TRANSPORTATION

PROBLEM
When Supply meets Demand
 X D1 P
S C11
C12

Transportation Problem
1
C13
C14 Q
Y D2
C21
S C22
C23
2 C24
◦ A subset of the LPP. Z C31 D3 R
C32
S C33
◦ There is a set of m sources each having a fixed unit supply of a 3
C34
product/service. D4 S
◦ And, another set of n destinations each having a fixed units of demand of
that product/service.
◦ We are dealing with a single product/service.
◦ The product/service has to be transported from all the source to all the
destination.
◦ A cost is associated to transport each unit of the product/service from a
source to a destination.
◦ Find out the minimum cost for distribution of the units.

Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,

05/04/2025 GCETTS 2
 Tabular Representation
Destinations
Suppl
P P Q R S
X D1 y
S C11
C12 X C11 C12 C13 C14 S1
1
C13
C14 Q

Sources
Y D2
C21
S C22 Y C21 C22 C23 C24 S2
C23
2 C24
Z C31 D3 R
S C32 Z C31 C32 C33 C34 S3
C33
3
C34

D4 S Demand D1 D2 D3 D4

Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,

05/04/2025 GCETTS 3
 LPP Representation
Destinations

d1 d2 d3 dn Supply

C11 C12 C13 C1n

s1 S1
X11 X12 X13 X1n

Balanced Transportation Problem

Sources

C21 C22 C23 C2n

s2 S2 A transportation problem is balanced if
overall supply meets overall demand
X21 X22 X23 X2n
exactly.

Cm1 Cm2 Cm3 Cmn

sm Sm
Xm1 Xm2 Xm3 Xmn

Demand D1 D2 D3 Dn
Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,
05/04/2025 GCETTS 4
 Theorems
◦ The necessary and sufficient condition for a transportation problem to have a solution is
that the problem is balanced (the total demand equals the total supply).
◦ If the problem is unbalanced, we can add a dummy row / dummy column accordingly to make it a
balanced problem to solve it.
◦ When the total demand is more than total supply, we can add a dummy source(row) to meet the
excess demand.
Destinations
◦ When the total supply is more than total demand, we
can add a dummy destination(column) to meet the Suppl
P Q R S
excess supply. y
◦ A balanced transportation problem X C11 C12 C13 C14 30
X C C C C 30
always has a feasible solution. 11 12 13 14

Sources
◦ Any basic feasible solution has m+n-1 Y C21 C22 C23 C24 50
non-negative allocations. Y C21 C22 C23 C24 50
Z C31 C32 C33 C34 40
◦ Integer Solution Property.
Z
D C0
?31 C0
?32 C0
?33 C0
?34 40
30

Demand 20 50 30 50 150
120
05/04/2025
Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,
GCETTS
150 5
 Solving Transportation Problem
◦ Check if the problem is balanced. If not, add a dummy row / column accordingly to
make it a balanced one.
◦ Find the initial basic feasible solution
◦ North-West Corner Method
◦ Least Cost / Matrix Minima Method
◦ Vogel’s Approximation Method
◦ Find the optimal solution from the basic feasible solution
◦ MOdified DIstribution (MODI Method)

Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,

05/04/2025 GCETTS 6
 North West Corner Rule
◦ Start with the upper left corner (north-west corner) cell.
◦ Allocate maximum possible items at that cell.
◦ What is the maximum possible items that can be allotted to a cell?
◦ Minimum of supply and demand of that cell.
◦ Update supply and demand associated with the allocated cell by reducing the
allocated units from them.
◦ This will make either or both the supply and demand of the allocated cell as 0.
◦ If the supply becomes 0, then move to the cell just below the allocated cell and
reiterate the process.
◦ If the demand becomes 0, then move to the cell just to the right of the allocated
cell and reiterate the process.
◦ Thinking it in a recursive way.

Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,

05/04/2025 GCETTS 7
 North West Corner Rule
Destinations

D1 D2 D3 D4 Total
Supply
Cost

1 2 3 4 6
6 4+4=
S1 6
2
2 8
Sources
4 2
0

4 3 2 0 8
8 12 + 8
S2 8
4
4 = 20
4 4
0

0 2 2 1 10
10 8+6=
S3 10
6
6 14
4 6
0

4 8 6 24
6 42
Demand 4 6 8 6 24
0 4 4 0 44 0 0
Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,
05/04/2025 GCETTS 8
 Least Cost Method
◦ A.k.a. Matrix Minima method.
◦ Determine the cell (i, j) associated with the smallest cost C ij in the matrix.
◦ Break ties arbitrarily.
◦ Allocate maximum possible items Xij at that cell.
◦ If Xij = Dj, then remove the j-th column. If Xij = Si, then remove the i-th row. If Xij = Dj
= Si, then remove either the column or the row, not both.
◦ We will mark a ‘X’ to all the cells of a row or column denote the row or column is deleted.
◦ Update the other (i.e demand or supply, as applicable) by reducing the allocated
units.
◦ Repeat the process with the reduced matrix.

Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,

05/04/2025 GCETTS 9
 Least Cost Method
Destinations

D1 D2 D3 D4 Total
Supply
Cost

X 1 X 2 X 3 X 4 6
S1 6 12
Sources 6 0

X 4 X 3 X 2 X 0 8 4+0=
S2 8
2 6 2 4

X 0 X 2 X 2 X 1 10 0 + 12
S3 10 = 12
4 6 6

8 24
Demand 4 6 8 6 24
28
2
Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,
05/04/2025 GCETTS 10
 Vogel’s Approximation Method
◦ For each row and each column, find out the difference between the smallest
and next smallest cost, known as penalty cost.
◦ If there are two or more smallest value in a row or column, the penalty cost is 0.
◦ Next, choose the row or column associated with maximum penalty cost.
◦ Ties are broken arbitrarily.
◦ In the selected row or column, find out the cell having the least cost.
◦ Allocate maximum possible value in this cell based on the demand and
supply of it.
◦ Update the demand or supply accordingly by reducing the allocated units.
◦ Delete the row or column which is fully exhausted and reiterate the process.
◦ We will mark a ‘X’ to all the cells of a row or column denote the row or column is
deleted.
Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,
05/04/2025 GCETTS 11
 Destinations
VAM D1 D2 D3 D4 Supply
Row Total
Penalty Cost

X 1 X 2 X 3 X 4
S1 6 1 12
6
Sources

X 4 X 3 X 2 X 0 8 4+0=
S2 8 X
2
1 4
2 6 2

X 0 X 2 X 2 X 1 10 0 + 12
S3 10 X
2
1
0 = 12
4 6 6

6 8 24
Demand 4 6 8 6 24 28
0 66 0
Column
Penalty X
1 0 1
0 X
1
Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,
05/04/2025 GCETTS 12
 Loop
◦ A loop is a closed path consists of the cells of the matrix.
◦ An ordered sequence of at least four different is called a loop if following conditions
hold –
◦ Any two consecutive cells of a loop should lie in either
same row or same column.
◦ No three or more consecutive cells of a loop lie in the
1 2 3 4
same row or same column.
◦ The last cell is in the same row or column as the first cell.
◦ The vertices or the turning points of a loop are 4 3 2 0
allocated cells or basic variables.
◦ Loops cannot turn from an unallocated cell.
◦ A loop will always have even number of edges and
vertices. 0 2 2 1

Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,

05/04/2025 GCETTS 13
 Loop Pivoting
◦ Determine the entering variable.
◦ An entering variable is non-basic variable or unallocated cell.
◦ Find a loop starting from the entering variable and involving some of the BVs.
◦ Starting from entering variable, label each vertices of
the loop as alternating (+) and (-) sign. (+ 1 2 (-) 3 4
◦ With the entering variable labelled as (+). )
◦ Find the (-) cell having smallest allocated units (Xij). + θ
θ
0
◦ Ties are broken arbitrarily.
4 3 2 (+
(+ 0
◦ This cell corresponds to the leaving variable. (-)
(-) 4
4 3
3 2
2 )) 0
0
◦ Let the allocated units in this cell is θ, i.e. Xij = θ for this cell.
◦ Decrease each (-) cell of the loop by θ, and, increase +
-θ
each (+) cell of the loop by θ. 0 2 (+ 2 θ
1
0 2 (+
)
2 (-) 1
0 2 ) 2 (-) 1
Preetam K Sur, Assistant Professor, Computer Science & Engineering Department, +
05/04/2025 GCETTS -θ 14
 Loop Pivoting
(-) (+)

10
35 25

(+) (-)

35
10 20 0
25

(+) (-)

35
10 5
30
Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,
05/04/2025 GCETTS 15
 Modified Distribution (MODI)
◦ After getting the initial basic feasible solution, the next step is to check if the
solution is optimal.
◦ MODI Method is used to test and find the optimal solution eventually.
◦ A.k.a. U-V Method or Transport Simplex method
◦ Find out a set of values U and V for each row and column respectively such that for
each occupied cell (i, j), Cij = Ui + Vj.
◦ Start with U1 = 0.
◦ Then, calculate rest of the values.
◦ For each unallocated cell (i, j), find out the value W ij = Ui + Vj - Cij
◦ If all the Wij ≤ 0, then the current allocated cells (Basic Feasible Solution) produces optimal
solution.
◦ Otherwise, choose the unallocated cell(i.e. NBV) associated with most positive W ij as
entering variable. Ties are broken arbitrarily.
◦ Use loop pivoting to obtain a new Basic Feasible Solution and reiterate the process.
Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,
05/04/2025 GCETTS 16
 MODI Method
V1 V2 V3 V4

1 (+ U1 = 0
(-) 1 ) 22 3
3 4
4
U1 U1 + V1 = 1 => V1
4
4 2
2 -2
-2 -4
-4
=1
U1 + V2 = 2 => V2
4 3 (+ 2 0 =2
U2 4 (-) 3 ) 2 0
-2 4 4 1 U2 + V2 = 3 => U2
-2 4 4 1 =1

(+
0 2 2 1 U2 + V3 = 2 => V3
U3 0 2 (-) 2 1 =1
)
2 1 4 6
05/04/2025
2 1 Preetam 4 6 Computer
K Sur, Assistant Professor, Science & Engineering Department,
GCETTS U + V = 2 => U 17
3 3 3
 MODI Method
V1 V2 V3 V4

1 2 3 4 U1 = 0
U1 U1 + V2 = 2 => V2
-2 6 -2 -4 =2

4 3 (-) 2 (+
2 ) 00 U2 + V2 = 3 => U2
U2
-4 0 8
8 1
1
=1
U2 + V3 = 2 => V3
0 2 (+ 2 1 =1
U3 )
2 (-) 1
4 1 0 6
0 6 U3 + V3 = 2 => U3
Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,
05/04/2025 GCETTS =1 18
 MODI Method
V1 V2 V3 V4

1 2 3 4 U1 = 0
U1 U1 + V2 = 2 => V2
-2 6 -2 -5 =2

4 (-) 3 (+
3 ) 22 0
U2 + V2 = 3 => U2
U2
-4 0
0 2
2 6 =1
U2 + V3 = 2 => V3
0 (+ 2 2 1 =1
U3 )
2 (-) 2 U2 + V4 = 0 => V4
4 1 6 -1 = -1
1 Preetam 6
K Sur, Assistant Professor, Computer Science & Engineering Department,
05/04/2025 GCETTS U3 + V3 = 2 => U3 19
 MODI Method
V1 V2 V3 V4
Optimal Solution
1 2 3 4 U1 = 0
U1 U0
1
+V12= 2 => V
+ 2 2
= 12
-1 6 -1 -4 =2

4 3 2 0
U2 U24++V30= 2 => U2
-4 -1 2 6 = 0= 4
U2 + V4 = 0 => V4
0 2 2 1 =0
0 + 12
U3
= 12
4 0 6 -1
U + V = 2 => U
3 2 3
Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,
05/04/2025 GCETTS=0 20
 Thank You!

Preetam K Sur, Assistant Professor, Computer Science & Engineering Department,

05/04/2025 GCETTS 21