THE TRANSPORTATION MODEL

Background
Suppose there is an item, which is kept in geographically dispersed or
separated locations and is required similarly at geographically dispersed
locations. Each source can only supply a certain quantity in a given period
while each destination has periodic minimum requirements.
To transport an item from a given source is to a given destination requires
some money per unit.
The network

C11 D1
S1
C12

C21

C22

S2 D2

The problem – How many units should be transported from source to a given
destination in order to minimize total transportation cost or maximize the
schedule

The structure
1 2 N Supply

1 C1 C1 C a1
1 2 1n

1
 X11 X12 X1n
2 C2 C2 C a2
1 2 2n
X21 X22 X2n

M C C am
m1 mn
Xm1 Xmn
Demand b1 b2 bn

The symbol.
C ij ≡ Unit transportation cost from source i to destination j
X ij≡ Number of units transported from source i to destination j
(the unknown or the decision variables)
ai ≡ maximum capacity of source i in a given period.
j≡ Minimum requirements by destination j
Let C ≡ total transportation cost – to be minimized (objective function value)
O F Min C = C11,X11+C12,X12+………………………..+CmnXmn
St
1) Source constraints (maximum supply)
X11+X12+…….Xm1≤a1
X21+X22+…….X2m≤a2 etc

2) Destination constraints (minimum demand)

X11+X21+…….Xm1≥b1
X12+X22+…….Xm2≥b2 etc

3) Non- negativity
All x1 ≥ 0

Steps followed in solving a Transportation problem.

2
1. Ensure that the problem is balanced i.e total supply = total demand. If it
is not balanced, creates a dummy / fictitious source or destination to take
up the surplus.
2. Set up the initial transportation table, which will have supply for every
source, demand for every source to every destination and unit cost of
transportation an item from every source to every destination.
3. Obtain an initial allocation (solution) by using one of the following
methods.
i) North west corner method (NWC)
ii) Least cost sell method (LCCM)
iii) Vogel’s Approximation Method (VAM)
4. Test for degeneracy: This a problem that arises when the number of
satisfied cells is less than r+c-1. The difference between the two is called
the degree of degeneracy. In practice it means the solution to the
problem is indeterminate since it oscillates in a manner such that the
shadow prices of the unsatisfied cells cannot be fully determined. It
occurs when there is a simultaneous satisfaction of a cell row wise and
column wise.
Remedy: Fill up hypothetical cell(s) to the extent of the degree of
degeneracy in the row/column(s) that gave rise to the problem to
facilitate the determination of shadow prices of the unfilled up cells. No
units can be transferred from such a cell(s) but units can be loaded there
into.
5. Evaluate the initial solution for optimality. This requires that every empty
cell is assessed to determine if the solution can be improved by moving
units to that cell. At optimality, all shadow prices must be ≥ 0 for a
minimization problem. Zero implies that the solution is not unique (not
the only one ie there are alternatives as many are the number of zeros)
If the solution is not optimal, take units to the cell that has the highest
negative improvement index (also called shadow price or opportunity
cost). In case of a tie, take units to the cell that can accommodate the
highest number of units & in case of a further tie allocate arbitrary
Repeat iterations 4 and 5 till optimality.
There are 2 optimality evaluation methods.
i) Stepping stone method (SSM)
ii) Modified distribution method(MOD)
North West corner Rule (NWCR).

3
 A B C Supply

W 4 8 8
56 56

X 16 24 16
16 66 82

Y 8 16 24
36 41 77

Demand 72 102 41 215

Least cost cell method (LCCM) INITIAL SOLUTION.

We begin allocation with the cell which has the least cost progressing to the
next least cost cell which can accept units until all cells which can accept
units are satisfied. If there is a tie, take units to the cell, which can accept
more but if there is still a tie. allocate arbitrarily. This method gives a better
initial solution than NWC since we explicitly take the cost into account.

A B C Supply

W 4 8 8 56
56
X 1 2 16 82
6 4 41
41
Y 8 1 2 77
16 6 4
61
Demand 72 106 41 215

Vogel’s Approximation method (VAM) (initial solution)

VAM is a technique that is based on the concept of minimizing opportunity
cost which is given as the difference between the lowest cost and the second
lowest cost for each row and column for cells which can accept units.

4
Principle: If we fail to put units in the lowest cost cell in a given row or
column, by how much will we be losing in possible cost savings if we put
units in the next lowest cost cell? This method gives a very good initial
solution.

Steps
1. For each eligible row / column, find the difference between the lowest cost
and the next lowest cost cell which can take units. These are known as
‘’penalties’’ or ‘’opportunity cost’’
2. Identify the largest penalty and allocate as many units as possible to the
cell with the lowest cost in the row or column in which the largest penalty
falls. If there is a tie in the cost, allocate units to the cell which can take
more, but if there is still a tie in number of units, allocate arbitrarily
3. Calculate the penalties again and reallocate, continue with the iteration
until all eligible cells are filled.
When recalculating the penalties, cells which cannot take units do not
participate since their opportunity costs are zero.

Initial solution by VAM

A B C Supply penalties

W 56,0 4
4 8 8
56
X 1 1 82,41,0 8,8
6 24 6
41 41
Y 2 77,5,0 8,8
8 16 4
72 5
Demand 72,0 102,46,5, 41,0 215
0
penalties 4,8 8,8 8,8
Optimality check
1. Stepping stone method (SSM)
SSM is a method for evaluating a transportation problem for optimality which
uses the concept of opportunity cost.
The question is by moving units to an unfilled cell (removed from a filled cell)
is it possible to lower total transportation cost? If it is possible then the

5
problem is not optimal. This would be the case if there are negative shadow
prices which represent possible cost savings, To minimize the number of
iterations/ tables we take units to the cell with the highest positive shadow
price (this is the entering variable). The leaving variable would be the cell
with the smallest quantity to be subtracted. This process is continued until all
shadow prices are zero or positive.
Initial solution by LCCM
A B C Supply

W 4 8 8
56 56

X 16 24 16
41 41 82

Y 8 16 24
16 61 77

Demand 72 102 41 215

Degeneracy check.
r +c-1 =+3+3-1 = 5
Filled up/satisfied=5 hence not degenerate
Shadow prices of unsatisfied cells
WB=8-4+8-16=-4
WC=8-16+24-16+8-4=4
XA=0
YC=16
Not optimal due to –ve shadow price. Therefore take 56 units from cell WA
(leaving cell) to cell WB (entering cell). To maintain the row and column
totals, remove 56 units from cell YB and add the same to cell YA. Test for
optimality again (shadow prices shown in blue)
A B C Supply
W 4 56 8 56
X 0 41 41 82
Y 72 5 16 77
Demand 72 102 41 215

6
All the shadow prices are ≥0 hence solution is optimal but not unique. The
solution is
From To units*cost TC
W B 56*8= 448
X B 41*24= 984
X C 41*16= 656
Y A 72*8= 576
Y B 5*16= 80
2744
2. Modified distribution method (MODI)
MODI is an alternative method for SSM for optimality check which makes use
of row and column indices. These are computed using the fact that the
opportunity cost of a filled cell is zero. This enables a series of equation to be
obtained so that the various row and column indices can be found. This then
makes it possible to evaluate the empty cells.

ILLUSTRATION
Obtain the initial solution using LCCM and check for optimality using MODI
K1=?) K2=? K3=?
A B C Supply

W (r1=0) 4 8 8
56 56

X (r2=?) 16 24 16
41 41 82

Y (r2=?) 8 16 24
16 61 77

Demand 72 102 41 215

Notes;
1. Let ri and kj represent row i and column j indices respectively.
2. The opportunity cost of any cell is given by cij –ri – kj,
cij ≡ unit transportation cost from source i to destination j.
3. For any filled cell opportunity cost / shadow price is zero.

7
4. Thus for any filled cell, cij – Ri – Kj = 0, Cij =R1 +Kj
1. R1+K1= 4 No of unknown = 6
2. R2+K2 = 24 no of equation = 5
3. R2+K3 = 16 shortage in equation = 1
4. R3+K1 = 8 traditionally R1=0
5. R3+K2 = 16
This means that R3 =4, k2=12, r2=12, k3=4 hence
Shadow the prices
WB=+8-12-0=-4
XA=0
WC=+8-4=4
YC=16
Therefore take 56 units from cell XA (leaving cell) to cell WB (entering cell).
To maintain the row and column totals, remove 56 units from cell YB and add
the same to cell YA. Test for optimality again

A (k1=0) B (k2=8) C Supply

(k3=0)
W (r1=0) 4 56 8 56
X (r2=16) 0 41 41 82
Y (r3=8) 72 5 16 77
Demand 72 102 41 215
The shadow prices are shown in blue (all >=0) hence solution is optimal but
not unique.

SPECIAL CASES IN TRANSPORTATION MODEL.

1. Unbalance problem.
Is where total demand and supply are not equal. If a transportation
problem is not balanced, create a dummy/ fictitious source or destination
to take the excess supply or demand.
Specifically if:
Supply > demand create dummy destination
Supply < demand create dummy source.
Since dummy cells are nonexistent, unit transportation for such cells are
zero.
When using LCCM or VAM for initial solution, we ignore the dummy cells
since the objective is not to allocate units to dummy cells. For VAM too,
dummy cells do not participate in calculation of the penalties.
2. Infeasible or impossible routes.
Is where it is not possible to transfer units from a particular source to a
certain destination due to restrictions. Ignore the route or attach a very
large unit transportation cost (for minimization) or a very low unit

8
 contribution (for maximization) such that it is prohibitive to put units in
such a cell and continue with the usual transportation procedures.

ILLUSTRATION.
Suppose source W can supply an additional 20 units i.e 76 units in total
and also W cannot supply A, Obtain the initial solution using VAM and
check for optimality using MODI.
Excess supply = 20 units ¿> ¿ create a dummy destination to take these
additional 20 units.

A B C Dummy Supply

W 76 76
X 0 21 41 20 82
Y 72 5 77
Demand 72 102 41 20 235

Degeneracy check
R+C=1=3+4+1 = 6
Number of official cell = 6
Remarks
Initial solution is optimal since all shadow prices are ≥0
The optimal solution is not unique since there is a zero shadow price in cell
Xa.
Dummy allocation – plant X will manufacture 82-20 = 62 units => .It will
have idle/ excess capacity of 20 units.
TC = Sh 2424.
3. Degeneracy
Degeneracy occurs when a row and column totals are simultaneously
satisfied by an entry or filling of a common cell.
ILLUSTRATION
Solve the following transportation problem for minimization use LCCM for
initial solution and MODI for optimal check.
TABLE 1
1 2 3 Suppl
y
A -3 40 40

9
B 35 +5 35
C 0 30 10 40
Demand 35 70 10 115
Degeneracy check.
R+c-1=3+3-1=5
Number of filled cells =4
Problem is degenerated – 1 degree
Solution not optimal since there are positive shadow prices
Entering variable = A1
Leaving variable = C3
TABLE 2
1 2 3 Suppl
y
A 10 30 +2 40
B 25 +2 10 35
C 03 40 +3 40
Demand 35 70 10 115
Degeneracy check.
R+c-1=3+3-1=5
Number of filled cells =5
Problem not degenerated – positive degree
Remarks
Solution is optimal since all shadow prices are positive.
Solution is unique since no zero shadow price
TC =10x3+30x4+25x4+10x2+40x4 =sh 380

10