Dr.
Paresh Prajapati
P-4 OPERATIONS RESEARCH [OR]
Areas of Operations Research:
1. Linear Programing Problem (LPP)
2. Transportation Problem
3. Assignment Problem
4. Replacement Problem
5. Decision Theory
Working of OR
a. Define the problem
b. Prepare a mathematical model
c. Solve the problem using models
d. Verify the solutions
e. Implement the solutions.
Unit-2
Transportation Problem (TP Problem)
It is a method of transporting goods from several manufacturing units (Origins)
to several places where goods are required (Destinations) in such a way that
requirement of all the destinations fulfils and stock is also completely used and
the total cost of transportation is minimized in order to maximise the total profit.
Methods of Transportation Problem:
1. North-West Corner Rule
2. Matrix Minima Method
3. Vogel’s Approximation Method (VAM)
1
Que: Explain the mathematical form of Transportation Problem:
Mathematical form of Transportation Problem:
Let 𝑂1 ,𝑂2 ,….𝑂𝑖 ……𝑂𝑚 be the m origins where the goods are manufactured and
𝐷1 ,𝐷2 ,……. 𝐷𝑗 ,…… 𝐷𝑛 be the n destinations where the goods are to be supplied.
Let 𝑎1 , 𝑎2 , …….𝑎𝑖 …..𝑎𝑚 be the stock of the origins 𝑂1 ,𝑂2 ,….𝑂𝑖 ……𝑂𝑚
respectively.
Similarly, the requirements (demands) of the destinations
𝐷1 ,𝐷2 ,……. 𝐷𝑗 ,…… 𝐷𝑛 are 𝑏1 , 𝑏2 , …….𝑏𝑗 .…..𝑏𝑛 respectively.
Let the cost of transferring one unit from ith origin to jth destination is 𝐶𝑖𝑗 .
Since, our objective is to find the number of units to be transferred from different
origins to different destinations in such a way that the total transportation cost is
minimized, let us define
𝑥𝑖𝑗 = Number of units transferred from ith origin to jth destination.
Destination
𝑫𝟏 𝑫𝟐 …. 𝑫𝒋 …… 𝑫𝒏 Supply
𝐎𝟏 𝑪𝟏𝟏 𝑪𝟏𝟐 …. 𝑪𝟏𝒋 …… 𝑪𝟏𝒏 𝑎1
𝒙𝟏𝟏 𝒙𝟏𝟐 𝒙𝟏𝒋 𝒙𝟏𝒏
𝐎𝟐 𝑪𝟐𝟏 𝑪𝟐𝟐 …. 𝑪𝟐𝒋 …… 𝑪𝟐𝒏 𝑎2
𝒙𝟐𝟏 𝒙𝟐𝟐 𝒙𝟐𝒋 𝒙𝟐𝒏
…..
…..
…..
…..
…..
…..
….. …..
Origin
𝐎𝒊 𝑪𝒊𝟏 𝑪𝒊𝟐 …. 𝑪𝒊𝒋 …… 𝑪𝒊𝒏 𝑎𝑖
𝒙𝒊𝟏 𝒙𝒊𝟐 𝒙𝒊𝒋 𝒙𝒊𝒏
…..
…..
….. …..
…..
…..
…..
…..
𝐎𝒎 𝑪𝒎𝟏 𝑪𝒎𝟐 …. 𝑪𝒎𝒋 …… 𝑪𝒎𝒏 𝑎𝑚
𝒙𝒎𝟏 𝒙𝒎𝟐 𝒙𝒎𝒋 𝒙𝒎𝒏
Demand 𝑏1 𝑏2 …. 𝑏𝑗 …… 𝑏𝑛 Total
2
Let us prepare mathematical model as under.
Our objective is to find the total minimum transportation cost(Z). Thus, the
objective function can be stated as –
Optimize (the total cost) 𝑍 = ∑𝑚 𝑛
𝑖=1 ∑𝑗=1 𝐶𝑖𝑗 . 𝑥𝑖𝑗 -------- (1)
Above objective function is to be optimized subject to the following
constraints/limitations/conditions.
𝑥11 + 𝑥12 + ⋯ … . +𝑥1𝑗 + ⋯ + 𝑥1𝑛 = 𝑎1
𝑥21 + 𝑥22 + ⋯ … . +𝑥2𝑗 + ⋯ + 𝑥2𝑛 = 𝑎2
….. ….. …... ….. ….. …..
𝑥𝑖1 + 𝑥𝑖2 + ⋯ … . +𝑥𝑖𝑗 + ⋯ + 𝑥𝑖𝑛 = 𝑎𝑖
….. ….. …... ….. ….. …..
𝑥𝑚1 + 𝑥𝑚2 + ⋯ … . +𝑥𝑚𝑗 + ⋯ + 𝑥𝑚𝑛 = 𝑎𝑚
Which is a set of m constraints, one for each row (origin).
This set of m constraints (equations) can represented by the one common
equation also, as under.
i.e. ∑𝑛𝑗=1 𝑥𝑖𝑗 = 𝑎𝑖 where, i=1,2….m
Similarly, we can write constraints for columns (Destinations) as well.
𝑥11 + 𝑥21 + ⋯ … . +𝑥𝑖1 + ⋯ + 𝑥𝑚1 = 𝑏1
𝑥12 + 𝑥22 + ⋯ … . +𝑥𝑖2 + ⋯ + 𝑥𝑚2 = 𝑏2
….. ….. …... …..
𝑥1𝑗 + 𝑥2𝑗 + ⋯ … . +𝑥𝑖𝑗 + ⋯ + 𝑥𝑚𝑗 = 𝑏𝑗
….. ….. …... …..
𝑥1𝑛 + 𝑥2𝑛 + ⋯ … . +𝑥𝑖𝑛 + ⋯ +𝑥𝑚𝑛 = 𝑏𝑛
Which is a set of n constraints, one for each column (destination)
This set of n constraints (equations) can also be represented by the one common
equation.
3
i.e. ∑𝑚
𝑖=1 𝑥𝑖𝑗 = 𝑏𝑗 where, j=1,2….n.
Since we are studying the balanced problem only, total demand and total supply
are always equal and this condition can be explained as under.
∑𝑚 𝑛
𝑖=1 𝑎𝑖 = ∑𝑗=1 𝑏𝑗
Therefore, every non-degenerated (i.e. demand and supply for particular cell are
not equal) transportation problem will have m+n-1 independent solutions.
Methods of solving Transportation Problem:
Methods
1. North-West Corner Rule
2. Matrix Minima Method
3. Vogel’s Approximation Method:
Assignment Problem.
In the assignment problem different Jobs/Machines are to be assigned to the
different workers with different efficiencies in such a way that the total coast or
time of assignment become minimum.
Mathematical Form of Assignment Problem.
In this method we shall remember that there is a one-one association
/correspondence between work and worker i.e. one job can be assigned to one
worker only. This implies that the number of origins and the number of
destinations are always equal in number in the assignment problem. i.e. the cost
matrix is of square type.
Let us define the assignment problem, i.e. let us try to understand the each and
every aspect of the assignment problem as under.
Let, the 𝑂1 ,𝑂2 ,….𝑂𝑖 ……𝑂𝑛 be the n origins/workers which are to be assigned to
the n Destinations (Jobs/Machines) denoted by 𝐷1 ,𝐷2 ,……. 𝐷𝑗 ,…… 𝐷𝑛 .
Let the supply from the origins 𝑂1 ,𝑂2 ,….𝑂𝑖 ……𝑂𝑛 are 𝑎1 , 𝑎2 , …….𝑎𝑖 …..𝑎𝑛
respectively.
4
Here it is to be noted that the values of each 𝑎𝑖 is 1 because only one worker is
to be assigned one job.
Similarly Requirements of the destinations 𝐷1 ,𝐷2 ,……. 𝐷𝑗 ,…… 𝐷𝑛 are 𝑏1 , 𝑏2 ,
…….𝑏𝑗 .…..𝑏𝑛 respectively.
Here also it is to be noted that the values of each 𝑏𝑗 is 1 because only one job is
to be assigned to one worker only.
𝐶𝑖𝑗 = Cost/Time of assigning one unit from ith origin to the jth destination.
𝑥𝑖𝑗 = Number of units assigned from ith origin to the jth destination.
Here,
𝑥𝑖𝑗 = 1 if assignment is given
= 0 if assignment is not given
Let us represent the above information in the tabular form as under.
Destination(Machines)
𝐷1 𝐷2 …. 𝐷𝑗 …… 𝐷𝑛 Supply
𝑂1 𝑪𝟏𝟏 𝑪𝟏𝟐 …. 𝑪𝟏𝒋 …… 𝑪𝟏𝒏 𝑎1 =1
𝒙𝟏𝟏 𝒙𝟏𝟐 𝒙𝟏𝒋 𝒙𝟏𝒏
𝑂2 𝑪𝟐𝟏 𝑪𝟐𝟐 …. 𝑪𝟐𝒋 …… 𝑪𝟐𝒏 𝑎2 =1
𝒙𝟐𝟏 𝒙𝟐𝟐 𝒙𝟐𝒋 𝒙𝟐𝒏
Origin(Worker)
…..
…..
…..
…..
…..
…..
….. …..
𝑂𝑖 𝑪𝒊𝟏 𝑪𝒊𝟐 …. 𝑪𝒊𝒋 …… 𝑪𝒊𝒏 𝑎𝑖 =1
𝒙𝒊𝟏 𝒙𝒊𝟐 𝒙𝒊𝒋 𝒙𝒊𝒏
…..
…..
….. …..
…..
…..
…..
…..
𝑂𝑛 𝑪𝒏𝟏 𝑪𝒏𝟐 …. 𝑪𝒏𝒋 …… 𝑪𝒏𝒏 𝑎𝑛 =1
𝒙𝒏𝟏 𝒙𝒏𝟐 𝒙𝒏𝒋 𝒙𝒏𝒏
Demand 𝑏1 =1 𝑏2 =1 …. 𝑏𝑗 =1 …… 𝑏𝑛 =1 Total
5
Let us prepare mathematical equations (model) from the above table.
Objective of the assignment problem is to optimize the total cost of assignment
thus the objective function can be stated as under.
𝑍 = ∑𝑛𝑖=1 ∑𝑛𝑗=1 𝐶𝑖𝑗 . 𝑥𝑖𝑗 --------------- (1)
Which is to be optimized subject to the following constraints.
𝑥11 + 𝑥12 + ⋯ … . +𝑥1𝑗 + ⋯ + 𝑥1𝑛 = 𝑎1 =1
𝑥21 + 𝑥22 + ⋯ … . +𝑥2𝑗 + ⋯ + 𝑥2𝑛 = 𝑎2 =1
….. ….. …... ….. ….. …..
𝑥𝑖1 + 𝑥𝑖2 + ⋯ … . +𝑥𝑖𝑗 + ⋯ + 𝑥𝑖𝑛 = 𝑎𝑖 = 1
….. ….. …... ….. ….. …..
𝑥𝑛1 + 𝑥𝑛2 + ⋯ … . +𝑥𝑛𝑗 + ⋯ + 𝑥𝑛𝑛 = 𝑎𝑛 =1
These n constraints can be represented by the single equation as under.
i.e. ∑𝑛𝑗=1 𝑥𝑖𝑗 = 𝑎𝑖 Where, i=1,2….n -------------(2)
Similarly, n constraints one for each column can be stated as under.
𝑥11 + 𝑥21 + ⋯ … . +𝑥𝑖1 + ⋯ + 𝑥𝑛1 = 𝑏1 =1
𝑥12 + 𝑥22 + ⋯ … . +𝑥𝑖2 + ⋯ + 𝑥𝑛2 = 𝑏2 =1
….. ….. …... …..
𝑥1𝑗 + 𝑥2𝑗 + ⋯ … . +𝑥𝑖𝑗 + ⋯ + 𝑥𝑛𝑗 = 𝑏𝑗 =1
….. ….. …... …..
𝑥1𝑛 + 𝑥2𝑛 + ⋯ … . +𝑥𝑖𝑛 + ⋯ +𝑥𝑛𝑛 = 𝑏𝑛 =1
This set on n constraints can also be represented by a single equation as under.
∑𝑛𝑖=1 𝑥𝑖𝑗 = 𝑏𝑗 where, j=1,2….n. _--------------- (3)
Also,
∑𝑚 𝑛
𝑖=1 𝑎𝑖 = ∑𝑗=1 𝑏𝑗 = n ---------------(4)
