Lecture Day 11

Specialized Linear Programming Models

Transportation Model

Modified Distribution Method

When Supply Is Not Equal To Demand
Degeneracy

Transshipment Model
Assignment Model

More Popular Heuristics

Row-Minimum Method
Vogels Approximation Method

The Cordoba Company

Warehouse

To
From
Cavite

Laguna

Batangas
Demand

Baguio

Cebu

50

40
14

Iloilo

11

12

Davao

Supply
16

X
5

40

10

X
6

50

80

40

40

30

50

120
70

90

70

40

50

250

40

30

40

The Cordoba Company

Warehouse

To
From
Cavite

Baguio

Demand

Davao

Iloilo

11

16

12

50

Laguna

Batangas

Cebu

Supply

40
14

40

10

40

30

50

90

70

40

50

K1

K2

K3

K4

50

R1

80

R2

120

R3

250

Ri + Kj = Cij
R1 + K1 = 8
R2 + K2 = 7

R2 + K3 =
R3 + K1 =
R3 + K2 =
R3 + K4 =

5
14
10
7

R1 = 0
R2 = 3
R3 = 6

K1
K2
K3
K4

=
=
=
=

8
4
2
1

Ri + Kj = Cij
R1 + K1 = 8
R2 + K2 = 7

R2 + K3 =
R3 + K1 =
R3 + K2 =
R3 + K4 =

5
14
10
7

R1 = 0
R2 = 3
R3 = 6

K1
K2
K3
K4

=
=
=
=

8
4
2
1

Cij (Ri + Kj) = OC

9 (0 + 4) = + 5
11 (0 + 2) = + 9

16 (0 + 1) = +15
12 (3 + 8) = + 1
8 (3 + 1) = + 4
6 (6 + 2) = - 2

The Cordoba Company

Warehouse

To
From
Cavite

Baguio
8

Demand

Davao

Iloilo

11

16

40
+ 6

50

Laguna

Batangas

Cebu

Supply

12

14

40
- 10

40

30

90

70

50
40

50

50

80

120
250

The Cordoba Company

Warehouse

To
From
Cavite

Baguio

Demand

Davao

Iloilo

11

16

12

50

Laguna

Batangas

Cebu

Supply

70
14

10

10

40

30

50

90

70

40

50

K1

K2

K3

K4

50

R1

80

R2

120

R3

250

Ri + Kj = Cij
R1 + K1 = 8
R2 + K2 = 7

R2 + K3 =
R3 + K1 =
R3 + K3 =
R3 + K4 =

5
14
6
7

R1 =
R2 =
R3 =
K1 =
K2 =
K3 =
K4 =

0
5
6
8
2
0
1

Ri + Kj = Cij
R1 + K1 = 8
R2 + K2 = 7

R2 + K3 =
R3 + K1 =
R3 + K3 =
R3 + K4 =

5
14
6
7

R1 =
R2 =
R3 =
K1 =
K2 =
K3 =
K4 =

0
5
6
8
2
0
1

Cij (Ri + Kj) = OC

9 (0 + 2) = + 7
11 (0 + 0) = +11

16 (0 + 1) = +15
12 (5 + 8) = - 1
8 (5 + 1) = + 2
10 (6 + 2) = + 2

The Cordoba Company

Warehouse

To
From
Cavite

Laguna

Batangas
Demand

Baguio

Cebu

Davao

50
+ 12

10

70
-

14

40
90

70

Supply
Iloilo

11

16

10
+ 6

30

50

40

50

50

80

120
250

The Cordoba Company

Warehouse

To
From
Cavite

Laguna

Batangas
Demand

Baguio

Cebu

Supply

Davao

Iloilo

11

16

12

10

50
10

70

14

30

40

50

90

70

40

50

K1

K2

K3

K4

50

R1

80

R2

120

R3

250

Ri + Kj = Cij
R1 + K1 = 8
R2 + K1 = 12

R2 + K2 =
R3 + K1 =
R3 + K3 =
R3 + K4 =

7
14
6
7

R1 =
R2 =
R3 =
K1 =
K2 =
K3 =
K4 =

0
4
6
8
3
0
1

Ri + Kj = Cij
R1 + K1 = 8
R2 + K1 = 12

R2 + K2 =
R3 + K1 =
R3 + K3 =
R3 + K4 =

7
14
6
7

R1 =
R2 =
R3 =
K1 =
K2 =
K3 =
K4 =

0
4
6
8
3
0
1

Cij (Ri + Kj) = OC

9 (0 + 3) = + 6
11 (0 + 0) = +11

16 (0 + 1) = +15
5 (4 + 0) = + 1
8 (4 + 1) = + 3
10 (6 + 3) = + 1

The Cordoba Company

Warehouse

To
From

Baguio

Cavite

Cebu

Supply

Davao

Iloilo

11

16

12

10

50

Laguna

10

Batangas
Demand

70

14

30
90

70

40

50

40

50

Minimum transportation cost = P2,020

Next Topic

50

80

120
250

The Cordoba Company

Warehouse

To
From
Cavite

Baguio
8

50

Laguna

Batangas
Demand

Cebu

90

Davao

Supply
Iloilo

11

16

12

14

10

70

40

50

40

50

80

120

250

The Cordoba Company

Warehouse

To
From
Cavite

Baguio
8

50

Laguna

Batangas
Demand

Cebu

90

Davao

Supply
Iloilo

11

16

12

14

10

70

40

50

40

50

80

120

250

The Cordoba Company

Warehouse

To
From
Cavite

Baguio
8

50

Davao

Iloilo

11

12

Laguna

16

X
5

70
14

Batangas
Demand

Cebu

Supply

10

X
90

70

40

50

40

50

80

10

120
250

2
1

The Cordoba Company

Warehouse

To
From
Cavite

Baguio
8

50

Davao

Iloilo

11

12

Laguna

16

X
5

70
14

Batangas
Demand

Cebu

Supply

10

X
90

70

40

50

40

50

80

10

120
250

3
1

The Cordoba Company

Warehouse

To
From
Cavite

Laguna

Batangas
Demand

Baguio

Cebu

50

70
14

Iloilo

11

12

Davao

Supply
16

X
5

10

10

X
6

40

30

50

90

70

40

50

40

30

50

80

10

120
250

90

3
1

40

The Cordoba Company

Warehouse

To
From
Cavite

Baguio

Demand

Davao

Iloilo

11

16

12

50

Laguna

Batangas

Cebu

Supply

70
14

10

10

40

30

50

90

70

40

50

K1

K2

K3

K4

50

R1

80

R2

120

R3

250

Ri + Kj = Cij
R1 + K1 = 8
R2 + K2 = 7

R2 + K3 =
R3 + K1 =
R3 + K3 =
R3 + K4 =

5
14
6
7

R1 = 0
R2 = 5
R3 = 6

K1
K2
K3
K4

=
=
=
=

8
2
0
1

Ri + Kj = Cij
R1 + K1 = 8
R2 + K2 = 7

R2 + K3 =
R3 + K1 =
R3 + K3 =
R3 + K4 =

5
14
6
7

R1 = 0
R2 = 5
R3 = 6

K1
K2
K3
K4

=
=
=
=

8
2
0
1

Cij (Ri + Kj) = OC

9 (0 + 2) = + 7
11 (0 + 0) = +11

16 (0 + 1) = +15
12 (5 + 8) = - 1
8 (5 + 1) = + 2
10 (6 + 2) = + 2

The Cordoba Company

Warehouse

To
From
Cavite

Laguna

Batangas
Demand

Baguio

Cebu

Davao

50
+ 12

10

70
-

14

40
90

70

Supply
Iloilo

11

16

10
+ 6

30

50

40

50

50

80

120
250

The Cordoba Company

Warehouse

To
From
Cavite

Laguna

Batangas
Demand

Baguio

Cebu

Supply

Davao

Iloilo

11

16

12

10

50
10

70

14

30

40

50

90

70

40

50

K1

K2

K3

K4

50

R1

80

R2

120

R3

250

Ri + Kj = Cij
R1 + K1 = 8
R2 + K1 = 12

R2 + K2 =
R3 + K1 =
R3 + K3 =
R3 + K4 =

7
14
6
7

R1 =
R2 =
R3 =
K1 =
K2 =
K3 =
K4 =

0
4
6
8
3
0
1

Ri + Kj = Cij
R1 + K1 = 8
R2 + K1 = 12

R2 + K2 =
R3 + K1 =
R3 + K3 =
R3 + K4 =

7
14
6
7

R1 =
R2 =
R3 =
K1 =
K2 =
K3 =
K4 =

0
4
6
8
3
0
1

Cij (Ri + Kj) = OC

9 (0 + 3) = + 6
11 (0 + 0) = +11

16 (0 + 1) = +15
5 (4 + 0) = + 1
8 (4 + 1) = + 3
10 (6 + 3) = + 1

The Cordoba Company

Warehouse

To
From

Baguio

Cavite

Cebu

Supply

Davao

Iloilo

11

16

12

10

50

Laguna

10

Batangas
Demand

70

14

30
90

70

40

50

40

50

Minimum transportation cost = P2,020

50

80

120
250

 For maximization problems, either change

the signs of the objective function values
and minimize, or choose the non-basic cell
with the most positive Cij (Ri + Kj) value
to enter the solution; all other steps
remain the same.

When Supply
Is Not Equal To
Demand

 When supply is not equal to demand . . .

. . . adjust accordingly.

Make supply always equal to demand

by adding either a dummy supplier
or a dummy demand point.

Example: If Supply > Demand or Demand < Supply

Warehouse

To
From

Baguio

Cavite

Laguna

Batangas
Demand

90

Cebu

Supply
Davao

11

12

14

10

70

40

50

80

120

Example: If Supply > Demand or Demand < Supply

Warehouse

To
From

Baguio

Cavite

Laguna

Batangas
Demand

90

Cebu

Supply

Davao

Dummy

11

12

14

10

70

40

50

50

80

120
250

Units assigned to cells in the dummy column show

which suppliers have excess supply and by how much.

Example: If Supply > Demand or Demand < Supply

Warehouse

To
From
Cavite

Laguna

Batangas
Demand

Baguio

Cebu

50

40

14

Dummy

11

12

40

Davao

Supply
0

X
5

10

X
6

50

80

40

30

40

50

120
90

90

70

40

50

250

40

30

50

Example: If Demand > Supply or Supply < Demand

Warehouse

To
From

Baguio

Cavite

Laguna
Demand

90

Cebu

Supply

Davao

Iloilo

11

16

12

70

40

50

50

80

Example: If Demand > Supply or Supply < Demand

Warehouse

To
From

Baguio

Cavite

Laguna

Dummy
Demand

90

Cebu

Supply

Davao

Iloilo

11

16

12

70

40

50

50

80

120
250

Units assigned to cells in the dummy row show which

demand points would have demands not fully met.

The Cordoba Company

Warehouse

To
From
Cavite

Laguna

Dummy
Demand

Baguio

Cebu

50
12

40

Davao

Supply
11

X
7

40
0

Iloilo
16

X
5

X
0

X
0

50

80

40

30

40

50

120
90

90

70

40

50

250

40

30

50

Degeneracy

 Degeneracy occurs whenever a basic variable assumes

a zero value. This condition violates the M + N 1
basic cells requirement.
Degeneracy can arise in the following situations:
1. In determining a feasible starting solution,
fewer than M + N 1 cells are used.

Example: Using Northwest Corner Rule

Warehouse

To
From
Cavite

Laguna

Batangas
Demand

Baguio

Cebu

70

60

14

Iloilo

11

12

20

Davao

Supply
16

X
5

10

X
6

70

80

60

40

60

100
60

90

60

40

60

250

20

 Degeneracy occurs whenever a basic variable assumes

a zero value. This condition violates the M + N 1
basic cells requirement.
Degeneracy can arise in the following situations:
1. In determining a feasible starting solution,
fewer than M + N 1 cells are used.
2. In working toward optimality, there may have
been more than one basic cell to leave the
solution at an iteration. This results in fewer
than M + N 1 cells being basic.

The Cordoba Company

Warehouse

To
From
Cavite

Laguna

Batangas
Demand

Baguio

Cebu

Davao

50
+ 12

10

70
-

14

40
90

70

Supply
Iloilo

11

16

10
+ 6

30

50

40

50

50

80

120
250

The Cordoba Company

Warehouse

To
From
Cavite

Laguna

Batangas
Demand

Baguio

Cebu

Davao

50
+ 12

10

70
-

14

10
60

70

Supply
Iloilo

11

16

10
+ 6

30

50

40

50

50

80

90
220

 In the first situation, degeneracy is handled

by never deleting a row and a column of
the transportation tableau at the same
time. When the maximum supply and
demand values that can be assigned in a
cell are equal, allocate zero units to another
cell at another step in the heuristic.

Example: Using Northwest Corner Rule

Warehouse

To
From
Cavite

Laguna

Batangas
Demand

Baguio

Cebu

70

60

14

Iloilo

11

12

20

Davao

Supply
16

X
5

10

X
6

70

80

60

40

60

100
60

90

60

40

60

250

20

 In the second case, degeneracy arises

whenever there is more than one basic cell
in the closed loop that is minus labeled and
has the same minimum shipment amount.
In this situation, care must be taken to
delete only one of the cells.

The Cordoba Company

Warehouse

To
From
Cavite

Laguna

Batangas
Demand

Baguio

Cebu

Davao

50
+ 12

10

70
-

14

10
60

70

Supply
Iloilo

11

16

10
+ 6

30

50

40

50

50

80

90
220

Transshipment Model

 The transshipment problem is sometimes called the

minimum-cost network flow problem.
It is the most general of the linear network models
and contains transportation and assignment models
as special cases.
Its solution methodology is similar to the
transportation MODI method.

 The transshipment model differs from the

transportation model in that it allows nodes or points
that are neither pure origins nor pure destinations.

The origins and destinations of a transportation

problem are sources and sinks, respectively.

Source nodes can be thought of as having supply
available; sink nodes typically represent demand
points.
Transshipment models allow the presence of a
transshipment point that has arcs leading into
and out of the node. Transshipment nodes can
have a supply, or demand; they can also have a
supply or demand of zero.

General Transshipment Model

Let:

N = set of all arcs in the network

n = number of nodes in the network
Xij = amount of flow from node i to node j
Cij = per unit cost of flow from node i to node j
lij = lower bound on flow from node i to node j
uij = upper bound on flow from node i to node j
bj = supply (demand) at node j
Additionally, let (i,j) denote an arc in the

General Transshipment Model

S cij xij

Minimize:

(i,j) e N

Subject to:
total flow out total flow in = node supply or demand

S xjk - S xkj
(j,k) e N

= bj, j = 1, . . . , n

(k,j) e N

lij < xij < uij, all (i,j) e N

Example:
Supply Sites
29

Distribution
Centers

Demand Sites
5

-20

-28

-17

4
36

2
Source

Transshipment
Point

Sink

Let Xij = amount to be shipped

from source i to
destination j
where i = 1, 2, 3, 4
j = 3, 4, 5, 6, 7
Min. Z = 5X13 + 2X14 + 8X24
+ 8X35 + 6X36 + 10X45
+ 7X46 + 2X47
Subject to:
X13 > 0
X13 + X14 = 29
8 < X14 < 20
X24 = 36
X24 > 0
X35 + X36 X13 = 0
0 < X35 < 15
X45 + X46 + X47 X14 X24 = 0
X36 > 5
- X35 - X45 = -20
0 < X45 < 15
- X36 - X46 = -28
X46 > 0
- X47 = -17
10 < X47 < 20

Assignment Model

 The assignment model is a special case of the

transportation problem wherein the basic idea is to
assign n single elements rather than many units
from each source to destination.
It can be described as a transportation model with
an equal number of sources and destinations with
all supply and demand values equal to one (1).
Typical applications of the assignment model
include the least-cost or least-time assignment of
jobs to machines, workers to tasks, salespersons to
territories, and contracts to contractors.

 In general, if there are n agents to be assigned to

n tasks, the assignment model is stated as:
n

Minimize:

S S cijxij
i=1 j=1
n

Subject to:

S xij = 1,

i = 1,2, . . . , n

j=1
n

S xij = 1,

i = 1,2, . . . , n

i=1

xij > 0

for all i and j

Example:
Vice Presidents

To
From

London

Sec. 1

Sec. 2

Sec. 3
Demand

New York

Supply
Shanghai

11

12

14

10

1
3

End of Lecture Day 11