Outline of the Handout

EE 477: Optimization Techniques 1. Introduction to the Transportation Problem

2. Formulation of the Transportation Problem

Handout #6 3. Balanced and Unbalanced Transportation Problems

4. Examples of Transportation Problems

The Transportation Problem 5. Determination of a Starting Basic Feasible Solution

6. Special Cases of the Transportation Problem

Prof. Mohamed Zribi 7. Solution of the Transportation Problem

8. Iterative Computations of the Transportation Algorithm

Updated 28 September 2017
9. Solved Transportation Problem Examples
1 2
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

The Transportation Problem

One of the first applications of linear programming

1. Introduction to the was the problem of minimizing the cost of
transporting materials. Problems of this type are
Transportation Problem referred to as transportation problems (TPs).

3 4
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Introduction to the TP The Transportation Model

The transportation problem is a special type of linear The transportation model is a special class of LPPs that
programming problems in which goods are deals with transporting (=shipping) a commodity from
transported from a set of sources to a set of sources (e.g. factories) to destinations (e.g. warehouses).

destinations subject to the supply and demand of the

The objective is to determine the shipping schedule that
source and the destination respectively, such that the
minimizes the total shipping cost while satisfying supply
total cost of transportation is minimized.
and demand limits. We assume that the shipping cost is
proportional to the number of units shipped on a given
route.
5 6
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Transportation Problem
Introduction to the TP • Basically a transportation problem is one in which you are
– The transportation Problem is a distribution-type problem asking:
in which supplies of goods that are held at various locations
are to be distributed to other receiving locations.
• How many of what should go where?
– How many goods should go from which factory to which
– The solution of a transportation problem will indicate to a warehouse?
manager the quantities and costs of various routes and the
resulting minimum cost. – How many goods should go from which warehouse to
which store?
– Inventory: How many goods should be made in month i to
– The transportation Problem is used to compare location
be available in month j? (i.e. how many goods should go
alternatives in deciding where to locate factories and
warehouses to achieve the minimum cost distribution from month i to month j?)
configuration. – Can you think of another example?
7 8
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Formulating the Model of a TP Formulating the Model of a TP
Transportation models deal can be formulated as a standard LP problem. Typical
– A transportation problem typically involves a set of situation shown in the following manufacturer example.
sending locations, which are referred to as sources, and a A manufacturer has three plants P1, P2, P3 producing same products.
set of receiving locations, which are referred to as From these plants, the product is transported to three warehouses
destinations. W1, W2 and W3.
Each plant has a limited capacity, and each warehouse has specific
– To develop a model of a transportation problem, it is demand. Each plant transport to each warehouse, but transportation
necessary to have the following information: cost vary for different combinations.

1. Supply quantity (capacity) of each source, The problem is to determine the quantity each warehouse in order to minimize
total transportation costs.
2. Demand quantity of each destination.
3. Unit transportation cost for each origin-destination
route.

9 10
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

The Transportation Model A schematic Representation of TP

Characteristics a1 S1 D1 b1

• A product is transported from a number of sources to a number a2 D2 b2

of destinations at the minimum possible cost.
•Each source is able to supply a fixed number of units of the
product, and each destination has a fixed demand for the product.
Si Dj
• The linear programming model has constraints for supply at each ai bj
source and constraints for demand at each destination.
• All constraints are equalities in a balanced transportation model
where supply equals demand. bn
Dn
• Constraints contain inequalities in unbalanced models where am Sm
supply does not equal demand. m- number of sources n- number of destinations
11 12
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 A schematic Representation of TP
A schematic Representation of TP
Sources Destinations

Supply a1 1 Demand b1
1

Supply a2 2
2 Demand b2

ai amount available at origin i

…

…
bj amount required at destination j
xij
Demand bn
xij amount to be shipped from origin i to
Supply am n
m
destination j
Costs cij cij cost of shipping one unit from origin i to
13 destination j 14
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 1: The Transportation Problem

• The problem of finding the minimum-cost distribution of a

Sources Destinations given commodity from a group of supply centers (sources)
factories, Markets i=1,…,m to a group of receiving centers (destinations)
finished goods warehouses , Finished goods ware house j=1,…,n
raw materials ware houses, raw materials ware houses, • Each source has a certain supply (ai)
suppliers etc. factories,
• Each destination has a certain demand (bj)
• The cost of shipping from a source to a destination is directly
proportional to the number of units shipped

15 16
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 General Description of a TP Example 2: Schematic of a Transportation Problem

1. A set of m supply points from which a good is shipped.

Supply point i can supply at most ai =si units.
2. A set of n demand points to which the good is shipped.
Demand point j must receive at least bi =di units of the
shipped good.
3. Each unit produced at supply point i and shipped to
demand point j incurs a variable cost of cij.

17 18
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 3: Transportation Problems

Find a minimal cost of transportation. • Since transportation problems are special, we

Supply Demand Transportation Cost formulate them in a different way.
Demand • Rather than write out all the constraints as in a
1 2
5 1 1 2 3 4 5 si
traditional LP (which involve a lot of zero
2 12 1 6 2 1 3 4 5
10 2 coefficients)
2 9 1 1 6 7 10
Supply

3 5
8 3 3 2 3 4 5 5 8 • We simply formulate them in table form.
4 9 4 1 6 3 8 2 7
7 4 dj 2 12 5 9 2 30
5 2
19 20
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Destinations
1 2 j n Supply Example 4:
c c12 c1j c1n a1
S 1 11 a2
O 2 Desti- Desti- Desti- Desti- Supply
U nation 1 nation 2 nation 3 nation 4
R
C Source 1 c11 c12 c13 c14 a1
E ai
S i ci1 ci2 cij cin Source 2 c21 c22 c23 c24 a2

Source 3 c31 c32 c33 c34 a3

m
cm1 cm2 cmn am Demand b1 b2 b3 b4
Demand b1 b2 bj bn
21 22
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Transportation Problem Example 5:

Parameter Table

These are the only Dest Dest Dest Dest Supply

1 2 3 4
parameters in a
Source 1 c11 c12 c13 c14 a1
transportation problem.
Source 2 c21 c22 c23 c24 a2
cij : the cost to move from
Source 3 c31 c32 c33 c34 a3
source i to destination j
Demand b1 b2 b3 b4
ai: the supply available at
source i
bj: the demand at
destination j.

23 24
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Decision variables
Number of millions energy (kWh) produced from sources and
sent to demands

xij = amount to be shipped from source i to destination j

i = number of sources
j = number of demands

sources demands
25 26
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Objective function Constraints

Minimize cost Constraints of number of Supply and demand
xij ≥ 0 (i = 1,2,3) and ( j=1,2,3,4)
cost =
Constraints of Supply

Constraints of Demands

j = number of demands

27 28
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 LP Formulation of Powerco’s Problem Example 6:

Min Z = 8x11+6x12+10x13+9x14+9x21+12x22+13x23+7x24
+14x31+9x32+16x33+5x34 - Problem: how many tons of wheat to transport from each grain elevator to each mill on
a monthly basis in order to minimize the total cost of transportation ?

S.T.: x11+x12+x13+x14 ≤ 35 (Supply Constraints) - Data: Grain Elevator Supply Mill Demand
1. Kansas City 150 A. Chicago 200
x21+x22+x23+x24 ≤ 50
2. Omaha 175 B. St.Louis 100
x31+x32+x33+x34 ≤ 40 3. Des Moines 275 C. Cincinnati 300
x11+x21+x31 ≥ 45 (Demand Constraints) Total 600 tons Total 600 tons
x12+x22+x32 ≥ 20
x13+x23+x33 ≥ 30 Transport cost from Grain Elevator to Mill ($/ton)
Grain Elevator A. Chicago B. St. Louis C. Cincinnati
x14+x24+x34 ≥ 30 1. Kansas City $6 8 10
2. Omaha 7 11 11
xij ≥ 0 (i= 1,2,3; j= 1,2,3,4) 3. Des Moines 4 5 12

29 30
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

minimize Z = 6x1A + 8x1B + 10x1C + 7x2A + 11x2B + 11x2C + 4x3A + 5x3B + 12x3C
Solution:
subject to x1A + x1B + x1C = 150
x2A + x2B + x2C = 175
You have 3 DCs, and need to deliver product to 3 customers.
x3A + x3B+ x3C = 275
x1A + x2A + x3A = 200
CH 200 x1B + x2B + x3B = 100
x1C + x2C + x3C = 300
KC 150
xij 0

O 175 St 100

DM 275
CI 300
Network of transportation routes for wheat shipments

where xij = tons of wheat from each grain elevator,

i, i = 1, 2, 3, to each mill j, j = A,B,C
Find the cheapest way to satisfy all demand?
31 32
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Example 7:
• Transportation problems are solved manually within a tableau
format. A computer manufacturing company has two assembly
• Each cell in a transportation tableau is analogous to a decision plants, plant A and plant B, and two distribution outlets,
variable that indicates the amount allocated from a source to a
destination. outlet I and outlet II. Plant A can assemble at most 700
computers a month, and plant B can assemble at most
900 computers a month. Outlet I must have at least 500
computers a month, and outlet II must have at least 1,000

The Transportation Tableau

computers a month.

33 34
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Transportation costs for shipping one computer from each Solution:

plant to each outlet are as follows: $6 from plant A to outlet I;
To form a shipping schedule, we must decide how many
$5 from plant A to outlet II: $4 from plant B to outlet I; $8
computers to ship from either plant to either outlet. This will
from plant B to outlet II.
involve 4 decision variables:

Formulate the transportation problem so that the total cost of x1 = number of computers shipped from plant A to outlet I

shipping the computers from the assembly plants to the x2 = number of computers shipped from plant A to outlet II
distribution outlets.
x3 = number of computers shipped from plant B to outlet I

x4 = number of computers shipped from plant B to outlet II

35 36
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Thus, we must solve the following linear programming problem:
The constraints are as follows:
x1 + x2 < 700 Available from A Minimize C = 6x1 + 5x2 + 4x3 + 8x4

x3 + x4 < 900 Available from B subject to

x1 + x3 > 500 Required at I
x1 + x 2 < 700 Available from A
x2 + x4 > 1,000 Required at II
x3 + x4 < 900 Available from B

Total shipping charges are: x1 + x3 > 500 Required at I

Cost = C = 6x1 + 5x2 + 4x3 + 8x4 x2 + x4 > 1,000 Required at II

Before we can solve this problem, we must multiply the first two
constraints by -1 so that all are of the > type.
37 38
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

1 1 0 0 700
The problem can now be stated as: 0 0 1 1 900
A 1 0 1 0 500
Minimize C = 6x1 + 5x2 + 4x3 + 8x4
0 1 0 1 1, 0 0 0
subject to 6 5 4 8 1
-x1 - x2 > -700
- x3 - x4 > -900 1 0 1 0 6
1 0 0 1 5
x1 + x3 > 500
A T
0 1 1 0 4
x2 + x4 > 1,000 0 1 0 1 8
x1, x2, x3, x4 > 0 700 900 500 1, 0 0 0 1

39 40
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Introduce slack variables x1, x2, x3, and x4 to form the initial system
for the dual:
The dual problem is;
-y1 + y3 + x1 =6
Maximize P = -700y1 - 900y2 +500y3 + 1,000y4
-y1 + y4 + x2 =5
subject to
-y2 + y3 + x3 =4
-y1 + y3 <6
-y2 + y4 + x4 =8
-y1 + y4 < 5
-700y1 - 900y2 +500y3 + 1,000y4 +P = 0
-y2 + y3 <4
-y2 + y4 < 8
If we form the simplex tableau for this initial system and solve, we
y1, y2, y3, y4 > 0 find that the shipping schedule that minimizes the shipping charges
is 0 from plant A to outlet I, 700 from plant A to outlet II, 500 from
plant B to outlet I, and 300 from plant B to outlet II. The total
shipping cost is $7,900.
41 42
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

The Transportation Problem

m resources, n destination

2. Formulation of the – ai：number of units supplied by source i

– bj：number of units required by destination j
– cij：transportation cost per unit shipped from source i
Transportation Problem to destination j
– Objective：minimize total transportation cost
– xij：number of product shipped from source i to
destination j
43 44
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 The Transportation Problem
The Transportation Problem We assume that there are m sources 1,2, …, m and n destinations
1, 2, …, n. The cost of shipping one unit from Source i to
Cost per unit Distributed Destination Destination j is cij.
Supply
1 2 3 n
We assume that the availability at source i is ai (i=1, 2, …, m) and
1
c11 c12 c13 c1n a1
2
c21 c22 c23 c2 n a2
the demand at the destination j is bj (j=1, 2, …, n). We make an
important assumption: the problem is a balanced one. That is
source
m n

m cm1 cm 2 cm 3 cmn am
ai bj
i 1 j 1
Demand b1 b2 b3 bn
That is, the total availability equals the total demand.
45 46
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Thus the transportation problem is such that: The Transportation Problem

m n
Minimize z c ij x ij
A problem’s argument An answer
i 1 j 1 Demand Demand
1 2 n s1 1 2 n s1
subject to: m n 1 c11 c12 c1n s1 1 x11 x12 x1n s1

n Minimize cij xij 2 c21 c22 c2 n s2 2 x21 x22 x2 n s2

Supply
Supply
x ij a i (i 1, 2 ,..., m ) i 1 j 1 m
dj
cm1 c42
d1 d 2
cmn
dn
sm
T
m
dj
xm1
d1
x42
d2
xmn
dn
sm
T

j 1 n

Subject to xij si i 1, ,m
m j 1

x ij bj ( j 1, 2 ,..., n ) m

i 1
xij dj j 1, ,n
i 1

x ij 0 47
xij 0 i 1, , m and j 1, ,n
48
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Remark:
The Transportation Problem

Thus there are m n decision variables xij and m+n constraints. Though we can solve the above TP by the Simplex

Since the sum of the first m constraints equals the sum of the last n
method, we will solve it by a special algorithm called
constraints (because the problem is a balanced one), one of the the transportation algorithm.
constraints is redundant and we can show that the other m+n-1
constraints are linearly independent.

We present the data in an m n tableau as explained

Thus any BFS will have only m+n-1 nonzero variables. below.

49 50
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Destination
1 2 . . n Supply
c11 c12 c1n
1 a1
S
o c21 c22 c2n
u
2 a2
3. Balanced and Unbalanced
r .
c . Transportation Problems
e
m cm1 cm2 cmn
am
Demand b1 b2 bn

51 52
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Balanced Transportation Problem Example 8: Balanced Transportation Problem

If the total supply equals to the total demand,

City 1 City 2 City 3 City 4 Supply
then the problem is said to be a balanced
8 6 10 9
Plant 1 35
transportation problem: 10 25
m n 9 12 13 7
ai b Plant 2 50
j 45 5
i 1 j 1
14 9 16 5
Plant 3 40
10 30
Demand 45 20 30 30 125
53 54
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 9: Balanced Transportation Problem Types of Transportation Problems

Source Destination Supply • Balanced transportation problems

1 2 3 m n

1 20 10 15 200 ai b j
i 1 j 1

2 10 12 9 300
• Unbalanced transportation problems
3 25 30 18 500 m n

Demand a i b j
i 1 j 1
200 400 400

55 56
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Balancing a TP if total supply exceeds
Types of Transportation Problems
total demand
• Key point in a transportation problem:
– Supply has to equal demand. • If total supply > total demand,
m n
we add a dummy demand point.
ai b j (Balanced Transportation Problem)
i 1 j 1

• When there isn’t equality: • Since shipments to the dummy demand point are
– We can use dummy demands (or supplies) in order capture not real, they are assigned a cost of zero.
the excess supply (or demand) to get equality.
– And then solve the problem.

57 58
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 10: TP with Excess Supply Solution: Add dummy demand

Ware Ware Ware Ware Output Ware Ware Ware Ware Unused Output
house 1 house 2 house 3 house 4 house 1 house 2 house 3 house 4 Capacity
Source 1 464 415 654 867 125 Source 1 464 415 654 867 125

Source 2 352 416 690 791 125 Source 2 352 416 690 791 125
Source 3 995 682 388 685 125
Source 3 995 682 388 685 125
80 65 70 85
80 65 70 85
We have 75 units of excess supply, so we create a dummy demand (this
Supply = 125 * 3 = 375; Demand = 80 + 65 + 70 + 85 = 30059 really represents supply capacity that won’t be used) 60
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Example 11: TP with Excess Supply

Ware Ware Ware Ware Unused Output

house 1 house 2 house 3 house 4 Capacity
Source 1 464 415 654 867 0 125
Source 2 352 416 690 791 0 125
Source 3 995 682 388 685 0 125
80 65 70 85 75

When supply exceeds demand, a dummy column is added to the tableau.

We have 75 units of excess supply, so we create a dummy demand (this
really represents supply capacity that won’t be used) 61 62
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 13:
Example 12: Demand Less than Supply
D1 D2 D3 Supply
Customer Dummy S1 3 7 8 21
1 Customer 2 Factory
Capacity S2 9 4 6 16
Factory 1 8 5 0 S3 10 11 5 13
170
Demand 18 7 15

Factory 2 15 10 0
130 Supply = 50, Demand=40 => Add a dummy destination

Factory 3 3 9 0 D1 D2 D3 D4 Supply
80 S1 3 7 8 0 18
S2 9 4 6 0 7
Customer S3 10 11 5 0 15
Requirements 150 80 150 380
Demand 21 16 13 10
63 64
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Balancing a transportation problem if Example 14: TP with Excess Demand
total supply is less than total demand
Ware Ware Ware Ware Output
house 1 house 2 house 3 house 4
• If total supply < total demand (no feasible
Source 1 464 415 654 867 75
solution) one or more of the demand will be left
unmet. Source 2 352 416 690 791 125

• a penalty cost is often associated with unmet Source 3 995 682 388 685 100
demand
80 80 80 80

65 Supply = 300 Demand = 320 66

EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Solution: Add dummy supply Example 15: Total demand ≠ Total supply

Ware Ware Ware Ware Output

house 1 house 2 house 3 house 4

Source 1 464 415 654 867 75

Source 2 352 416 690 791 125

Source 3 995 682 388 685 100

Unused 0 0 0 0 20 Note that, the total demand=650, and the total supply = 600
storage How to solve it?
Allocation 80 80 80 80
We need to add a dummy row and assign o cost to each cell as such ..
67 68
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Example 16: TP with Excess Demand

Extra row, since Demand > supply When demand exceeds supply a dummy row is added to the tableau.
69 70
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 18:
Example 17: Supply Less than Demand
D1 D2 D3 Supply
Customer Customer Customer Factory S1 3 7 8 18
1 2 3 Capacity S2 9 4 6 7
Factory 1 8 5 16 S3 10 11 5 15
170
Demand 21 16 13

Factory 2 15 10 7
130 Supply= 40, Demand=50 => Add a dummy source

0 0 0 D1 D2 D3 Supply
Dummy 80 S1 3 7 8 18
S2 9 4 6 7
Customer S3 10 11 5 15
Requirements 150 80 150 380
S4 0 0 0 10
71
Demand 21 16 13 72
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Remark: The Balanced Transportation Problem:
represented as an LP model
Quantities in dummy routes in the optimal m n
M inim ize : Z = c ij x ij
solution are not shipped and serve to indicate i =1 j=1

which supplier will hold the excess supply, and n

subject to x ij = a i i = 1, 2, ...., m
how much, or which destination will not j=1

receive its total demand, and how much it will m

x ij = b j j = 1, 2, ....., n
be short. i =1

x ij 0 for i = 1, ...m and j = 1, ..n

73 74
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 19:
MG Auto has three plants in Los Angeles, Detroit, and New
Orleans, and two major distribution centers in Denver and

4. Examples of Miami. The capacities of the three plants during the next quarter
are 1000, 1300 and 1200 cars. The quarterly demands at the two
distribution centers are 2300 and 1400 cars. The transportation
Transportation Problems cost per car from Los Angeles to Denver and Miami are $80 and
$215 respectively. The corresponding figures from Detroit and
New Orleans are 100, 108 and 102, 68 respectively.

Formulate this transportation problem.

75 76
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Solution:
Destination
Denver Miami Supply
Since the total demand = 3700 > 3500 (Total supply) we
80 215
introduce a dummy supply with availability 3700-3500=200
S Los Angeles 1000
units to make the problem a balanced one.
o 100 108
1300
Detroit
If a destination receives u units from the dummy source, it u
102 68
means that that destination gets u units less than what it r New Orleans 1200
demanded. c 0 0
200
Dummy
e
Demand 2300 1400
We usually put the cost per unit of transporting from a dummy
source as zero (unless some restrictions are there). Thus we get We write inside the (i,j) cell the amount to be shipped from source
i to destination j. A blank inside a cell indicates no amount was
the transportation tableau
shipped.
77 78
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Destination
Example 20:
Denver Miami Supply
In the previous problem, penalty costs are levied at the rate of
80 M
$200 and $300 for each undelivered car at Denver and Miami S Los Angeles 1000
respectively. Additionally no deliveries are made from the Los o 100 108
1300
Detroit
Angeles plant to the Miami distribution center. Set up the u
102 68
transportation model. r New Orleans 1200
Solution: c 200 300
200
Dummy
e
The above imply that the "cost" of transporting a car from the
Demand 2300 1400
dummy source to Denver and Miami are respectively 200 and
300. The second condition means we put a "high" transportation Note: M indicates a very "big" positive number.
cost from Los Angeles to Miami. We thus get the tableau
79 80
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Example 21: Distribution Area
1 2 3
Three refineries with daily capacities of 6,5, and 8 million 1 120 180 -
gallons, respectively, supply three distribution areas with daily Refinery 2 300 100 80
demands of 4,8, and 7 million gallons, respectively. Gasoline
3 200 250 120
is distributed to the three distribution areas through a network
of pipelines. The transportation cost is 10 cents per 1000 Construct the associated transportation model.
gallons per pipeline mile. The table below gives the mileage
between the refineries and the distribution areas. Refinery 1 is
not connected to the distribution area 3.

81 82
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Solution: Destination
Distribution Area Example 22:
S 1 2 3 Supply
o 12 18 M
1 6 In the previous problem, suppose that the daily demand at area
u 30 10 8
3 drops to 4 million gallons. Surplus production at refineries 1
r Refinery 2 5 and 2 is diverted to other distribution areas by truck. The
transportation cost per 100 gallons is $1.50 from refinery 1 and
c 20 25 12
8 $2.20 from refinery 2. Refinery 3 can divert its surplus
3
e production to other chemical processes within the plant.
Demand 4 8 7
Formulate the problem as a transportation model.
The problem is a balanced one. M indicates a very "big"
positive number.
3 3 Solution: We introduce a dummy destination.
The total cost will be 1000* c ij x ij
i 1 j 1 83 84
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Destination
Distribution Area Example 23:
S 1 2 3 Dummy Supply
o 1 12 18 M 15
6
Three orchards supply crates of oranges to four retailers. The
u daily demand at the four retailers is 150,150,400, and 100
30 10 8 22
r Refinery 2 5 crates, respectively. Supply at the three orchards is dictated by
c 3 20 25 12 0 8 available regular labor and is estimated at 150, 200, and 250
e crates daily. However, both orchards 1 and 2 have indicated
Demand 4 8 4 3
that they could supply more crates, if necessary by using
M indicates a very "big" positive number. overtime labor. Orchard 3 does not offer this option. The
transportation costs (in dollars) per crate from the orchards to
3 3 the retailers are given in Table below.
The total cost will be 1000* c ij x ij
i 1 j 1 85 86
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Retailer Destination
1 2 3 4 Retailer
S O 1 2 3 4 Dummy Supply
1 1 2 3 2 o r
1
1 2 3 2 0
350
Orchard 2 2 4 1 2 u c
3 1 3 5 3 h 2 2 4 1 2 0
400
r a
c 1 3 5 3 M
Formulate the problem as a transportation model. r 3 250
e d
Solution: Demand 150 150 400 100 200
Since the orchards 1, 2 can supply more crates with overtime
labor, we increase their capacities to 150+200=350 and
200+200=400 respectively (as initially the total supply fell short
by 200). But then to balance the problem we add a dummy
destination. The tableau follows.
87 88
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Finding a BFS for the TP

In any transportation model, we first determine a starting BFS

5. Determination of a and then iteratively move towards the optimal solution which has
the least shipping cost.

Starting Basic Feasible Unlike other Linear Programming problems, a balanced TP with
m supply points and n demand points is easier to solve, although

Solution it has m + n equality constraints. The reason for that is, if a set of
decision variables (xij’s) satisfy all but one constraint, the values
for xij’s will satisfy that remaining constraint automatically.

89 90
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Determination of the Starting Solution Starting BFS for the Transportation Problem

• Transportation models do not start at the origin where all

Any BFS will have only m+n-1 basic variables (which may decision values are zero; they must instead be given an
assume non-zero = positive values) and the remaining variables initial basic feasible solution.
will all be non-basic and so have zero values.
•There are a number of methods for generating an initial
feasible solution for a transportation problem. We will
In any transportation tableau, we only indicate the values of cover the following 3 methods:
basic variables. The cells corresponding to non-basic - northwest corner method
variables will be blank.
- Lowest cost method

- Vogel’s Approximation Method

91 92
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 1. The North-West Corner Method for
Determining a Starting BFS
Remark 1: The quantities so allocated are circled to indicated, The simplest of the procedures used to generate an initial feasible
the value of the corresponding variable. solution is NWCM. It is so called because we begin with the North
West or upper left corner cell of our transportation table. The various
steps of this method can be summarized as under.
Remark 2: Empty cells indicate the value of the corresponding Step 1:
Select the North West (upper left-hand) corner cell of the
variable as zero, I.e., no unit is shipped to this cell. transportation table and allocate as many units as possible equal to
the minimum between the available supply and demand requirement
i.e., min (S1, D1).

Step 2:

Adjust the supply and demand numbers in the respective rows

93
and columns allocation. 94
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Step 3: Example 24: Find an initial BFS using the NW corner

method for the following transportation problem.
(a) If the supply for the first row is exhausted, then move down to
the first cell in t he second row and first column and go to step
2.
5
(b) If the demand for the first column is satisfied, then move
horizontally to the next cell in the second column and first 6

row and go to step 2.

2
Step 4: 3 5 2 3

If for any cell, supply equals demand, then the next allocation can
be made in cell either in the next row or column.
Step 5:
Continue the procedure until the total available quantity is fully
allocated to the cells as required.
95 96
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Solution: After we check the east and south cells, we see that we can go
east (meaning supply point 1 still has capacity to fulfill some
We can set x11=3 (meaning demand point 1 is satisfied by
supply point 1). demand).

5 3 2 X

6 6

2 2

3 5 2 3 X 3 2 3

3 2 3 2 X

6 3 3

2 2

X 5 2 3 X X 2 3
97 98
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

After applying the same procedure, we see that we can go south

Finally, we will have the following BFS, which is:
this time (meaning demand point 2 needs more supply by supply
x11=3, x12=2, x22=3, x23=2, x24=1, x34=2
point 2).

3 2 X

3 2 1

3 2 X
2

X X X 3 3 2 1 X

3 2 X 2 X

X X X X
3 2 1 X

X X X 2
99 100
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Example 25: Find an initial BFS for the following Solution:
transportation problem. Destination Destination
1 2 3 4 Supply 1 2 3 4 Supply
3 7 6 4 3 7 6 4
1 5 1 52
3 2
Source 2 4 3 2 Source 2 4 3 2
2 2 2 1 1 21

4 3 8 5 4 3 8 5
3 3 3 1 2 3 2

Demand 3 3 2 2 Demand 3 3 1 2 1 2

101 Total shipping cost = 48 102

EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 26: Find an initial BFS using the NW corner

method for the following transportation problem. Example 27: Solve the Transportation Table to find Initial Basic
Feasible Solution using North-West Corner Method.

Total Cost =19*5+30*2+30*6+40*3+70*4+20*14

Destination
= 1015
1 2 3 4 5 Supply ui

1
1
30
1
20
1 2 1
50
D1 D2 D3 D4 Supply
19 30 50 10
S1 7
1 1 1 1 1 5 2
2 0 60 60
source 70 30 40 60
S2 9
1 1 2 2 M 6 3
3 10 30 10 50
40 8 70 20
S3 18
M 0 M 0 0 4 14
4(D) 50 50
Demand 5 8 7 14 34
Demand 30 20 70 30 60
vj
103 104
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Example 28: Find an initial BFS using the NW corner Example 29: Find an initial BFS using the NW corner
method for the following transportation problem. method for the following transportation problem.

- The initial solution is complete when all rim requirements are satisfied.
- Transportation cost is computed by evaluating the objective function:
Z = $6x1A + 8x1B + 10x1C + 7x2A + 11x2B + 11x2C + 4x3A + 5x3B + 12x3C
= 6(150) + 8(0) + 10(0) + 7(50) + 11(100) + 11(25) + 4(0) + 5(0) + 12(275) Total sipping cost = 2250
= $5,925
105 106
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 30: Find an initial BFS using the NW corner

method for the following transportation problem. Solution:
Cust 1 Cust 2 Cust 3 Cust 3 Cust 4 Output
excess Cust 1 Cust 2 Cust Cust 3 Cust 4 Output
3 excess
plant 1 60 plant 1 40 20 60
plant 2 80 plant 2 40 30 10 80
plant 3 40 0 40
plant 3 40
Dummy plant 50 50
Dummy 50
plant Demand 40 60 30 50 50
Demand 40 60 30 50 50

107 108
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Example 31: Find an initial BFS using the NW corner
method for the following transportation problem. Solution:

D1 D2 D3 D4 Supply D1 D2 D3 D4 Supply
S1 5 S1 2 3 5
S2 1 S2 1 1
S3 3 S3 0 2 1 3
Demand 2 4 2 1 9 Demand 2 4 2 1 9

109 110
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 32: Find an initial BFS for the following Solution:

transportation problem using the NW corner method.

111 112
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 2. The Least Cost Method for
Determining the Starting BFS
Step2: Then eliminate the row or column in which supply or
demand is exhausted. If both the supply and demand values
Step1: In this method we start assigning as much as are the same, either of the row or column can be eliminated.
possible to the cell with the least cost (ties are broken In case, the smallest unit cost is not unique, then select the
arbitrarily) and the associated amounts of supply and cell where maximum allocation can be made.

demand are adjusted by subtracting the allocated amount.

Step3: Repeat the process with the next lowest unit cost and
continue until the entire available supplies at various sources
and demands at various destinations are satisfied.
113 114
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 33: Find a starting BFS using the Minimum Step 2: Cross-out column 2
Cost Method

Step 1: Select the cell with minimum cost. 2 3 5 6

5

2 3 5 6 2 1 3 5
5 2
8

2 1 3 5 3 8 4 6
10 15

3 8 4 6 12 X 4 6
15

12 8 4 6

115 116
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Step 3: Find the new cell with minimum shipping cost and cross- Step 4: Find the new cell with minimum shipping cost and cross-
out row 2 out row 1

2 3 5 6 2 3 5 6
5 X
5
2 1 3 5 2 1 3 5
X X
2 8 2 8
3 8 4 6 3 8 4 6
15 15

10 X 4 6 5 X 4 6

117 118
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Step 5: Find the new cell with minimum shipping cost and cross- Step 6: Find the new cell with minimum shipping cost and cross-
out column 1 out column 3

2 3 5 6 2 3 5 6
X X
5 5
2 1 3 5 2 1 3 5
X X
2 8 2 8
3 8 4 6 3 8 4 6
10 6
5 5 4

X X 4 6 X X X 6

119 120
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Step 7: Finally assign 6 to last cell. The BFS is found as: x11=5, Example 34: Find a starting BFS using the Minimum
x21=2, x22=8, x31=5, x33=4 and x34=6 Cost Method

2 3 5 6
X
5
2 1 3 5
X
2 8
3 8 4 6
X
5 4 6

X X X X

121 122
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Total sipping cost = 2065

123 124
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Example 35: Find a starting BFS using the Minimum
Cost Method - The complete initial minimum cell cost solution; total cost = $4,550.
- The minimum cell cost method will provide a solution with a lower cost than
the northwest corner solution because it considers cost in the allocation process.

The Initial Minimum Cell Cost Allocation

The Initial Solution

The Second Minimum Cell Cost Allocation

125 126
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 36: Find a starting BFS using the Minimum Solution:

Cost Method

D1 D2 D3 D4 Supply D1 D2 D3 D4 Supply
2 3 5 6 2 3 5 6
S1 5 S1 5
5
2 1 3 5 2 1 3 5
S2 10 S2 10
2 8
3 8 4 6 3 8 4 6
S3 15 S3 15
5 4 6
Demand 12 8 4 6 30 Demand 12 8 4 6 30
127 128
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Example 37: Find an initial BFS for the following Solution:
transportation problem. Destination Destination
1 2 3 4 Supply 1 2 3 4 Supply
3 7 6 4 3 7 6 4
1 5 1 1 0 2 2 542

Source 2 4 3 2 Source 2 4 3 2
2 2 2 2 2

4 3 8 5 4 3 8 5
3 3 3 3 3

Demand 3 3 2 2 Demand 3 1 3 0 2 2

EE 477 Dr. Mohamed Zribi

129 Total shipping costEE=47736Dr. Mohamed Zribi
130

Example 38: Find an initial BFS for the following Solution:

transportation problem using the least cost method.

131 132
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 3. The Vogel’s Approximation method
If both a row and a column are simultaneously satisfied then
(VAM)
Step 1. For each row (column), determine a penalty by If exactly one row or column is left uncrossed make the
subtracting the smallest unit cost in the row (column) from the obvious allocations and stop. Else cross out one only (either
next smallest unit cost in the same row (column). ( If two unit
the row or the column) and leave a zero supply (demand) in
costs tie for being the smallest unit cost, then the penalty is
0). the uncrossed out row (column). (But omit that row or
column for calculating future penalties).
Step2. Identify the row or column with the largest penalty. Break
ties arbitrarily. Allocate as much as possible to the cell with the
least unit cost in the selected row or column. (Again break the Step 3. If all allocations are made, stop. Else go to step 1.
ties arbitrarily.) Adjust the supply and demand and cross out
the satisfied row or column.

133 134
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Vogel’s Approximation Method Example 39:

1 2
Destination
3 4 Supply Row Penalties
Step1: Calculate penalty for each row and column by taking the
difference between the two smallest unit costs. 3 7 6 4 1 1
1 3 0 0 2 5
S
Step2: Select the row or column with the highest penalty and select the o 2 4 3 2
minimum unit cost of that row or column. Then, allocate the 2 2 2 0 -
minimum of supply or demand values in that cell. If there is a tie, u
then select the cell where maximum allocation could be made. r 4 3 8 5
1 1
3 3 3
Step3: Adjust the supply and demand and eliminate the satisfied row or c
column. If a row and a column are satisfied simultaneously, only one
of them is eliminated and the other one is assigned a zero value. Any e Demand 3 3 0 2 0 2
row or column having zero supply or demand, can not be used in
calculating future penalties.
1 1 3 2
Step4: Repeat the process until all the supply sources and demand Column
destinations are satisfied. Penalties 1 4 - 1

EE 477 Dr. Mohamed Zribi

135
Total shipping cost = 32
EE 477 Dr. Mohamed Zribi
136
Example 40: Vogel’s Method Step 2: Identify the largest penalty and assign the highest
possible value to the variable.
Step 1: Compute the penalties.

Supply Row Penalty Supply Row Penalty

6 7 8 6 7 8
10 7-6=1 5 8-6=2
5

15 80 78 15 80 78
15 78-15=63 15 78-15=63

Demand 15 5 5 Demand 15 X 5

Column Penalty 15-6=9 80-7=73 78-8=70 Column Penalty 15-6=9 _ 78-8=70

137 138
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Step 3: Identify the largest penalty and assign the highest Step 4: Identify the largest penalty and assign the highest
possible value to the variable. possible value to the variable.

Supply Row Penalty Supply Row Penalty

6 7 8 6 7 8
0 _ X _
5 5 0 5 5

15 80 78 15 80 78
15 _ 15 _

Demand 15 X X Demand 15 X X

Column Penalty 15-6=9 _ _ Column Penalty _ _ _

139 140
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Step 5: Finally the BFS is found as x11=0, x12=5, x13=5, and Example 41:
x21=15.
Find an initial BFS for the following transportation problem.

Supply Row Penalty

6 7 8
X _
0 5 5

15 80 78
X _
15

Demand X X X

Column Penalty _ _ _

141 142
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

- VAM allocates as much as possible to the minimum cost cell in the row or column with the - After each VAM cell allocation, all row and column penalty costs are recomputed.
largest penalty cost.

The Second
The Initial VAM AM Allocation
Allocation

143 144
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 - Re-compute penalty costs after the third allocation. - The initial VAM solution; total cost = $5,125
- VAM and minimum cell cost methods both provide better initial solutions than does the
northwest corner method.

The Third VAM

Allocation The Initial VAM
Solution

145 146
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 42: Find a starting BFS using the Vogel’s Method Solution:

Suppl Row
D1 D2 D3
D1 D2 D3 Supply y penalty
6 7 8 6 7 8
S1 10 S1 10 7-6=1

15 80 78 15 80 78
S2 15 S2 15 8-15=63

Demand 15 5 5 25 Demand 15 5 5
Column
15-6=9 80-7=73 78-8=70
Penalty
147 148
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Suppl Row Suppl Row
D1 D2 D3 D1 D2 D3
y penalty y penalty
6 7 8 6 7 8
S1 5 8-6=2 S1 0 —
5 5 5
15 80 78 15 80 78
S2 15 8-15=63 S2 15 —

Demand 15 X 5 Demand 15 X X
Column Column
15-6=9 — 78-8=70 15-6=9 — —
Penalty Penalty
149 150
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Suppl Row Suppl Row

D1 D2 D3 D1 D2 D3
y penalty y penalty
6 7 8 6 7 8
S1 X — S1 X —
0 5 5 0 5 5
15 80 78 15 80 78
S2 15 8-15=63 S2 X —
15
Demand 15 X X Demand X X X
Column Column
— — — — — —
Penalty Penalty
151 152
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Example 43: Find an initial BFS for the following
transportation problem using the Vogel’s method.

D1 D2 D3 Supply
6 7 8
S1 10
0 5 5
15 80 78
S2 15
15
Demand 15 5 5 25

153 154
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

6. Special Cases of the

5

Transportation Problem

155 156
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Degenerate BFS Degenerate BFS

When solving the transportation problem, the number of

possible routes should be m+n-1. If in a cell we find a zero mentioned, it means that that cell
corresponds to a basic variable which assumes a value of
If it is <m+n-1, the BFS is called a degenerate solution. zero.

In order to evaluate the routes, we need to introduce some In simplex language, we say that we have a degenerate BFS.
cell (s) with zero values. That cell should be selected at
the correct place.

157 158
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 44: Degeneracy Example 45: Degeneracy

We have only 4 basic variables => The BFS is not degenerate.

m + n – 1 = 5 the number of cells with allocations
(five basic variables, and above has 5 as well!)
If not matched, then we select an non-basic variable with the least
The BFS is not degenerate.
cost and considered it as a new basic variable with assigned 0 allocation to it
159 160
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Example 46: Degeneracy
Customer Customer Customer Factory
1 2 3 Capacity
Factory 1 5 4 3
100 100

Factory 2 0 8 4 3
100 20 120

Factory 3 9 7 5
80 80
Added this
Note: we pick this over others because it has the least cost for the Min Z problem! Customer
Requirements 100 100 100 300

161 162
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 47: Degeneracy The Minimum Cell Cost Initial Solution To rectify a degenerate tableau, an empty cell must artificially
be treated as an occupied cell.

The Initial Solution

We have only 4 basic variables => The BFS is not degenerate.

163 164
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 The TP and Unacceptable Routes Example 48: Unacceptable Routes

– Certain origin-destination routes may be unacceptable due When one route cannot be used then Assign a big
If 1C Cannot
to weather factors, equipment breakdowns, labor problems, M cost to its cell delivered
Then we assign
or skill requirements that either prohibit, or make M value here

undesirable, certain routes.

– These routes will have a very large associated costs.
– A prohibited route is assigned a large cost such as M.
– When the prohibited cell is evaluated, it will always
contain the cost M, which will keep it from being
selected as an entering variable.
165 166
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

The General Transportation Model

7. Solution of the • Objective function

m n
– Transportation cost: cij xij
i 1 j 1

Transportation Problem • Constraints

n

– Production capacity: xij ai i 1...m

j 1
m

– Demand: xij bj j 1...n

i 1

– Variable limits: xij 0 i 1...m, j 1...n

167 168
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 The Balanced Transportation Model The Transportation Algorithm
m n
min f x cij xij
i 1 j 1 • Since any transportation problem can be formulated as an
s.t.
n
LP, we can use the simplex method to find an optimal
xij ai ; i 1, 2, ,m
j 1
solution
m
xij bj ; j 1, 2, ,n • Because of the special structure of a transportation LP, the
i 1

xij 0; i 1, 2, , m; j 1, 2, ,n iterations of the simplex method have a very special form

• The transportation simplex method is nothing but the

m n
Note that ai bj. original simplex method, but it streamlines the iterations
i 1 j 1

Thus, there are m n -1 basic variables. 169 given this special form 170
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

The Transportation Algorithm The Transportation Algorithm

• Transportation problems are a special type of LP. • The Transportation Algorithm has fewer, less complicated,
– They can be solved through the normal simplex method,
but this way is much faster for these special problems. computations than with the regular simplex method.
• We use a streamlined version of the simplex method to solve. • The Transportation Algorithm requires Less computer
Thus, it is important to recognize them.
• Transportation problems always have integer solutions (as • The Transportation Algorithm produces integer solutions
long as the parameters are integer).
• Classic problem involves transporting items from one set of
nodes to another.
• But other problems can be modeled as transportation problems
and solved more easily.
171 172
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Theorem: Constraint Structure of a TP

A transportation problem always has a solution, but

x11 x12 x1n a1
there is exactly one redundant equality constraint. x21 x22 x2 n a2

When any one of the equality constraints is dropped,

xm1 xm 2 xmn am
the remaining system of n+m-1 equality constraints is x11 x21 xm1 b1
x12 x22 xm 2 b2
linearly independent.
x1n x2 n xmn bn

173 174
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Problem Structure Model Parameters

a
Ax r
b
1T Since the problem structure is fixed, it is
T
1
A only necessary to specify a, b and
1T c11 c1n
In n In n In n

x x11 x12 x1n x21 x2 n xm1 xmn

T C
T
T cm1 cmn
1 1 1 1 1 1 1 1 1
n
T T
a a1 a2 am ;b b1 b2 bn 175 176
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Transformation of Standard Form of Asymmetric Form of Duality
z b A
Transportation Problem into the Primal Form let y b A
Given w b A
T
min c x
s.t. Dual problem can be written as
Ax b max y T b zT b w T b
x 0 s.t.
We can write it in the equivalent form
yT A cT ; y 0
min cT x T T
or z A w A cT ; z 0; w 0
s.t.
Ax b
Ax b Let y z w
T
x 0 max y b
which is in the primal form but with s.t.
A yT A cT
coefficient matrix
A 177 y not restricted to be nonnegative 178
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Interpretation of the Dual

Dual of the Transportation Problem
Transportation Problem
Let us imagine an entrepreneur who, feeling that he can ship
m n
more efficiently, come to the manufacturer with the offer to
max ai ui bjv j
i 1 j 1 buy his product at origins and sell it at the destinations. The
entrepreneur must pay -u1, -u2, …, -um for the product at the m
s.t.
origins and then receive v1, v2, …, vn at the n destinations. To
ui vj cij be competitive with the usual transportation modes, his prices
i 1, 2, , m; j 1, 2, ,n must satisfy ui+vj ≤ cij for all i and j, since ui+vj represents the
net amount the manufacturer must pay to sell a unit of product
ui and v j unbounded
at origin i and but it back again at the destination j.

179 180
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 The Transportation Simplex Method
Properties of Transportation Problem
• The transportation problem is solved in two phases:
• Feasible Solutions Property :
m n m n – Phase I – Finding an initial basic feasible solution
ai bj xij – Phase II – Iterating to the optimal solution
i 1 j 1 i 1 j 1
• In Phase I, an initial basic feasible solution is determined.

If the total supply total demand, it mean either ai or bj • In Phase II, the Stepping Stone Method, using the Modified
represent a bound rather than an exact requirement, in this
Distribution Method (MODI) for evaluating the reduced costs
case introduce “dummy source” or “dummy destination”
as the slack variable. may be used to move from the initial feasible solution to the
optimal one.
• Integer solutions property: Not only ai and bj must be
integer values. But also all the B.F.S. have integer values.
181 182
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Solution Procedure for Transportation

Problem
Conceptually, the transportation problem is solved in a
8. Iterative Computations similar manner as the simplex method.
Begin with an initial basic feasible solution.
This initial feasible solution may or may not be optimal. The
of the Transportation only way you can find out is to test it.
If the solution is not optimal, it is revised and the test is
repeated. Each iteration should bring you closer to the
Algorithm optimal solution.

183 184
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 The Transportation Simplex Method Solution of the Transportation Problem
• The transportation problem is solved in two phases: Once an initial basic feasible solution is obtained, the next

– Phase I – Finding an initial basic feasible solution step is to check its optimality.
– Phase II – Iterating to the optimal solution

• In Phase I, an initial basic feasible solution is determined. An optimal solution is one where there is no other set of
• In Phase II, the Stepping Stone Method, using the Modified transportation routes (allocations) that will further reduce the
Distribution Method (MODI) for evaluating the reduced costs total transportation cost. Thus, we have to evaluate each un-
may be used to move from the initial feasible solution to the
occupied cell in the transportation table in terms of an
optimal one.
opportunity of reducing the total transportation cost.

185 186
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Solution of the Transportation Problem Optimality Test

If we have a B.F.S. consisting of (m+ n–1) independent positive A basic feasible solution is optimal if and
allocations and a set of arbitrary number ui and vj (i=1,2,...m; only if rij cij ui vj 0 for every ( i , j )
j=1,2,...n) such that cij = ui + vj for all occupied cells (i, j) then the
such that xij is non-basic.
value rij corresponding to each empty cell (i, j) is given by :
rij = cij – ui – vj

This value is also called the opportunity cost for un-occupied cells.

187 188
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 The Modified Distribution (MODI) or
Remark: the u-v Method
Step 1: Start with B.F.S. consisting of (m+ n–1) allocations in
independent positions.
Since cij – ui – vj = 0 if xij is a basic variable,
Step 2: Determine a set of m+n numbers ui (i=1,2,....m) and
cij = ui + vj if xij is a basic variable,
vj (j=1,2,...n) for all the rows and columns such that for each
There are m+n-1 basic variables. occupied cell (i,j), the following condition is satisfied :
cij = ui + vj
m+n-1 equations, but m+n unknowns. We arbitrarily choose one of them and equate to zero and determine the
remaining m+n-1 of them. We usually, choose u1 =0
A convenient choice is to select ui that has the largest
number of allocations in its row (tie broken Step 3: Calculate cell evaluations (opportunity cost) rij for
each empty cell (i,j) by using the formula : rij = cij – ui – vj for all i & j.
arbitrarily), and to assign it to be zero.
189 190
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Step 4: Examine the matrix of cell evaluation rij for negative Step 5: Construct a closed path (loop) for the unoccupied cell
entries and conclude that having largest negative opportunity cost. Mark a (+) sign in
this cell and move along the rows (or columns) to find an
i) If all rij > 0 , then solution is optimal and unique.
occupied cell. Mark a (-) sign in this cell and find out another
ii) If at least one rij = 0 , then solution is optimal and alternate
occupied cell. Repeat the process and mark the occupied cells
solution also exists. with (+) and (-) signs alternatively. Close the path back to the
iii) If at least one rij < 0 ,then solution is not optimal and an selected unoccupied cell.
improved solution can be obtained.
In this case, the un-occupied cell with the largest negative value of rij
is considered for the new transportation schedule.

191 192
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Step 6: Select the smallest quantity amongst the cells marked
with (-) sign. Allocate this value to the unoccupied cell of the
loop and add and subtract it in the occupied cells as per their
signs.

Thus an improved solution is obtained by calculating the total

9. Solved Transportation
transportation cost by this method.
Problem Examples
The method described by step 5 and step 6 is called the
stepping stone method.

Step 7: Test the revised solution further for optimality. The

procedure terminates when all rij ≥ 0 , for unoccupied cells.
193 194
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 49: Solve the transportation problem with the Solution: The first step consists of determining a starting
following cost function. BFS. In this case we use the Northwest method.
Destination Supply Destination Supply
3 7 6 4 3 7 6 4
S 5 S 5
3 2 0 1
o o
2 4 3 2 2 4 3 2
u 2 u 1 1 2
r 2
r
4 3 8 5 c 4 3 8 5
c 3 1 2 3
e e
Demand 3 3 2 2 Demand 3 3 2 2

195 196
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
The second step consists of calculate the ui, vi and rij to check the
optimality. Determining the Leaving Variable
Destination
Iteration 1
v1=3v2=7 v3=6 v4=3 Supply
3 7 6 4
S 5 We first construct a closed loop that starts and ends at the entering
u1=0 3 2
o 0 1 variable cell. The loop consists of connected horizontal and
u 2 4 3 2
u2= -3 1 1 2 vertical segments only (no diagonals are allowed). Except for the
r 2 2 entering variable cell, each vertex (or corner) of the closed loop
c 4 3 8 5
u3 = 2 3 must correspond to a basic variable cell. The loop can cross itself
e 1 2
-1 -6 and bypass one or more basic variables. The amount to be

Demand 3 3 2 2 allocated to the entering variable cell is such that it satisfies all the
demand and supply restrictions and must be non-negative. Usually
Thus x32 enters the basis. Total Cost =48 197 198
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

is the minimum of the amounts allocated to the basic cells Destination

adjacent to the entering variable cell. Having decided about the v2=7 v3=6 v4=3 Supply
v1=3
amount to be allocated to the entering cell, for the supply and 3 7 6 4
S u1=0 5
demand limits to remain satisfied, we must alternate between 3 2 0 1
o
subtracting and adding the amount at the successive corners of
u 2 4 - 3 + 2
u2= -3 1 1 2
the loop. In this process one of the basic variables will drop to r 2 2
zero. In simplex language, we say it leaves the basis. We repeat c 4 3 8 - 5
u3= 2 +-6 1 2 3
this process till optimality is reached. We illustrate with a e -1
numerical example.
Demand 3 3 2 2

EE 477 Dr. Mohamed Zribi

199
Total Cost =48 EE 477 Dr. Mohamed Zribi
200
 Iteration 2 Total Cost =42
Thus will become 1 and in the process both the basic Destination
variables x22 and x33 will become simultaneously zero. v1=3 v2=7 v3=12 v4=9 Supply
3 7 6 4
Since only one of them should leave the basis we make S u1=0 3 2 + -5 5
o - -6
x22 leave the basis and keep x33 in the basis but with value
u 2 4 3 2
zero. u2= -9 2 2
r 8 6 2
Thus the transportation cost reduces by 6 (as x23 increases c 4 3 8 5
u3= -4 1 0 - 2 3
by 1) and we say one iteration is over. The resulting new e 5 +
tableau is on the next slide. Demand 3 3 2 2

EE 477 Dr. Mohamed Zribi

201 Thus x13 enters the basis.
EE 477 Dr. Mohamed Zribi
202

Destination
Iteration 3 Total Cost =42
Thus will become 0 and x32 leaves the basis. Again the
BFS is degenerate . But the transportation cost remains v1=3 v2=7 v3=6 v4=9 Supply
3 7 6 4
the same and we say the second iteration is over. The S 3 2 0 5
u1=0
o -5
resulting new tableau is on the next slide.
u 2 4 3 2
u2= -3 2 2
r 2 0 -4
c 4 3 8 5
u3= -4 1 2 3
e 5 6

Demand 3 3 2 2

EE 477 Dr. Mohamed Zribi

203 Thus x14 enters the basis.
EE 477 Dr. Mohamed Zribi
204
 Sometimes it might be difficult to find the closed loop from
the entering cell by inspection. In that case the following
method can be used to find the closed loop. We sketch the v1=3
flowchart of the sequence in which the variables ui and vj
were determined. For example in the above case the flowchart
u1=0 v2=7 u3= -4 v4 = 9
is on the next slide. Now to find the loop emanating from the
non-basic cell (1,4), join u1 and v4 by a dotted line (as
v3=6 u2= -9
shown). Then the closed loop is:
Thus the closed loop is
(1,4) (1,2) (2,3) (3,4) (1,4) (1,4) (1,2) (2,3) (3,4) (1,4) 206
205
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Iteration 4 Total Cost =32

Thus will become 2 and in the process both the basic
variables x12 and x32 will become simultaneously zero. v1=3 v2=7 v3=6 v4=4 Supply
3 7 6 4
Since only one of them should leave the basis we make x32 S 3 0 0 2 5
u1=0
leave the basis and keep x12 in the basis but with value o
u 2 4 3 2
zero. Also x32 becomes 3. u2= -3 2 2
r 2 0 1
c 4 3 8 5
u3= -4 3 3
e 5 6 5
Thus the transportation cost reduces by 1*2+4*2=10 and
Demand 3 3 2 2
we say third iteration is over. The resulting new tableau is
on the next slide. 207 Thus this is the optimal tableau. Alternative Optimal solutions exist.
208
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Example 50: Acme Block Co. (ABC) Solution:
Acme Block Company has orders for 80 tons of concrete Step 1: Initial Transportation Tableau
blocks at three suburban locations as follows: Northwood -- Since total supply = 100 and total demand = 80, a
25 tons, Westwood -- 45 tons, and Eastwood -- 10 tons. dummy destination is created with demand of 20 and 0
Acme has two plants, each of which can produce 50 tons per unit costs.
week. Delivery cost per ton from each plant to each suburban
location is shown on the following table. Westwood Dummy Supply
Northwood Eastwood
24 30 40 0
How should end of week shipments be made to fill Plant 1 50
the above orders? 30 40 42 0
Northwood Westwood Eastwood Plant 2 50
Plant 1 24 30 40
Demand 25 45 10 20
Plant 2 30 40 42

209 210
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Step 2: Finding a starting BFS using the Least

Cost Starting Procedure Hence, a starting BFS using the Least Cost
• Iteration 1: Tie for least cost (0), arbitrarily select Starting Procedure is such that:
x14. Allocate 20. Reduce s1 by 20 to 30 and delete the
Dummy column.
• Iteration 2: Of the remaining cells the least cost is Westwood Dummy Supply
24 for x11. Allocate 25. Reduce s1 by 25 to 5 and Northwood Eastwood
eliminate the Northwood column. 24 30 40 0
Plant 1 25 5 20 50
• Iteration 3: Of the remaining cells the least cost is
30 for x12. Allocate 5. Reduce the Westwood column 30 40 42 0
Plant 2 40 10 50
to 40 and eliminate the Plant 1 row.
• Iteration 4: Since there is only one row with two Demand 25 45 10 20
cells left, make the final allocations of 40 and 10 to x22
and x23, respectively.
211 212
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Step 3: Finding the Optimal Solution:
Calculate the reduced costs (circled numbers on the
next slide) by cij - ui - vj.
Iteration 1
Unoccupied Cell Reduced Cost
• The MODI Method
(1,3) 40 - 0 - 32 = 8
1. Set u1 = 0
(2,1) 30 - 24 -10 = -4
2. Since u1 + vj = c1j for occupied cells in row 1, then
(2,4) 0 - 10 - 0 = -10
v1 = 24, v2 = 30, v4 = 0.
3. Since ui + v2 = ci2 for occupied cells in column 2,
then u2 + 30 = 40, hence u2 = 10.
4. Since u2 + vj = c2j for occupied cells in row 2, then
10 + v3 = 42, hence v3 = 32.

213 214
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Iteration 1 Tableau
The Stepping Stone Method
The stepping stone path for cell (2,4) is (2,4), (1,4),
(1,2), (2,2). The allocations in the subtraction cells
Westwood Dummy
Northwood Eastwood ui are 20 and 40, respectively. The minimum is 20,
24 30 40 0 and hence reallocate 20 along this path. Thus for
Plant 1 25 5 +8 20 0
the next tableau:
30 40 42 0 x24 = 0 + 20 = 20 (0 is its current allocation)
Plant 2 -4 40 10 -10 10
x14 = 20 - 20 = 0 (blank for the next tableau)
vj 24 30 32 0 x12 = 5 + 20 = 25
x22 = 40 - 20 = 20
Note that the most negative reduced cost is = -10 determined by x24.
Therefore, x24 will be the entering variable. The stepping stone method The other occupied cells remain the same.
will help us determine the leaving variable.
215 216
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Iteration 1 Tableau (Stepping Stone Method)
New BFS Tableau 1
Westwood Dummy Supply Westwood Dummy Supply
Northwood Eastwood Northwood Eastwood
24 30 40 0 24 30 40 0
Plant 1 25 5+ 20 - 50 Plant 1 25 25 50

30 40 42 0 30 40 42 0
Plant 2 40- 10 + 50 Plant 2 20 10 20 50

Demand 25 45 10 20 Demand 25 45 10 20

217 218
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Iteration 2 Calculate the reduced costs (circled numbers on the

next slide) by cij - ui - vj.

• The MODI Method Unoccupied Cell Reduced Cost

The reduced costs are found by calculating the ui's (1,3) 40 - 0 - 32 = 8
and vj's for this tableau. (1,4) 0 - 0 - (-10) = 10
1. Set u1 = 0. (2,1) 30 - 10 - 24 = -4
2. Since u1 + vj = cij for occupied cells in row 1, then
v1 = 24, v2 = 30.
3. Since ui + v2 = ci2 for occupied cells in column 2,
then u2 + 30 = 40, or u2 = 10.
4. Since u2 + vj = c2j for occupied cells in row 2, then
10 + v3 = 42 or v3 = 32; and, 10 + v4 = 0 or v4 = -10.

219 220
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 • The Stepping Stone Method
The most negative reduced cost is = -4 determined
Iteration 2 Tableau by x21. The stepping stone path for this cell is
Westwood (2,1),(1,1),(1,2),(2,2). The allocations in the
Dummy ui
Northwood Eastwood subtraction cells are 25 and 20 respectively.
24 30 40 0 Thus the new solution is obtained by reallocating 20
Plant 1 25 25 +8 +10 0
on the stepping stone path. Thus for the next
30 40 42 0 tableau:
Plant 2 -4 20 10 20 10
x21 = 0 + 20 = 20 (0 is its current allocation)
vj 24 30 36 -6 x11 = 25 - 20 = 5
x12 = 25 + 20 = 45
Note that the most negative reduced cost is = -4 determined by x21.
x22 = 20 - 20 = 0 (blank for the next tableau)
Therefore, x21 will be the entering variable. The stepping stone method
will help us determine the leaving variable.
The other occupied cells remain the same.
221 222
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Iteration 2 Tableau (Stepping Stone Method) New BFS Tableau 2

Westwood Dummy Supply Westwood Dummy Supply
Northwood Eastwood Northwood Eastwood
24 30 40 0 50 24 30 40 0 50
Plant 1 25- 25+ Plant 1 5 45

30 40 42 0 30 40 42 0
Plant 2 + 20- 10 20 Plant 2 20 10 20
50 50

Demand 25 45 10 20 Demand 25 45 10 20

223 224
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Iteration 3
Calculate the reduced costs (circled numbers on the
next slide) by cij - ui - vj.
• The MODI Method
The reduced costs are found by calculating the ui's Unoccupied Cell Reduced Cost
and vj's for this tableau. (1,3) 40 - 0 - 36 = 4
1. Set u1 = 0 (1,4) 0 - 0 - (-6) = 6
2. Since u1 + vj = c1j for occupied cells in row 1, then (2,2) 40 - 6 - 30 = 4
v1 = 24 and v2 = 30.
3. Since ui + v1 = ci1 for occupied cells in column 2,
then u2 + 24 = 30 or u2 = 6.
4. Since u2 + vj = c2j for occupied cells in row 2, then
6 + v3 = 42 or v3 = 36, and 6 + v4 = 0 or v4 = -6.

225 226
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

The Optimal Solution

Iteration 3 Tableau
From To Amount Cost
Since all the reduced costs are non-negative, this is the
optimal tableau. Plant 1 Northwood 5 120
Plant 1 Westwood 45 1,350
Westwood Dummy ui Plant 2 Northwood 20 600
Northwood Eastwood
24 30 40 0 Plant 2 Eastwood 10 420
Plant 1 5 45 +4 +6 0 Total Cost = $2,490
30 40 42 0
Plant 2 20 +4 10 20 6

vj 24 30 36 -6

227 228
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Example 51: Solution:
1 2 3 Supply
In the unbalanced transportation problem given in the table
1 1 2 1 20
below, if a unit from a source is not shipped out (to any of the
destinations) a storage cost is incurred at the rate of $5, $4, and
2 3 4 5 40
3 2 3 3 30
$3 per unit for sources 1,2, and 3 respectively. Additionally all
the supply at source 2 must be shipped out completely to make
Demand 30 20 20
room for a new product. Use VAM to determine the starting To balance the problem, we introduce a dummy
solution and determine the optimum solution. destination with transportation costs $5, $M, $3
respectively.

229 230
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Destination Destination
1 2 3 Dummy Supply Row Penalties Starting Tableau Total Cost =240
1 2 1 5 0 - - -
1 20 20 v2=3 v3=3 v4=3 Supply
v1=2
S
1 2 1 5
o 3 4 5 M
40 10 1 1 1 1 S u1=-2 20 20
2 30 10 1 1 4
u o
r 2 3 3 3 u 3 4 5 M
3 10 0 20 30 10 1 11 0 u2= 1 30 10 40
c r 1 -4+M
e Demand 30 20 10 20 0 20 c 2 3 3 3
u3 = 0 10 0 20 30
e 0
1 1 2 2
Column
Penalties 1 1 - M-3 Demand 30 20 20 20
1 1 - -
- 1 - - 231 Thus this is the optimal tableau. Alt Opt solutions exist. 232
EE 477 Total shipping cost = 240
Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Example 52: If each task is to be allocated to a person (i.e. no splitting of the

The table below gives the times taken by 3 persons to task between 2 or more persons is allowed) and if each person can

complete 4 tasks ( i.e. cell (i,j) is the time taken by person i to be assigned at most two tasks, find the optimum allocation of the

complete the task j). jobs to the persons to minimize the total time taken to complete all
the 4 tasks.
Tasks
1 2 3 4 Solution: This can be formulated as a transportation model with
1 4 1 2 6 three sources (persons) and 4 destinations (tasks). The demands at
Person 2 6 4 3 5 the three destinations are bj = 1 for j=1,2,3,4. But the availabilities
3 5 2 6 4 are ai = 2 for i=1, 2, 3 as each person can be assigned a maximum
of two tasks. Thus to balance the problem, we introduce a dummy
233 task with demand 2 and time 0. Thus we get the starting tableau: 234
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Destination
Starting Tableau Destination Starting Tableau
1 2 3 4 Dummy Supply v1=5 v2=2 v3=3 v4=4 v5=0 Supply
S S
o 1 4 11 2 1 6 0 2 1 1 1 3 o u1= -1
4 1 2 6 0
2
0 1 1
-3 -1
u 2 6 4 3 5 0 2 2 3 - -
u
6 4 3 5 0
r r u2= 0 2 2
c 3 5 1
2 0
6 4 1 0 0 2 2 22 c -1 -2 0 -1
e e 5 2 6 4 0
2
0
Demand 1 1 0 1 1 2 0 u3= 0 1 0 1
-3

1 1 1 1 0 Demand 1 1 1 1 2
1 1 4 2 -
1 1 - 2 - Thus this is the optimal tableau. Optimal cost = 12
235 236
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Example 53: Medical Supply Transportation Problem

• A Medical Supply company produces catheters in packs FROM

TO WAREHOUSE

at three productions facilities. PLANT Seattle New York Phoenix Miami

Juarez $19 $7 $3 $21

• The company ships the packs from the production Seoul 15 21 18 6
facilities to four warehouses. K.L 11 14 15 22

• The packs are distributed directly to hospitals from the

Capacity Demand
warehouses. Juarez 100 Seattle 150
Seoul 300 New York 100
• The table on the next slide shows the costs per pack to KL 200 Phoenix 200
Miami 150
ship to the four warehouses.
237 238
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Northwest Corner Method

To
TO WAREHOUSE S N P M
From Plant Capacity
Plant Capacity From
S N P M
19 7 3 21

J x11 x12 x13 x14 100 J 100

S x21 x22 x23 x24 300 15 21 18 6

S 300
KL x31 x32 x33 x34 200
11 14 15 22
Warehouse
Demand
150 100 200 150 600 KL 200

-- Number of constraints = number of rows + number of columns Demand 150 100 200 150 600
--Total plant capacity must equal total warehouse demand.
Although this may seem unrealistic in real world application, it Begin with a blank shipment schedule. Note the shipping
is possible to construct any transportation problem using this model. costs in the upper right hand corner of each cell.
239 240
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Northwest Corner Method Northwest Corner Method

To To
S N P M S N P M
Capacity Capacity
From From
19 7 3 21 19 7 3 21

J 100 100 J 100 100

15 21 18 6 15 21 18 6

S 300 S 50 300

11 14 15 22 11 14 15 22

KL 200 KL 200

Demand 150 100 200 150 600 Demand 150 100 200 150 600

Start in the upper left-hand corner, “northwest corner” of the schedule Since Juarez capacity is depleted move down to repeat the
and place the largest amount of capacity and demand available in that process for the Seoul to Seattle cell. Seoul has sufficient
cell. Seattle demands 150 and Jaurez has a capacity of 100. capacity but Seattle can only take another 50 packs of demand.
241 242
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Northwest Corner Method Northwest Corner Method

To
S N P M To
Capacity S N P M
From Capacity
From
19 7 3 21

J 100 100
19 7 3 21 1900
J 1900 100
15 21 18 6 750
15 21 18 6
S 50 100 150 300
S 750 2100 2700 300 2100
11 14 15 22

KL 50 150 200
11 14 15 22 2700
KL 750 3300 200
750
Demand 150 100 200 150 600
Demand 150 100 200 150 600 3300
Now move to the next cells to the right and assign capacity for
Seoul to warehouse demand until depleted. Then move down The previous slides show the process of satisfying all constraints
and allows us to begin with a starting feasible solution. Multiply C =11,500
to the KL row and repeat the process.
243
the quantity in each cell by the cost. 244
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 For non empty cells: cij = ui+ vj For non empty cells: cij = ui+ vj
Assign zero as the row number for the first row. Assign zero as the row number for the first row.

To
S N P M
To Capacity
S N P M From
Capacity
From 19 7 3 21
19 7 3 21 J 100 100 u1 = 0 a
u1 = 0 a
J 100 100
15 21 18 6
15 21 18 6 S 50 100 150 300 u2 = -4 c
S 50 100 150 300
11 14 15 22
11 14 15 22 T 50 150 200
KL 50 150 200

Demand 150 100 200 150 600

Demand 150 100 200 150 600
15 = u2 + 19 v1 = 19
v1 = 19 b
u2 = -4
19 = (0) + v1 b
245 246
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

For non empty cells: cij = ui+ vj

For non empty cells: cij = ui+ vj Assign zero as the row number for the first row then
Assign zero as the row number for the first row. use the newest r value to compute the next k.

To To
S N P M S N P M
Capacity Capacity
From From
19 7 3 21 19 7 3 21
J 100 100 u1 = 0 a J 100 100 u1 = 0 a

15 21 18 6 15 21 18 6
S 50 100 150 300 u2 = -4 c S 50 100 150 300 u2 = -4 c
* *
11 14 15 22 11 14 15 22
T 50 150 200 T 50 150 200 u3 = -7 e

Demand 150 100 200 150 600 Demand 150 100 200 150 600

v1 = 19 v3 = 22 v1 = 19 v3 = 22
18 =-4 + v3 b d
15 = u3+ 22 b d
Skip cell SN, mark it * for later
v3 = 22 u3= -7
and move on to cell SP .
247 248
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 For non empty cells: cij = ui+ vj For non empty cells: cij = ui+ vj
Assign zero as the row number for the first row then Assign zero as the row number for the first row then
use the newest r value to compute the next k. use the newest r value to compute the next k.

To To
S N P M S N P M
Capacity Capacity
From From
19 7 3 21 19 7 3 21
J 100 100 u1 = 0 a J 100 100 u1 = 0 a

15 21 18 6 15 21 18 6
S 50 100 150 300 u2 = -4 c S 50 100 150 300 u2 = -4 c
* *
11 14 15 22 11 14 15 22
T 50 150 200 u3 = -7 e T 50 150 200 u3 = -7 e

Demand 150 100 200 150 600 Demand 150 100 200 150 600

v1 = 19 v3 = 22 v4 = 29 v1 = 19 v2 = 25 v3 = 22 v4 = 29
22 = -7 + v4 b d f
21= -4 + v2 b g d f
v4= 29 v2= 25
249 250
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Next calculate empty cells using: cij - ui - vj Next calculate empty cells using: cij - ui - vj

Improvement Difference >> r12 = 7 – 0 – 25 = -18 Improvement Difference >> r13 = 3 – 0 – 22 = -19

To To
S N P M S N P M
Capacity Capacity
From From
19 7 3 21 19 7 3 21
J 100 100 u1 = 0 a J 100 100 u1 = 0 a
-18 -18 -19
15 21 18 6 15 21 18 6
S 50 100 150 300 u2 = -4 c S 50 100 150 300 u2 = -4 c

11 14 15 22 11 14 15 22
T 50 150 200 u3 = -7 e T 50 150 200 u3 = -7 e

Demand 150 100 200 150 600 Demand 150 100 200 150 600

v1 = 19 v2 = 25 v3 = 22 v4 = 29 v1 = 19 v2 = 25 v3 = 22 v4 = 29
b g d f b g d f

251 252
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Next calculate empty cells using: cij - ui - vj Next calculate empty cells using: cij - ui - vj

Improvement Difference >> r14 = 21 – 0 – 29 = -8 Improvement Difference >> r24 = 6 – (-4) – 29 = -19

To To
S N P M S N P M
Capacity Capacity
From From
19 7 3 21 19 7 3 21
J 100 100 u1 = 0 a J 100 100 u1 = 0 a
-18 -19 -8 -18 -19 -8
15 21 18 6 15 21 18 6
S 50 100 150 300 u2 = -4 c S 50 100 150 300 u2 = -4 c
-19
11 14 15 22 11 14 15 22
T 50 150 200 u3 = -7 e T 50 150 200 u3 = -7 e

Demand 150 100 200 150 600 Demand 150 100 200 150 600

v1 = 19 v2 = 25 v3 = 22 v4 = 29 v1 = 19 v2 = 25 v3 = 22 v4 = 29
b g d f b g d f

253 254
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Next calculate empty cells using: cij - ri - kj Next calculate empty cells using: cij - ui - vj

Improvement Difference >> r31 = 11 – (-7) – 19 = -1 Improvement Difference >> r32 = 14 – (-7) – 25 = -4

To To
S N P M S N P M
Capacity Capacity
From From
19 7 3 21 19 7 3 21
J 100 100 u1 = 0 a J 100 100 u1 = 0 a
-18 -19 -8 -18 -19 -8
15 21 18 6 15 21 18 6
S 50 100 150 300 u2 = -4 c S 50 100 150 300 u2 = -4 c
-19 -19
11 14 15 22 11 14 15 22
T 50 150 200 u3 = -7 e T 50 150 200 u3 = -7 e
-1 -1 -4

Demand 150 100 200 150 600 Demand 150 100 200 150 600

v1 = 19 v3 = 25 v3 = 22 v4 = 29 v1 = 19 v2 = 25 v3 = 22 v4 = 29
b g d f b g d f

255 256
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Improvement Difference >> Improvement Difference >>

To To
S N P M S N P M
Capacity Capacity
From From

19 7 3 21 19 7 3 21

J 100 100 u1 = 0 a J 100 100 u1 = 0 a

-18 -19 -8 + -8
-
15 21 18 6 15 21 18 6

S 50 100 150 300 c S 50 100 150 300 c

-19 u2 = -4 + -19 u2 = -4
-
11 14 15 22 11 14 15 22

T -1 -4 50 150 200 u3 = -7 e T -1 -4 50 150 200 u3 = -7 e

Demand 150 100 200 150 600 Demand 150 100 200 150 600

v1 = 19 v2 = 25 v3 = 22 v4 = 29 v1 = 19 v2 = 25 v3 = 22 v4 = 29
b g d f b g d f

257 258
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Next calculate the entering cell by finding the empty

cell with the greatest absolute negative improvement
difference. Continue around the closed loop until all tradeoffs are
Cells JP and SM are tied for the greatest improvement at
completed.
$19 per pack. Break the tie and arbitrarily choose JP. JP
becomes the entering cell. Place a + sign in cell JP Previous cost was $11,500 and the new is:

To To
S N P M S N P M
Capacity Capacity 300
From From
19 7 3 21 19 7 3 21 2250
J 100 100 100 u1= 0 J 100 100 u1 = 0 2100
(-) -18 -19 (+) -8
(-) -18 -19 (+) -8
15 21 18 6 15 21 18 6
900
S 50 100 150 300 S 150
50 100 50 300 750
(+) (-) -19 u2 = -4 (+) (-) -19 u2 = -4 3300
11 14 15 22 11 14 15 22
C = $9,600
T 50 150 200 T 50 150 200
-1 -4 u3 = -7 -1 -4 u3= -7

Demand 150 100 200 150 600 Demand 150 100 200 150 600

v1= 19 v2 = 25 v3 = 22 v4 = 29 v1 = 19 v2 = 25 v3 = 22 v4 = 29
Note: Except for the entering cell all changes must Note: Except for the entering cell all changes must involve nonempty cells.
involve nonempty cells.
259 260
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Begin another iteration choosing the empty cell with the Begin another iteration choosing the empty cell with the
greatest absolute negative improvement difference. greatest absolute negative improvement difference.
>>>>>SM SM
Previous cost was $9,600, now the new is:

To To
S N P M S N P M
Capacity Capacity 300
From From
19 7 3 21 19 7 3 21 2250
J 100 100 u1 = 0 J 100 100 u 1= 0 2100
19 1 11 19 1 11 300
15 21 18 6 15 21 18 6
S 150
50 100 50 300 S 150
50 100 50 50 300 1500
(-) -19
(+) u2 = 15 (-) -19
(+) u2 = 15 2200
11 14 15 22 11 14 15 22
150 100 150
100
C = $8,650
T 50 200 T 50 200
-1 -4 (+) (-) u3 = 12 -1 -4 (+) (-) u3 = 12

Demand 150 100 200 150 600 Demand 150 100 200 150 600

v1 = 0 v2 = 6 v3 = 3 v4 = 10 v1 = 0 v 2= 6 v3 = 3 v4 = 10

Note: The r and k values and the improvement Note: The r and k values and the improvement
difference values have changed. 261 difference values have changed. 262
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Begin another iteration choosing the empty cell with the Begin another iteration choosing the empty cell with the
greatest absolute negative improvement difference. greatest absolute negative improvement difference.
SM SM
Previous cost was $8,650, now the new is: $6,350

To To
S N P M S N P M
Capacity 300 Capacity 300
From From
19 7 3 21 2250 19 7 3 21 2250
J 100 100 u1 = 0 900 J 100 100 u1 = 0 900
0 -18 11 1400 20 5 31 1400
15 21 18 6 15 21 18 6
100 50
150 1500 150 150 1500
S 150
50 300 S 300
(-) -19
(+) u2 = -4 (-) 3 -1
(+) u2 = 16
11 14 15 22 11 14 15 22
100 100 0
100 C = $6,350 0 100 100 C = $6,350
T 200 T 200
-20
(+)
-23
(-) u3 = 12 (+) (-) 20 u3= 12

Demand 150 100 200 150 600 Demand 150 100 200 150 600

v1= 19 v2= 2 v3 = 3 v4 = 10 v1 = -1 v2 = 2 v3 = 3 v4 = 10

Note: The u and v values and the improvement Note: The u and v values and the improvement
difference values have changed. 263 difference values have changed. 264
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Optimal Solution In five iterations the shipping cost has moved Example 54:
from $11,500 to $6,250. There are no
remaining empty cells with a negative value. Solve the following transportation problem.
$6,250

Client Supply
To
S N P M
Capacity
From
300 1 2 3 4
19 7 3 21 750
J 100 100 u1 = 0 1800
19 4 30 900
1 4 6 8 3 700
15 21 18 6
S
50 100 150
300 1100
(-) (+) u2 = 15 1400
11
3
14 15 22
C = $6,250 Factories 2 5 5 7 4 900
100 100
T 200
(+) 1 (-) 20 u3= 11

Demand 150 100 200 150 600 3 2 3 11 9 1600

v1 = 0 v2 = 3 v3= 3 v4 = -9

Note: The u and v values and the improvement difference values have changed. Demand 800 500 600 1300
265 266
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Solution: Iteration 1:

Client Supply Client Supply

1 2 3 4 1 2 3 4
x11 x12 x13 x14 700
1 700
1 4 6 8 3 700 4 6 8 3

300 600
x21 x22 x23 x24 Factories 2 900
5 5 7 4 5 5 7 4
Factories 2 900
500 500 600
3 1600
x31 2 x32 3 x33 11 x34 9 2 3 11 9
3 1600
Demand 800 500 600 1300
Demand 800 500 600 1300 267 268
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
BASIC: Set x1j = 0; Solve variables using cij = ui + vj NON-BASIC: xij = cij – ui - vj
vj 4 5 13 3 vj 4 5 13 3
ui Client1 Client 2 Client 3 Client 4 ui Client 1 Client 2 Client 3 Client4
700 0 1 -5 700
0 F1 4 6 8 3
700 0 F1 4 6 8 3
700
300 600 300 -1 -7 600
1 F2 900 1 F2 900
5 5 7 4 5 5 7 4

500 500 600 500 500 600 8

-2 F3 1600 -2 F3 1600
2 3 11 9 2 3 11 9

800 500 600 1300 800 500 600 1300

269 270
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

vj 4 5 13 3 vj 4 5 13 3
ui 1 2 3 4 ui 1 2 3 4
0 1 -5 700 0 1 -5 700
0 1 700
0 1 4 6 8 3 700 4 6 8 3

300- θ -1 -7+θ 600

300 -1 -7 600 1 2 900
5 5 7 4 5 5 7 4
1 2 900
500+ θ 500 600- θ 8
-2 3 1600
500 2 500 3 600 11 8 9 2 3 11 9
-2 3 1600
800 500 600 1300
800 500 600 1300 271 272
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Iteration 2:

vj vj -3 -2 6 3
ui 1 2 3 4 ui 1 2 3 4
700 700
1 4 6 8 3 700 0 1 4 6 8 3 700

0 5 5
300 7
600 4
0 5 5
300 7
600 4
2 900 1 2 900

800 2 500 3 300 11 9 800 2 500 3 300 11 9

3 1600 5 3 1600

800 500 600 1300 273 800 500 600 1300 274
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

The solution is optimal since all the rij are non-

negative.
vj -3 -2 6 3
ui 1 2 3 4 The optimal solution is
7 8 2 700 x14= 700, x23=300, x24= 600, x31=800, x32=500,
0 1 4 6 8 3 700 x33=300.
7 5
6 5
300 7
600 4
1 2 900 The optimal cost is Z = .
800 2 500 3 300 11 1 9
5 3 1600

800 500 600 1300 275 276

EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Example 55: Solution:
Solve the following transportation problem. Finding a starting BFS using the Northwest rule.

Destination Destination
W1 W2 W3 W4 W1 W2 W3 W4
12 10 5 8 12 10 5 8
P1 90 P1 90
65 25
7 4 9 6 7 4 9 6
P2 120 P2 120
1 1 55 65
9 11 7 5 9 11 7 5
P3 75 P3 75
5 9 5 9 25 50
65 80 90 50 65 80 90 50

277 278
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Destination
W1 W2 W3 W4 ui W1 W2 W3 W4 ui
12 10 5 8 12 10 5 8
P1 90 0 P1 90 0
65 25 10 5 65 10 25 5
7 4 9 6 7 4 9 6
P2 120 6 P2 120 4
1 55 65 1 9 80 40 1
9 11 7 5 9 11 7 5
P3 75 8 P3 75 2
5 9 25 50 5 9 25 50
65 80 90 50 Z 2, 260 65 80 90 50 Z 2, 010
vj 12 10 15 13 vj 12 0 5 3

279 280
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 W1 W2 W3 W4 ui W1 W2 W3 W4 ui
12 10 5 8 12 10 5 8
P1 90 0 P1 90 0
25 1 65 5 0 1 90 0
7 4 9 6 7 4 9 6
P2 120 5 P2 120 5
40 80 9 8 40 80 9 3
9 11 7 5 9 11 7 5
P3 75 2 P3 75 3
5 0 25 50 25 5 5 50
65 80 90 50 Z 1, 625 65 80 90 50 Z 1,525
vj 12 9 5 3 vj 12 9 5 8

281 282
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 56:
The solution is optimal since all the rij are non-
negative. • Holiday shipments of iPods to distribution centers
• Production at 3 facilities,
The optimal solution is – A, supply 200k
– B, supply 350k
x13= 90, x21=40, x23= 80, x31=25, x34=80. – C, supply 150k
• Distribute to 4 centers,
The optimal cost is Z = 1525. – N, demand 100k
– S, demand 140k
– E, demand 300k
– W, demand 250k
• Total demand vs. total supply
283 284
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Step 2: Finding a starting BFS using the
Step 1: Initial Transportation Tableau
Northwest Method
Destination Destination Suppl
Supply
Source N S E W Source N S E W y
16 13 22 17 16 13 22 17
A 200 A 200
100 100

14 13 19 15 14 13 19 15
B 350 B 350
40 300 10
9 20 23 10 9 20 23 10
C 150 C 150
150

0 0 0 0 0 0 0 0
Dummy 90 Dummy 90
90

Demand 100 140 300 250 285

Demand 100 140 300 250 286
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Step 3: Finding a starting BFS using Vogel’s

Destination
Method Supply diff
Destination Source N S E W
Supply diff
Source N S E W 16 13 22 17
A 200 3
16 13 22 17
A 200 3
14 13 19 15
B 350 1
14 13 19 15
B 350 1
9 20 23 10
C 150 1
9 20 23 10
C 150 1
0 0 0 0
Dummy 90 90 0
Dumm 0 0 0 0
90 0
y
Demand 100 140 300 250
Demand 100 140 300 250
287 diff 9 13 19 10 288
diff 9 13
EE 477
19 10
Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Destination Destination
Supply diff Supply diff
Source N S E W Source N S E W
16 13 22 17 16 13 22 17
A 200 3 A 200 3

14 13 19 15 14 13 19 15
B 350 1 B 350 1

9 20 23 10 9 20 23 10
C 150 1 C 100 150 1

0 0 0 0 0 0 0 0
Dummy
90 --- --- Dummy --- ---
90

Demand 100 140 210 250 Demand 100 140 210 250

diff 5 0 3 5 diff 5 0 3 5
289 290
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Destination Destination
Supply diff
Supply diff Source N S E W
Source N S E W
16 13 22 17
16 13 22 17 A 200 4
A 200 4
14 13 19 15
14 13 19 15 B 350 2
B 350 2
9 20 23 10
9 20 23 10 C 50 10
C 100 50 10 100 50

0 0 0 0 0 0 0 0
Dummy --- ---
Dummy --- ---
90 90

Demand --- 140 210 250

Demand --- 140 210 250

diff --- 0 3 5
diff --- 0 3 5
291 292
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Destination Destination
Supply diff Supply diff
Source N S E W Source N S E W
16 13 22 17 16 13 22 17
A 200 4 A 200 4
140

14 13 19 15 14 13 19 15
B 350 2 B 350 2

9 20 23 10 9 20 23 10
C --- --- C --- ---
100 50 100 50

0 0 0 0 0 0 0 0
Dummy --- --- Dummy --- ---
90 90

Demand --- 140 210 200 Demand --- 140 210 200
diff --- 0 3 2 diff --- 0 3 2
293 294
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Destination
Supply diff Destination
Source N S E W Supply diff
Source N S E W
16 13 22 17
A 60 5 16 13 22 17
140 A 60 5
140 60
14 13 19 15 14 13 19 15
B 350 4
B 350 4
9 20 23 10
C --- --- 9 20 23 10
100 50 C --- ---
100 50
0 0 0 0 0 0 0 0
Dummy --- ---
90 Dummy --- ---
90

Demand --- --- 210 200 Demand --- --- 210 200
diff --- --- 3 2 diff --- --- 3 2
295 296
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Destination Destination
Supply diff Supply diff
Source N S E W Source N S E W
16 13 22 17 16 13 22 17
A --- --- A --- ---
140 60 140 60

14 13 19 15 14 13 19 15
B 350 4 B ---
210 140

9 20 23 10 9 20 23 10
C 50
--- --- C --- ---
100 100 50

0 0 0 0 0 0 0 0
Dummy --- --- Dummy --- ---
90 90

Demand --- --- 210 140 Demand --- --- --- --- Z = 10330

diff --- --- 297 diff --- --- 298

EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Step 4: Finding the Optimal Solution

Destination
Supply ui
Destination Source N S E W
Supply ui
Source N S E W 16 13 22 17
A 140 60 200 0
16 13 22 17
A 140 60 200
14 13 19 15
B 210 140 350 -2
14 13 19 15
B 210 140 350
9 20 23 10
C 150 -7
9 20 23 10 100 50
C 100 50 150
0 0 0 0
Dummy 90 90 -21
0 0 0 0
Dummy 90 90
Demand 100 140 300 250
Demand 100 140 300 250 299
vj 16 13 21 17 300
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Destination The solution is optimal since all the rij are non-negative.
Supply ui
Source N S E W
16 13 22 17
A 140 60 200 0 The optimal solution is
0 1
14 13 19 15 x12= 140, x14=60, x23= 210, x24=140, x31=100. x34=150,
B 210 140 350 -2
0 2 x43=90.
9 20 23 10
C 100 50
150 -7
14 9
0 0 0 0 The optimal cost is Z = 10330.
Dumm
90 90 -21
y 5 8 4
vj 16 13 21 17
301 302
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 57: Solution: Finding a starting BFS using the

Northwest Method
Solve the following transportation problem.
Destination
Supply Destination
Source N S E W Supply ui
Source N S E W
16 13 22 17
A 200 16 100 13 100 22 17
A 200 0
3 2
14 13 19 15 40 300 10
B 350 14 13 19 15
B 350 0
-2 150
9 20 23 10
C 150 9 20 23 10
C 150 -5
-2 12 9 90
0 0 0 0
Dummy 90 0 0 0 0
Dummy 90 -15
-1 2 -4
Demand 100 140 300 250 303 vj 16 13 19 15 304
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Iteration 1

Destination Destination
Supply ui Supply ui
Source N S E W Source N S E W
16 100 13 100 22 17 16 100 13 100 22 17
A 200 0 A 200
- +
40 300 10 40 210 100
14 13 19 15 14 13 19 15
B 350 0 B 350
150 150
9 20 23 +
10 9 20 23 10
C - 150 -5 C 150
? 90 90

0 0 0 0 0 0 0 0
Dummy 90 -15 Dummy 90

vj 16 13 19 15 305 vj 306
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Optimality Test Iteration 2

Destination Destination
Supply ui Supply ui
Source N S E W Source N S E W
- +
16 100 13 100 22 17 16 100 13 100 22 17
A 200 0 A 200 0
3 2 + -
40 210 100 ? 40 210 100
14 13 19 15 14 13 19 15
B 350 0 B 350 0
-2 150 150
9 20 23 10 9 20 23 10
C 150 -5 C 150 -5
-2 12 9
90 90

0 0 0 0 0 0 0 0
Dummy 90 -19 Dummy 90 -19
3 6 4
307 vj 16 13 19 15 308
vj 16 13477
EE 19
Dr. Mohamed Zribi15 EE 477 Dr. Mohamed Zribi
 Optimality Test

Destination Destination
Supply ui Supply ui
Source N S E W Source N S E W
16 60 13 140 22 17 16 60 13 140 22 17
A 200 A 200 2
1 0
40 210 100 40 210 100
14 13 19 15 14 13 19 15
B 350 B 350 0
150
2 150
9 20 23 10 9 20 23 10
C 150 C 150 -5
90 0 14 990
0 0 0 0 0 0 0 0
Dummy 90 Dummy 90 -19
5 8 4 Z = 10330
vj 309 vj 14 11 19 15 310
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 58: Solution:

Solve the following transportation problem.
Step1: Determine a Starting BFS Using the Northeast
Corner Method

311 312
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Step 2: Determine a Starting BFS Using the Least Cost
Method

--

--
200

200 75

The min cost, so allocate as much resource as possible here

313 314
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Step 3: Determine a Starting BFS Using the Vogel’s

Method

(8-6)

(11-7)
Stop. The above tableau is an initial tableau because all (5-4)
allocations have been assigned
(6-4) (8-5) (11-10)

315 316
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Highest penalty
cost

--- ---

this has the least cost

317 318
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

--- --- ---

--- --- ---

---

319 320
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Step 4: Check the Optimality of the solution

Initial solution from:

Northeast cost, total cost =$5,925
The min cost, total cost =$4,550
VAM, total cost = $5,125
(note: here, we are not saying the second one always
better!)

Hence, the second one has the minimum cost, but is it the
optimal solution?
The Initial tableau for VAM
321 322
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Using the MODI method to check for optimality.

Consider to this initial tableau:

Step 1: (a) form a set of equations that uses to

represent all basic variables
C3A ui + vj = cij
323 324
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Because we
added an
non-basic variable

(b) solve variables by assign

one variable = 0
325 326
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

-1

-1 +2

+5

6-0-7
Step 3: Select the cell that has the most –ve value in 2b
Note this may look difficult and complicated, however, we can add these 327 328
EE 477 asDr.
V=values into the above tableau Mohamed Zribi
well EE 477 Dr. Mohamed Zribi
 Use the stepping-stone method

-1
+ -
-1 +2

+5

- +

Select either one, (Why?)

From here we have ….
These cells mean, introduce it will reduce the min z to -1 cost unit
329 330
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

331 332
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Example 59:
Solve the following transportation problem.

All positives
STOP

333 334
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Solution: Check the optimality.

The solution is optimal since all the rij ≥ 0.

335 336
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Example 60:
- Formulas for cells containing allocations:

Solve the following transportation problem. x1B: u1 + vB = 8

x1C: u1 + vC = 10
x2C: u2 + vC = 11
ui + vj = cij = unit transportation cost for cell ij.
x3A: u3 + vA = 4
x3B: u3 + vB = 5

The Initial Solution with All ui and vj Values

The Minimum Cell Cost
Initial Solution
- Five equations with 6 unknowns, therefore let u1 = 0 and solve to obtain:
vB = 8, vC = 10, u2 = 1, u3 = -3, vA= 7

337 338
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

- After each allocation to an empty cell, the ui and vj values must be recomputed.

- Each MODI allocation replicates the stepping-stone allocation.

- Use following to evaluate all empty cells:
cij - ui - vj = kij
where kij equals the cost increase or decrease that would occur by allocating to a cell.
- For the empty cells in Table 26:
x1A: k1A = c1A - u1 - vA = 6 - 0 - 7 = -1
x2A: k2A = c2A - u2 - vA = 7 - 1 - 7 = -1
x2B: k2B = c2B - u2 - vB = 11- 1 - 8 = +2
x3C: k3C = c3C - u3 -vC = 12 - (-3) - 10 = +5
The Second Iteration of the MODI Solution Method

339 340
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 - Re-computing ui and vj values:
x1A: u1 + vA = 6, vA = 6 x1C: u1 + vC = 10, vC = 10 x2C: u2 + vC = 11, u2 = 1
x3A: u3 + vA = 4, u3 = -2 x3B: u3 + vB = 5, vB = 7

- Cost changes for the empty cells, cij - ui - vj = kij;

x1B: k1B = c1B - u1 - vB = 8 - 0 - 7 = +1
x2A: k2A = c2A - u2 - vA = 7 - 1 - 6 = 0
x2B: k2B = c2B - u2 - vB = 11 - 1 -7 = +3
x3C: k2B = c2B - u3 - vC = 12 - (-2) - 10 = +4

- Since none of the values are negative, solution obtained is optimal.

- Cell 2A with a zero cost change indicates a multiple optimal solution.

The New ui and vj Values for the Second Iteration

341 342
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Example 61: a) Find the associated optimal cost

In a 3 3 transportation problem, let xij be the amount shipped
b) Determine the smallest values of cij associated with
from source i to destination j and cij be the corresponding
each non-basic variable that will maintain the
transportation cost per unit. The amounts of supply at sources
optimality of the northwest corner solution.
1, 2, and 3 are 15, 30, and 85 units, respectively; and the
demands at destinations 1, 2, and 3 are 20, 30, and 80 units,
respectively. Assume that the northwest corner solution is
optimal and that the associated values of the multipliers are

given by: u1 = -2, u2 = 3, u3 = 5, v1 = 2, v2 = 5, and v3 = 10.

343 344
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
Solution:
v1=2 v2=5 v3=10 Supply
Example 62:
0 ≥3 ≥8 15
u1=-2
15 The Onenote Co. produces a single product at three plants for
four customers. The three plants will produce 60, 80, and 40
5 8 ≥13
u2=3 5 25 30 25
units respectively. The firm has made a commitment to sell 40
≥7 10 5 15 80 units to customer 1, 60 units to customer 2, and at least 20
u3=5 85
units to customer 3. Both customers 3 and 4 also want to buy
20 5 30 5 80 as many of the remaining units as possible. The net profit
Demand
associated with shipping a unit from plant i to customer j is
Associated cost = 1475 given by the following table.

345 346
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Customer Solution: There are 3 sources (Plants 1, 2 and 3). Right now there
1 2 3 4
are 4 destinations (customers 1, 2, 3, and 4). The supplies ai at the
1 $800 $700 $500 $200
three sources are 60, 80, and 40 respectively. The demands at the
Plant 2 $500 $200 $100 $300 three destinations are: b1 = 40, b2 = 60, b3 20, b4 = ?
3 $600 $400 $300 $500
Since in a transportation model, all constraints are equalities, we
Management wants to know how many units to sell to
shall put b3 = 80 ( since customer 3 must get at least 20 units) and
customers 3 and 4 and how many units to ship from each of b4= 60 as the supply remaining after satisfying the three customers 1,
the plant to each of the customers to maximize profits. 2, and 3 is 60 and since customers 3 and 4 will buy as much as
Formulate the problem as a transportation model and solve it. possible.

347 348
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi
 Destination
But now the demand has become 240 and so we introduce a dummy 1 2 3 4 Supply
source SF with supply 60. Since the customers 1, 2 must definitely 1 -8 -7 -5 -2
60 0 60 0 1 - - -
get 40 and 60 units respectively, the dummy source cannot send any S -5 -2 -1 -3
2 80 40
o 40 40 2 2 2 -
amount to these destinations. This is achieved by putting the cost
from the dummy to these destinations as big M. Now the cost from u 3 -6 -4 -3 -5
20 20 40 20 1 1 2 2
dummy to the destinations 3 and 4 are put as zero. Also since these
r
SF M M 0 0
c 60 60 0 0 0 0
are actually profits, and since the transportation model is a
e
minimization problem, to maximize the total profit we take cij as Demand 40 60 80 6020

negative of the profits given. The starting tableau is given below. 2 3 2 2

1 - 2 2
- - 2 2
- - 3 5
349 350
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi

Destination
v1= -9 v2= -7 v3= -5 v4= -7

u1= 0 -8 -7 -5 -2
S -1 60 0 -5
o u2= 4 -5 -2 -1 -3
u 40 -5 0 40
r -6 -4 -3 -5
u3= 2 20 20
c -1 -9
e M M 0 0
u4= 5 60
-4-M -2-M -2

This is the optimal tableau.

Max Profit = - min z = - (-900) = $900.

351 352
EE 477 Dr. Mohamed Zribi EE 477 Dr. Mohamed Zribi