Transportation

F. M. Kapepiso
Learning objectives

At the end of the lecture, you should be able to:

• Explain a transportation problem
• Formulate a transportation problem
• Set up the initial transportation table and make the initial
feasible allocations using various methods
• Calculate optimal solution using various methods of solving
transportation problems
• Solve transportation problems having unequal availability and
requirements
Introduction
 Transportation problems consist of decision making
concerning the transportation (conveyance) or allocation of
goods from various sources to various destinations or
points of demand
 Transportation models deal with minimizing transportation
costs by finding an optimum route for supply of goods from
multiple sources to multiple destinations.
 a step by step iterative process, which starts with initial
feasible solution and each succeeding solution is also
feasible.
 At each stage it is evaluated whether transportation costs
can be reduced to optimize the solution.
 An optimum solution is reached when no further cost
reduction is possible.
Key Terms Used
 Balanced Transportation Problem: a transportation
problem in which total demand is equal to total supply
 Degeneracy: A condition that occurs when the number of
occupied squares in any solution is less than the number of
rows plus the number of columns minus 1
 Dummy Destination: An artificial destination added to a
transportation table when total supply is greater than total
demand.
 Dummy Source: An artificial source added to a
transportation table when total demand is greater than
total supply.
 Facility Location Analysis: An application of the
transportation method to help a firm decide where to
locate a new factory, warehouse, or other facility.
Key Terms Used…
 Improvement Index: The net cost of shipping one unit on a
route not used in the current solution.

 Modified Distribution Method: A technique that is used to

evaluate the empty cells in a transportation problem.

 Northwest Corner Rule: A systematic procedure for

establishing an initial feasible solution to the transportation
problem.

 Transportation Problems: A specific type of LP concerned

with scheduling shipments from sources to destinations so that
total transportation cost is minimized.

 Vogel's Approximation Method (VAM): An algorithm used

to find a relatively efficient initial feasible solution which is often
optimal.
Transportation problem defined
 Is a special type of linear programming problem which
attempts to regulate the transportation of goods from
sources to destinations in such a way that the transport or
total cost of a product complies with certain requirements.
 Every source has its capacity and every destination has a
specific demand.
Formulation of the transportation problem
1) The problem can be formulated as a linear programming
model, consisting of an objective function and subject to
constraints.
2) Linear programming can be solved quickly by means of a
computer, hence it will not be covered in this chapter.
3) Secondly, the problem can be formulated in a tabular form. This
chapter concentrate on this.
4) Consist of cells for each possible allocation from the
sources to the destination points.
5) The cost of transporting one unit from the source to the
destination concerned is shown in a square in the upper-
right hand corner.
Initial feasible solution
Once the information is tabulated, the initial feasible solution
can be found, by applying the following methods.
1. Northwest corner method
2. Lowest cost entry method
3. Vogel approximation method
Initial feasible solution…
Steps using- Northwest corner (NWC) method and lowest
cost method
1. Begin by placing as many units as possible in the upper left-
hand corner for NWC and start at the lowest cost as
possible for lowest cost entry method.
2. Utilize the full capacity in a row before moving to the next
row
3. Fully satisfy the demand in column before moving to the
next column
4. Ensure that total demand and capacity agree
Initial feasible solution…
EXAMPLE 1
Ministry of health operates three blood banks in Namibia which are
centrally managed from the Windhoek office. The blood banks supply blood
to four of state run hospitals located at Rundu, Swakop, Tsumeb and
Windhoek. The blood banks are located at Katima, Luderitz and Mariental
where volunteers donate blood that is stored under prescribed standard
conditions and transported to various hospitals as per demand.
The cost of shipping a standard container of blood from each bank to
hospital is given below in a matrix (shaded area). You are also given the
number of blood units available and demand of units at various hospitals
for a normal month.
Required: How many blood units should be shipped from each
blood bank to various hospitals so that the transportation cost
is minimized?
Initial feasible solution…
EXAMPLE 1…

NB: Last column and last row values are quantity while all other values
in the table are cost (N$). We are dealing with a balanced problem
Initial feasible solution (Lowest cost entry)
Initial feasible solution (Lowest cost entry)

Total transport cost

Initial feasible solution…
EXAMPLE 2

Required: Find the initial solution using all three methods

so that the transportation cost is minimized?
Vogel approximation method (VAM)
 This method takes costs into account, the initial solution is
closer to an optimal solution
Steps for solving transportation using VAM
 Calculate the difference between lowest and second lowest
cost in each row and column
 Select the largest difference for allocation
 Place as many units as possible in a cell with the lowest cost
in a row or column
 Where more than one row/ column have same largest
difference, the allocation is made to the cell with lowest cost
in a row or column
 Repeat the procedure until all the available capacity has been
allocated and demand fully satisfied
Vogel approximation method (VAM)…
Example
To explain application of VAM, let me use the previous example
1. The blood bank – hospital case is reproduced here.
Vogel approximation method (VAM)…
Initial Feasible Solution

To Hospitals Supply
From Rundu Swakop Tsumeb Windhoe
k
Marienta 50 50
l
Luderitz 70 10 80
Katima 40 30 50 120
Demand 90 70 40 50 250
Vogel approximation method (VAM)…

Occupied Route No of units Cost per unit Total cost

shipped
MR 50 40 2000
LS 70 35 2450
LT 10 25 250
KR 40 70 2800
KT 30 30 900
KW 50 35 1750
10150
Testing the solution for possible improvement

 Using the initial solution as point of departure, the optimal

solution can be determined by one of two methods, namely
◦ The stepping stone method or
◦ Modified distribution method.

Stepping Stone Method:

 Under this method each unused route is evaluated to see
whether it can reduce the overall cost. Since it is time
consuming we will not use this method.
Modified Distribution Method:
 A technique that is used to evaluate the empty cells in a
transportation problem.
 In an attempt to reduce the large number of calculations
by determining a relative value in respect of each row
and each column.
Steps for using modified distribution method
Set transportation cost in a linear equation for occupied cells
as Cij = Ri + Kj where
C = cost of that cell
R = row heading which indicates source of supply
K = column heading which indicates destination or
demand center.
After writing equations for all occupied cells set value of first
source of supply equal to zero.
Solve all equations and calculate row values and column
values
Add the row value and column value to the table
Modified Distribution Method…
Find improvement index for unoccupied cells using the formula:
Actual cost – (corresponding row value +
corresponding column value) or Cᵢⱼ - Rᵢ - Kⱼ
Select the best largest negative index and identify
corresponding unoccupied cell.
Modify supply for the closed path. For this we have to put a +
sign in identified cell and put a – sign in corresponding cell to
meet the constraints. We put + or – sign in appropriate cells till
the demand – supply constraints are balanced.
This procedure is repeated until none of the cells has a
negative. A negative index indicates that the current solution
can be improved upon.
Modified Distribution Method…
Using the previous initial solution of VAM in example 1 on slide 18 to find
the optimal solution by modified method. Cost equation for occupied routes:
R1 + K1 = 40
R2 + K2 = 35
R2 + K3 = 25
R3 + K1 = 70
R3 + K3 = 30
R3 + K4 = 35
To simplify calculations an arbitrary value of nil (0) is allocated to R1. We calculate
values for each variable in the above equations., by substituting values already found.
R1 = 0
K1 = 40 (40 – 0)
R3 = 30 (70 – 40)
K3 = 0 (30 – 30)
K4 = 5 (35 –30)
R2 = 25 (25 – 0)
K2 = 10 (35 – 25)
Modified Distribution Method…
Improvement Index
Modified Distribution Method…
Improvement Index

The above improvement index indicates that rerouting units to

L - R route (cell) will reduce total shipment cost.

According to procedure explained above we have to put a +

sign in L + R cell and – sign in appropriate cell and complete
this procedure of re routing.

Rerouting is done in the previous table which is reproduced

here for convenience. (See next slide)
Modified Distribution Method…
Transportation table

Looking at the above +, - signs and quantities involved we have

to reroute 10 units to L  route at the same time maintaining
demand and supply requirements. Accordingly the following
solution table will be prepared. Shaded area represents
movement of units from one to other destination.
Modified Distribution Method…
Second Solution Table
Modified Distribution Method…
Now we can calculate total transportation cost according to
new shipment plan.
Modified Distribution Method…
It is clear that re routing will save N$ 50 in transportation cost.
(10150 – 10100)
Note that saving in transportation cost is equal to number of units
rerouted times highest negative improvement index(10x-5)
Now we have to repeat step 2 to ensure whether the above
solution is optimum. You will find that most of values in this step are
going to remain unchanged.
Transportation cost for occupied cells
R1 + K1 = 40
R2 + K2 = 35
R2 + K1 = 60 (new route)
R3 + K1 = 70
R3 + K3 = 30
R3 + K4 = 35
Modified Distribution Method…
To simplify calculations an arbitrary value of nil (0) is allocated to R1. We calculate
values for each variable in the above equations., by substituting values already
found.
R1 = 0
K1 = 40 (40 – 0)
R3 = 30 (70 – 40)
K3 = 0 (30 – 30)
K4 = 5 (35 –30)
R2 = 20 (60 – 40)
K2 = 15 (35 – 20)
Modified Distribution Method…
Improvement Index

The above improvement index indicates that rerouting will not

reduce total shipment cost as none of the Improvement
indexes is negative, hence; the above solution is optimum.
Degeneracy and Transportation:

If you look at initial and further solution tables you will

observe that in each case the number of allocations made
is 1 less than the number of rows added to the number of
columns (R + K – 1) or (3 + 4 – 1 = 6). However, at times
number of allocations may be less than rows + columns - 1. This
condition is known as degeneracy. It happens when a source of
supply has capacity which equals demand at any destination. In
such situations it is necessary to make one or more zero
allocations to routes to bring up the number of allocations to
(R + K – 1). This means that, for shadow cost calculation
purposes, one or more cells with zero allocation are treated as
occupied
Example 3

Hi-Fi Corporation has 3 warehouses in Namibia where

electronics goods are stored before delivery to customers.
Recently Hi –Fi Corporation has received orders for supply
of 150 special configuration personal computers from four
municipalities namely A, B, C and D. There are 150 such
special configuration personal computers available in the
three warehouses of the company located at different places.
The management wishes to minimise delivery costs by
dispatching personal computers (PC) from the appropriate
warehouse to each customer.
Availability – Demand and transportation cost for
PCs at Hi – Fi Corporation

Customers
Availability
Warehouse A B C D (PCs)

X 130 110 150 200 20

Y 170 140 120 130 60
Z 180 180 150 140 70
Demand
30 20 40 60 150
(PCs)
Step 1: Develop an initial feasible solution (Lowest cost)

Ware Customers Availabilit

Houses A B C D y (PCs)
X 20 20
Y 40 20 60
Z 30 40 70
Demand
30 20 40 60 150
(PCs)
Initial Feasible Solution after Zero Allocation

Ware Customers Availability

Houses A B C D (PCs)
X 0 20 20
Y 40 20 60
Z 30 40 70
Demand
30 20 40 60 150
(PCs)
 Now we can calculate total transportation cost for the
above shipment plan.

Route No of units Cost per Total cost

shipped unit
XA 0 130 0
XB 20 110 2200
YC 40 120 4800
YD 20 130 2600
ZA 30 180 5400
ZD 40 140 5600
20600
Step 2:Testing the solution for possible improvement
Transportation cost of occupies cells
X + A = 130
X + B = 110
Y + C = 120
Y + D = 130
Z + A = 180
Z +D = 140

Setting value of X = 0; values of other variables will be as

follows:
X=0 A = 130
B = 110 Y = 40
C= 80 D = 90
Z = 50
 Improvement Index
Unoccupied Actual cost Shadow Difference
Route cost (Row (Improveme
value + nt Index)
Column
value)
X+C 150 0 + 80 = 80 70
X+D 200 0 + 9 0= 90 110
Y+A 170 4 0+ 130 = 0
170
Y+B 140 4 0+ 110 = - 10
150
Z+B 180 5 0+ 110 = 20
160
Z+C 150 50 + 80 = 20
130
 Re Routing

Ware Customers Avail

house abilit
A B C D y
(PCs)
(20) -
X 0 + 130 150 200 20
110
+ (20) -
Y 170 60
140 120(40) 130
(30) -
Z 150 (40) +140
180 180 70
Deman
30 20 40 60 150
d (PCs)
In the above table shaded area represent rescheduling of
shipment. The number of units that will be moved is 20.
 Second Solution

Ware Customers Availabilit

Houses A B C D y (PCs)
X 20 0 20
Y 20 40 60
Z 10 60 70
Demand
30 20 40 60 150
(PCs)
 Now we can calculate total transportation cost according
to new shipment plan.

Route No of units Cost per Total cost

shipped unit
XA 20 130 2600
YB 20 140 2800
YC 40 120 4800
ZA 10 180 1800
ZD 60 140 8400
20400

It is clear that re routing will save N$ 200 in transportation

cost. (20600 – 20400)
 Improvement Index
Unoccupied Actual cost Shadow cost Difference
Route (Row value + (Improveme
Column nt Index)
value)
X+C 150 0 + 90 = 90 60
X+D 200 0 + 90 = 90 110
Y+A 170 30 + 130 = 10
160
Y+D 130 30 + 90 = 120 10
Z+B 180 50 + 110 = 20
160
Z+C 150 50 + 90 = 140 10
Degeneracy and Subsequent Solutions:

Sometimes when 2 or more cells have a negative sign and compete

for closed path movements, it may cause degeneracy at later stage
of solution. To deal with such problem we place a zero in previously
occupied cell having the least cost.

Two or More optimal Solutions:

if any of the improvement indices is 0. It indicates that supply
can be rearranged on such routes without affecting total cost
and provides management greater flexibility to choose
alternative routes .
Transportation and Maximization:

Sometimes a transportation problem may indicate profit (per

unit transported) on a particular route. In such cases the
objective function becomes maximizing profit as against
minimizing transportation costs. Since an improvement index
shows further improvement in solution, in maximization
problems optimum solution is reached when all the indices are
zero or negative.
Prohibited Routes:
Assign a very high cost so that it never enters the solution mix

Unequal Availability and Requirements

Availability >Requirements
 A dummy destination, with zero transport costs, is inserted
in the table to absorb the surplus available.
 The units shipped to such dummy destination are actually
not shipped at all.
Example 4 (Dummy Destination)

To Retail Outlets Availabl

From A B C D e
Quantity
Wareho 300 1000 900 200 80
use X
Wareho 100 900 300 800 60
use Y
Wareho 400 500 200 600 100
use Z
240
200
Deman 50 50 80 20
d
As you can observe from the above table the demand and
supply are not balanced (demand < supply)hence, we have to
insert a dummy retail outlet with a total demand of 40 washing
machines (240 – 200) and zero transportation costs.

After inserting a dummy destination the above table will look

like the following:

To Retail Outlets Availabl

From A B C D E e
Quantity
X 300 1000 900 200 0 80
Y 100 900 300 800 0 60
Z 400 500 200 600 0 100
Deman 50 50 80 20 40 240
d
 Initial Feasible Solution (Lowest cost entry)

To Retail Outlets Available

From A B C D E Quantity
X 20 20 40 80
Y 50 10 60
Z 20 80 100
Dema 50 50 80 20 40 240
nd
 Quantity Cost per
Route Shipped unit Total Cost
X+B 20 1000 20000
X+D 20 200 4000
X+E 40 0 0
Y+A 50 100 5000
Y+B 10 900 9000
Z+B 20 500 10000
Z+ C 80 200 16000
64000
ACTIVITY: Evaluate whether the above solution is optimum. If
not, find optimum solution.

To Retail Outlets Availa

From A B C D E ble
Quant
ity
X 20 20 40 80
Y 30 30 60
Z 50 50 100
Dema 50 50 80 20 40 240
nd
 Qty Cost per
From To Shipped unit Total cost
X A 20 300 6000
X D 20 200 4000
X E 40 0 0
Y A 30 100 3000
Y C 30 300 9000
Z B 50 500 25000
Z C 50 200 10000
57000
Now we calculate improvement indices again and you will find
that none of these indices in negative now, hence, the above
solution is optimum.
Example 5 (Dummy Source):

Hi – Fi Corporation has 3 assembly plants located at three

places namely X, Y and Z where it assembles its popular brand
of music systems. Hi – Fi sells these music systems to retail
customers through its 3 retail outlets situated at Windhoek,
Swakop and Oshakati. The following table provides details of
monthly capacity of three assembly plants, demand at various
retail outlets and transportation cots.
 To Retail Outlets Plant
From Windhoek Swakop Oshakati Capacity
Plant X 60 40 90 200
Plant Y 100 50 80 175
Plant Z 120 70 60 75
450
500
Demand 250 100 150
After inserting a dummy source row the above table will look
like the following.

To Retail Outlets Plant

From Capacity
Windhoek Swakop Oshakati
Plant X 60 40 90 200
Plant Y 100 50 80 175
Plant Z 120 70 60 75
Dummy 0 0 0 50
Plant D
 Initial Feasible Solution (lowest cost entry)

To Retail Outlets Plant

From Windhoek Swakop Oshakati Capacity
Plant X 100 100 200
Plant Y 100 75 175
Plant Z 75 75
Dummy 50 50
Plant D
Demand 250 100 150 500
The cost of above initial feasible solution is given in the
following table.

Initial Solution Cost

Shipment Cost per
From To (Qty) unit Total cost
X W 100 60 6000
X S 100 40 4000
Y W 100 100 10000
Y Os 75 80 6000
Z Os 75 60 4500
D W 50 0 0
500 30500
 Final Optimum Solution

Retail Outlets Plant

Windhoek Swakop Oshakati Capacity
Plant X 200 200
Plant Y 100 75 175
Plant Z 75 75
Dummy 50 50
Plant D
Demand 250 100 150 500
The total cost of above optimum shipment plan is given in the
following table which is N$ 3000 less (30500 – 27500) than
the previous solution.

Cost of Optimum Solution (Hi Fi Corporation)

Cost per Shipment
From To Shipment unit cost
X W 200 60 12000
Y S 100 50 5000
Y Os 75 80 6000
Z Os 75 60 4500
D W 50 0 0
500 27500
Now we have to calculate improvement indices again and you
will find that none of these indices in negative, hence, the above
solution is optimum.

However, it should be noted that Windhoek outlet will be 50

units short of the given demand.

END OF CHAPTER
Practice question
Angel Air conditioning manufacturers produce room air conditioners
at plants located in Khomas, Hardap and Omusati regions, these are
sent to regional distributors in Windhoek, Rundu and Swakop. The
shipping costs vary and the company would like to find the least cost
way to meet the demands at each of the distribution centers.
Windhoek needs to receive 800 air conditioners per month, Rundu
need 600 and Swakop needs 200. Khomas plant has 850 air
conditioners available each month, Hardap plant has 650 air
conditioners each month, and Omusati plant has 300 air
conditioners available each month. The shipping cost per unit from
Khomas to Windhoek is N$ 80, to Rundu is N$ 120, and to Swakop
is N$ 120. The cost per unit from Hardap to Windhoek is N$ 100 to
Rundu is N$ 140 and to Swakop is N$ 90. The cost per unit from
Omusati to Windhoek is N$ 110, to Rundu is N$ 85 and to Swakop
is N$ 120.
Required:
 How Many units should be shipped from each plant to each
regional distribution center? (Use Vogel`s approximation method)
and determine how much is the total cost for this shipment.
Thanks