Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

3.

Transportation Problem
It is a special case of LP model that deals with commodities shipped from sources (e.g factories) to
destinations (e.g warehouses). The objective of a transportation problem is to determine the amount
shipped from each source to each destination where the total shipping cost is minimized and the
supply and demand requirements are satisfied.

General description of a transportation problem


i Has m supply points. Supply point i can supply at most si units.

ii Has n demand points. Demand point j must receive at least d j units.

iii Each unit produced at supply point i and shipped to demand point j incurrs a cost ci j

This information can be presented in a diagram as shown below

·-·I
I •

J .

. -- - - - li

I I
'
)l '
1
.:I I
-(

l I
I •

The arcs linking the sources to destinations represent the routes between them. Arc (ij) joining
source i to destination j contains the transportation cost per unit ci j and the amount shipped xi j . The

1
general LP model will be given as
m n
Min ∑ ∑ ci j xi j
i=1 j=1
n
s.t ∑ xi j ≤ si(i = 1, 2, ..., m)supply constraint
j=1
m
∑ xi j ≥ d j ( j = 1, 2, ..., n)demand constraint
i=1
with xi j ≥ 0

If ∑m n
i=1 si = ∑ j=1 d j then supply=demand and the problem is said to be a balanced transportation
problem. The general LP model of a balanced transportation problem is given as
m n
Min ∑ ∑ ci j xi j
i=1 j=1
n
s.t ∑ xi j = si(i = 1, 2, ..., m)supply constraint
j=1
m
∑ xi j = d j ( j = 1, 2, ..., n)demand constraint
i=1
with xi j ≥ 0

Example 3.1
Power Co. has 3 power plants that supply the needs of 4 cities. Power plants 1,2 and 3 can supply
35M,50M and 40M Kwh respectively. Peak power demand in the cities are 45M,20M,30M and 30M
Kwh respectively. The costs for sending 1M Kwh of electricity from plant to city is given in the table
below.

From city 1 city 2 city 3 city 4 Supply


Plant 1 8 6 10 9 35
Plant 2 9 12 13 7 50
Plant 3 14 9 16 5 40
Demand 45 20 30 30
This is a balanced transportation problem since demand=supply. The LP model will be given as

Min 8x11 + 6x12 + 10x13 + 9x14 + 9x21 + 12x22 + 13x23 + 7x24 + 14x31 + 9x32 + 16x33 + 5x34
s.t x11 + x12 + x13 + x14 = 35
x21 + x22 + x23 + x24 = 50
x31 + x32 + x33 + x34 = 40
x11 + x21 + x31 = 45
x12 + x22 + x32 = 20
x13 + x23 + x33 = 30
x14 + x24 + x34 = 30
with xi j ≥ 0

2
Balancing a transportation problem
Total supply > Total demand
In this case, create a dummy demand point with demand equal to the excess supply. Shipments to the
dummy point are assigned a cost of zero.
In example 3.1 suppose the demand for city 1 is 40M Kwh. Total demand=120 and total supply=125.
Add a dummy demand point with a demand of 5M Kwh.

Example 3.2
Using example 3.1, the table will be given as

From city 1 city 2 city 3 city 4 Supply


Plant 1 8 6 10 9 0 35
Plant 2 9 12 13 7 0 50
Plant 3 14 9 16 5 0 40
Demand 40 20 30 30 5

Total supply < Total demand


In this case, the problem has no feasible solution. However we may allow the possibility of leaving
some demand unmet then a penalty is associated with the unmet demand. In this case a dummy supply
point is created.

Basic feasible solutions for balanced transportation problems


Balanced transportation problems with m supply and n demand points have m+n equality constraints.
NOTE: 1. When solving a balanced transportation problem we may omit any one of the problem’s
constraints and solve an LP having m+n-1 constraints. Therefore a basic feasible solution must have
m+n-1 basic variables. 2. The costs are not needed to find a basic feasible solution.
There are 3 methods for obtaining a basic feasible solution (bfs) in a transportation problem. These
are;
1. North West Corner method

2. Minimum/Least Cost Method

3. Vogel’s Approximation Method

North West Corner method


The steps are;
i Begin at the north-west corner cell and allocate the minimum of s1 and d1

ii If the first row is exhausted proceed to the second row of the same column. However if the first
column is the one exhausted proceed to the next cell in the same row.

iii Proceed in the same manner until all supply and demand is satisfied.

3
Least cost method
The steps are;
i Select the cell with the lowest cost and allocate the minimum of si and d j

ii Repeat this until all the supply and demand is satisfied.

Vogel’s Approximation method


The steps are;
i Determine the penalty for each row and column by subtracting the smallest cost from the next
smallest cost.

ii Identify the row or column with the largest penalty and allocate the minimum of si and d j to
the cell with the least cost in the identified row or column.

iii Repeat the first two steps until all supply and demand is satisfied.

Example
Using Example 3.1, we determine the basic feasible solution using the 3 methods.
North-west corner method

From city 1 city 2 city 3 city 4 Supply


Plant 1 8 (35) 6 10 9 35
Plant 2 9 (10) 12 (20) 13 (20) 7 50
Plant 3 14 9 16 (10) 5 (30) 40
Demand 45 20 30 30
The solution is x11 = 35, x21 = 10, x22 = 20, x23 = 20, x33 = 10, x34 = 30. The total cost will be 1180.

Least cost method

From city 1 city 2 city 3 city 4 Supply


Plant 1 8 6 (20) 10 (15) 9 35
Plant 2 9 (45) 12 13 (5) 7 50
Plant 3 14 9 16 (10) 5 (30) 40
Demand 45 20 30 30
The solution is x12 = 20, x13 = 15, x21 = 45, x23 = 5, x33 = 10, x34 = 30. The total cost will be 1050.

Vogel’s Approximation method

From city 1 city 2 city 3 city 4 Supply Penalty


Plant 1 8 6 (10) 10 (25) 9 35 3,2,2,2,
Plant 2 9 (45) 12 13 (5) 7 50 2,3,3,4,
Plant 3 14 9 (10) 16 5 (30) 40 4,5,
Demand 45 20 30 30 125
Penalty 1,1,1,1, 3,3,6, 3,3,3,3, 2,

4
The solution is x12 = 10, x13 = 25, x21 = 45, x23 = 5, x32 = 10, x34 = 30. The total cost will be 1020.

NOTE: The bfs must have m+n-1 basic variables.

Exercise
1. MG Auto has 3 plants in Los Angeles, Detroit and New Orleans and 2 major distribution cen-
ters in Denver and Miami. The capacities of the 3 plants are 1000, 1500 and 1200 cars. The
demands at the 2 distribution centers are 2300 and 1400 cars.The transportation cost per car on
the different routes are given below in dollars.
Denver Miami
Los Angeles 80 215
Detroit 100 108
New Orleans 102 68
Find the bfs for the problem using the 3 methods. Which method gives the minimum cost?

2. The SunRay transport company ships truckloads of grain from 3 silos to 4 mills. The supply and
demand (in truckloads) together with the unit transportation cost per truckload on the different
routes are summarized below.The unit transportation costs are in hundreds of dollars.
Mill 1 Mill 2 Mill 3 Mill 4 Supply
Silo 1 10 2 20 11 15
Silo 2 12 7 9 20 25
Silo 3 4 14 16 18 10
Demand 5 15 15 15
Find the bfs for the problem using the 3 methods. Which method gives the minimum cost?

Transportation Simplex method


To obtain the pivot element we only use addition and subtraction. The steps in obtaining the
optimal solution of a transportation problem are as follows;

(a) Determine a starting basic feasible solution. This is mostly determined using the north-
west corner method.

(b) To determine the entering variable we use method of multipliers. Associate multiplier ui
and v j with row i and column j

(c) For each current basic variable xi j the multiplier should satisfy ui + v j = ci j .

(d) Use the computed value of the multipliers to evaluate the non-basic variables by com-
puting ui + v j − ci j .The non-basic variable with the most positive value is the entering
variable.

(e) To determine the leaving variable construct a closed loop that starts and ends at the en-
tering variable cell.The loop consists of connected horizontal and vertical lines only. No
diagonals are allowed. Each corner of the resulting loop with the exception of the entering
variable cell must coincide with a current basic variable.

5
(f) Alternate between + and - in each corner of the loop starting with a + at the entering vari-
able cell. For the cells with - identify the smallest value. let this value be θ .The variable
corresponding to value θ becomes the leaving variable.

(g) Add θ to the cells with + and subtract θ from the cells with -. The other cells remain
unchanged.

(h) Repeat process (b) to (g) until for all non-basic variables ui + v j − ci j are all negative or
zero.

Example
Using example 3.1, the starting basic feasible solution using the North-west corner method is given
as;
From city 1 city 2 city 3 city 4 Supply
Plant 1 8 (35) 6 10 9 35
Plant 2 9 (10) 12 (20) 13 (20) 7 50
Plant 3 14 9 16 (10) 5 (30) 40
Demand 45 20 30 30
We obtain m+n-1 basic variables. Using these basic variables we obtain our multipliers by ui + v j =
ci j . Begin by letting u1 = 0

x11 u1 + v1 = 8 v1 = 8
x21 u2 + v1 = 9 u2 = 1
x22 u2 + v2 = 12 v2 = 11
x23 u2 + v3 = 13 v3 = 12
x33 u3 + v3 = 16 u3 = 4
x34 u3 + v4 = 5 v4 = 1
Having obtained the multipliers, we use them to determine the entering variable from the non-basic
variables. Thus using ui + v j − ci j we obtain;

x12 u1 + v2 − 6 5
x13 u1 + v3 − 10 2
x14 u1 + v4 − 9 −8
x24 u2 + v4 − 7 −5
x31 u3 + v1 − 14 −2
x32 u3 + v2 − 9 4
The most positive variable is x12 and this becomes the entering variable.
To determine the leaving variable we construct a closed loop that starts and ends at the entering vari-
able cell such that the corners coincide with a basic variable.

6
r

I'
~
l3
U~ I ~ -
-
'
~===--'--~
- --
' .
--
______,_,__,.,
--- ~ ----
~ s
.

Smallest value with negative is 20 thus the leaving variable is x22 and θ = 20. The new table
becomes

6
20
13

30 (20)
16
(30

We compute the multipliers again using the new basic variables, Begin by letting u1 = 0 and de-
termine the leaving variable from the non-basic variables. Thus

7
Basic variables Non-basic variables
x11 u1 + v1 = 8 v1 = 8 x13 u1 + v3 − 10 2
x12 u1 + v2 = 6 v2 = 6 x14 u1 + v4 − 9 -8
x21 u2 + v1 = 9 u2 = 1 x22 u2 + v2 − 12 -5
x23 u2 + v3 = 13 v3 = 12 x24 u2 + v4 − 7 -5
x33 u3 + v3 = 16 u3 = 4 x31 u3 + v1 − 14 -2
x34 u3 + v4 = 5 v4 = 1 x32 u3 + v2 − 9 1
The most positive variable is x13 and this becomes the entering variable.
To determine the leaving variable we construct a closed loop that starts and ends at the entering vari-
able cell such that the corners coincide with a basic variable.
t I

--
8 G
1

I · IU

Smallest value with negative is 15 thus the leaving variable is x11 and θ = 15. The new table
becomes

ll ~ 13
5- \

8
We compute the multipliers again using the new basic variables, Begin by letting u1 = 0 and de-
termine the leaving variable from the non-basic variables. Thus

Basic variables Non-basic variables


x12 u1 + v2 = 6 v2 = 6 x11 u1 + v3 − 8 -2
x13 u1 + v3 = 10 v3 = 10 x14 u1 + v4 − 9 -10
x21 u2 + v1 = 9 v1 = 6 x22 u2 + v2 − 12 -3
x23 u2 + v3 = 13 u2 = 3 x24 u2 + v4 − 7 -5
x33 u3 + v3 = 16 u3 = 6 x31 u3 + v1 − 14 -2
x34 u3 + v4 = 5 v4 = −1 x32 u3 + v2 − 9 3
The most positive variable is x32 and this becomes the entering variable.
To determine the leaving variable we construct a closed loop that starts and ends at the entering vari-
able cell such that the corners coincide with a basic variable.

V6 V 6 V

-20
- +

19 12
H5
l6 S
Ug 6
30

Smallest value with negative is 10 thus the leaving variable is x33 and θ = 10. The new table
becomes

V:G V6 V3 10 V2
6

1025
12
h3 H5 5
16
Ua 3 10 30

We compute the multipliers again using the new basic variables, Begin by letting u1 = 0 and de-
termine the leaving variable from the non-basic variables. Thus

9
Basic variables Non-basic variables
x12 u1 + v2 = 6 v2 = 6 x11 u1 + v3 − 8 -2
x13 u1 + v3 = 10 v3 = 10 x14 u1 + v4 − 9 -7
x21 u2 + v1 = 9 v1 = 6 x22 u2 + v2 − 12 -3
x23 u2 + v3 = 13 u2 = 3 x24 u2 + v4 − 7 -2
x32 u3 + v2 = 9 u3 = 3 x31 u3 + v1 − 14 -5
x34 u3 + v4 = 5 v4 = 2 x33 u3 + v3 − 16 -3
All the variables are negative thus the solution is optimal.The solution is given as x12 = 10, x13 =
25, x21 = 45, x23 = 5, x32 = 10, x34 = 30 The total cost is 1020.

Exercise
Obtain the optimal solution for the MG Auto and SunRay company exercises.

Sensitivity analysis for transportation problem


Sensitivity analysis is performed on the optimal solution. The sensitivity analysis will focus on

1. Changing the objective function of a non-basic variable.

2. Changing the objective function of a basic variable.

3. Increasing a single supply by ∆ and a single demand by ∆

Changing the objective function coefficient of a non-basic variable


In the Simplex method changing the objective function of a non-basic variable leaves the RHS of the
optimal table unchanged and in Row 0 only the coefficient of the variable changes. As long as the
coefficient of the variable in Row 0 remains ≤ 0 the current solution remains optimal.

Example: In Example 3.1, for what range of values of the cost for shipping a unit of power from
plant 1 to city 1 will the current solution remain optimal?

Solution: x11 is non-basic. Cost of x11 =8, but c11 changes to 8 + ∆. For the solution to remain
optimal,

u1 + v1 − c11 ≤ 0
0 + 6 − (8 + ∆) ≤ 0
−2 − ∆ ≤ 0
∆ ≥ −2
Therefore
c11 ≥ −2 + 8
c11 ≥ 6

10
Changing the objective function coefficient of a basic variable
This affects the whole of Row 0 in the Simplex method therefore we obtain new u0i s and v0j s. The
current solution remains optimal as long as for all the non-basic variables ui + v j − ci j ≤ 0

Example: In Example 3.1, for what range of values of the cost for shipping a unit of power from
plant 1 to city 3 will the current solution remain optimal?

Solution: x13 is a basic variable. Cost of x13 =10, but c13 changes to 10 + ∆. The ui , v j equations
become, (let u1 = 0)

Basic variables Non-basic variables


x12 u1 + v2 = 6 v2 = 6 x11 u1 + v3 − 8 ∆−2
x13 u1 + v3 = 10 + ∆ v3 = 10 + ∆ x14 u1 + v4 − 9 -7
x21 u2 + v1 = 9 v1 = 6 + ∆ x22 u2 + v2 − 12 −∆ − 3
x23 u2 + v3 = 13 u2 = 3 − ∆ x24 u2 + v4 − 7 −∆ − 2
x32 u3 + v2 = 9 u3 = 3 x31 u3 + v1 − 14 ∆ − 5
x34 u3 + v4 = 5 v4 = 2 x33 u3 + v3 − 16 ∆ − 3
Solving for ∆ we obtain ∆ ≤ 2, ∆ ≥ −3, ∆ ≥ −2, ∆ ≤ 5, ∆ ≤ 3. This simplifies to −2 ≤ ∆ ≤ 2 and thus
8 ≤ c13 ≤ 12

Increasing both supply si and demand d j by ∆


This change maintains a balanced transportation problem. If the current solution remains optimal
then
Newtotalcost = oldtotalcost + ∆ui + ∆v j
Example: Increase plant 1 supply by 1 and city 2 demand by 1 then the new cost will be 1020 +
1(0) + 1(6) = 1026

Example: Increase plant 1 supply by 2 and city 2 demand by 2 then the new cost will be 1020 +
2(0) + 2(6) = 1032

Exercise: In example 3.1

i Determine the range of values of c14 for which the current solution remains optimal.

ii Determine the range of values of c34 for which the current solution remains optimal.

iii If s2 and d3 are both increased by 3 what is the new optimal solution?

iv If s3 and d3 are both decreased by 2 what is the new optimal solution?

11

You might also like