Service for Solving Linear Programming Problems

and other interesting typical problems
Ðóññêèé

Example ¹4. Transportation Problem by North-West Corner Method (Balanced problem)

This solution was made using the calculator presented on the site.
Problem:
The cost of delivery of a unit of production from the supplier to the consumer is located in the lower right corner of the cell.
Supplier Consumer   Supply  
B 1 B 2 B 3
A 1
5
3
1
  10  
A 2
3
2
4
  20  
A 3
4
1
2
  30  
  Customer  
needs
15 20 25
It is necessary to find a transport plan in which the total cost of delivery will be the lowest.
Solution:
This is a necessary condition for solving the problem:
the total supply of suppliers should be equal to the total needs of consumers.

Let's check it.
The total supply of suppliers: 10 + 20 + 30 = 60 units.
The total needs of consumers: 15 + 20 + 25 = 60 units.
The total supply of suppliers equals the total needs of consumers.
This is a necessary condition for solving the problem:
number of used routes = number of suppliers + number of consumers - 1.

Therefore, if we have a situation where it is necessary to exclude a column and a row at the same time, we will exclude one thing.
We will start filling the table from the upper left corner and gradually move to the lower right corner.
From the North-West to South-East.
Supplier Consumer   Supply  
B 1 B 2 B 3
A 1
?
5
3
1
  10  
A 2
3
2
4
  20  
A 3
4
1
2
  30  
  Customer  
needs
15 20 25
10 = min { 15, 10 }
Supplier Consumer   Supply  
B 1 B 2 B 3
A 1
10
5
3
1
  10   No  
A 2
?
3
2
4
  20  
A 3
4
1
2
  30  
  Customer  
needs
15
5
20 25
5 = min { 5, 20 }
Supplier Consumer   Supply  
B 1 B 2 B 3
A 1
10
5
3
1
  10   No  
A 2
5
3
?
2
4
  20   15  
A 3
4
1
2
  30  
  Customer  
needs
15
5
No
20 25
15 = min { 20, 15 }
Supplier Consumer   Supply  
B 1 B 2 B 3
A 1
10
5
3
1
  10   No  
A 2
5
3
15
2
4
  20   15   No  
A 3
4
?
1
2
  30  
  Customer  
needs
15
5
No
20
5
25
5 = min { 5, 30 }
Supplier Consumer   Supply  
B 1 B 2 B 3
A 1
10
5
3
1
  10   No  
A 2
5
3
15
2
4
  20   15   No  
A 3
4
5
1
?
2
  30   25  
  Customer  
needs
15
5
No
20
5
No
25
25 = min { 25, 25 }
Supplier Consumer   Supply  
B 1 B 2 B 3
A 1
10
5
3
1
  10   No  
A 2
5
3
15
2
4
  20   15   No  
A 3
4
5
1
25
2
  30   25   No  
  Customer  
needs
15
5
No
20
5
No
25
No
Let's calculate the total cost of delivery for the initial solution.

10*5 + 5*3 + 15*2 + 5*1 + 25*2 = 150

Is the initial solution optimal?
Let's check this using the MODI method (UV method).
To each supplier A i we associate a some number U i
To each consumer B j we associate a some number V j
For the routes used: U + V = cost of delivery.
Let's find U i and V j step by step.
To do this, we need to enter the value of one of them. Let u2 = 0.
A2B1 :   v1 + u2 = 3     v1 = 3 - 0 = 3
A2B2 :   v2 + u2 = 2     v2 = 2 - 0 = 2
A3B2 :   v2 + u3 = 1     u3 = 1 - 2 = -1
A3B3 :   v3 + u3 = 2     v3 = 2 - (-1) = 3
A1B1 :   v1 + u1 = 5     u1 = 5 - 3 = 2
  Supplier   Consumer   U  
B 1 B 2 B 3
A 1
10
5
3
1
  u1 = 2  
A 2
5
3
15
2
4
  u2 = 0  
A 3
4
5
1
25
2
  u3 = -1  
  V   v1 = 3 v2 = 2 v3 = 3
Let's find evaluations of unused routes (cij - cost of delivery). ?
A1B2 :   Δ12 = c12 - ( u1 + v2 ) = 3 - ( 2 + 2 ) = -1
A1B3 :   Δ13 = c13 - ( u1 + v3 ) = 1 - ( 2 + 3 ) = -4
A2B3 :   Δ23 = c23 - ( u2 + v3 ) = 4 - ( 0 + 3 ) = 1
A3B1 :   Δ31 = c31 - ( u3 + v1 ) = 4 - ( -1 + 3 ) = 2
There are negative evaluations. Therefore it is possible to reduce the total cost of delivery.
STEP ¹1.
The cell A1B3 (unused route) is selected ? because its evaluation is negative.
Please, move the mouse cursor to the selected cell A1B3. Only horizontal and vertical cursor movements can be used.
Connect the filled cells with a continuous line so that you return to the cell A1B3.
The cells located at the vertices of the plotted line form a loop for the selected cell (see highlighted cells in the table below).
There is only one loop for cell A1B3.
Supplier Consumer   Supply  
B 1 B 2 B 3
A 1
10
5
3
-4
1
  10  
A 2
5
3
15
2
4
  20  
A 3
4
5
1
25
2
  30  
  Customer  
needs  
  15     20     25  
10 = min { 10, 15, 25 } ?
Supplier Consumer   Supply  
B 1 B 2 B 3
A 1
10
5
3
-4
1
  10  
A 2
5
3
15
2
4
  20  
A 3
4
5
1
25
2
  30  
  Customer  
needs  
  15     20     25  
This transformation will not change the balance.
But this transformation will change the total cost of delivery by:
1 * 10 - 5 * 10 + 3 * 10 - 2 * 10 + 1 * 10 - 2 * 10 = ( 1 - 5 + 3 - 2 + 1 - 2 ) * 10 = -4 * 10
You correctly noted that   -4 * 10 = Δ13 * 10 ?
Supplier Consumer   Supply  
B 1 B 2 B 3
A 1
10 - 10
5
3
+10
-4
1
  10  
A 2
5 + 10
3
15 - 10
2
4
  20  
A 3
4
5 + 10
1
25 - 10
2
  30  
  Customer  
needs  
  15     20     25  
We have a new solution. ?
Supplier Consumer   Supply  
B 1 B 2 B 3
A 1
5
3
10
1
  10  
A 2
15
3
5
2
4
  20  
A 3
4
15
1
15
2
  30  
  Customer  
needs  
  15     20     25  
We will calculate the total cost of delivery of the new solution.

S = 150 + Δ13 * 10 = 150 -4 * 10 = 110

Is the new solution optimal?
Let's check this using the MODI method (UV method).
To each supplier A i we associate a some number U i
To each consumer B j we associate a some number V j
For the routes used: U + V = cost of delivery.
Let's find U i and V j step by step.
To do this, we need to enter the value of one of them. Let u2 = 0.
A2B1 :   v1 + u2 = 3     v1 = 3 - 0 = 3
A2B2 :   v2 + u2 = 2     v2 = 2 - 0 = 2
A3B2 :   v2 + u3 = 1     u3 = 1 - 2 = -1
A3B3 :   v3 + u3 = 2     v3 = 2 - (-1) = 3
A1B3 :   v3 + u1 = 1     u1 = 1 - 3 = -2
  Supplier   Consumer   U  
B 1 B 2 B 3
A 1
5
3
10
1
  u1 = -2  
A 2
15
3
5
2
4
  u2 = 0  
A 3
4
15
1
15
2
  u3 = -1  
  V   v1 = 3 v2 = 2 v3 = 3
Let's find evaluations of unused routes (cij - cost of delivery). ?
A1B1 :   Δ11 = c11 - ( u1 + v1 ) = 5 - ( -2 + 3 ) = 4
A1B2 :   Δ12 = c12 - ( u1 + v2 ) = 3 - ( -2 + 2 ) = 3
A2B3 :   Δ23 = c23 - ( u2 + v3 ) = 4 - ( 0 + 3 ) = 1
A3B1 :   Δ31 = c31 - ( u3 + v1 ) = 4 - ( -1 + 3 ) = 2
There are no negative evaluations. Therefore it is not possible to reduce the total cost of delivery.
Result:

X opt =

Çíàê ñèñòåìû 0 0 10Çíàê ñèñòåìû
15 5 0
0 15 15

Smin = 110







© 2010-2024

If you have any comments, please write to matematika1974@yandex.ru