginning

with which a closed path cannot be
drawn.

4. is an independent cell, beginning

with which a closed path can be
drawn.
Multiple Choice Questions

Which of the following conditions

is not required to be satisfied by ,
used to remove degeneracy:

1. k   = 0

2. k -  = k

3.  -  = 0

4. k   = k
Multiple Choice Questions

Mark the correct statement:

1. For drawing a closed path,
movement has always to be clock-
wise, never anti clock-wise.

2. If all ui and vj values are known, one

and only one closed path can be
drawn starting with a given
unoccupied cell.

3. Any of the r + c – 1 number of

occupied cells would allow
determining whether a given
solution is optimal or not.

4. Degeneracy can arise only in initial

solution to a problem.
Multiple Choice Questions
Mark the wrong statement:
1. The number of occupied cells
involved in a closed path is always
even.

2. A closed path need not be a

square/rectangular and can have
any configuration.

3. The maximum quantity, which can

be reallocated in a closed path, is
equal to the minimum quantity in
the cells bearing negative sign.

4. The ij = ui + vj – cij value of an

unoccupied cell indicates the net
change in cost of reallocating one
Multiple Choice Questions
Which of the following is not true
for a transhipment problem?
1. A transhipment problem allows for
shipment of goods from one source to
another as also from one destination
point to another.
2. An m-origin n-destination
transportation problem, when
converted into a transhipment
problem, would become an m  n-
origin and an equal number of
destinations problem.
3. There is no real distinction between
sources and destinations in a
transhipment problem.
4. A transhipment problem is likely to
involve a lower cost than a
transportation problem, in a given
situation.