3rd + 10th Aug'2024 Lecture
3rd + 10th Aug'2024 Lecture
5
Problem Statement (2)
The following table shows the coordinates of the candidate warehouse sites and the fixed cost of
building the warehouse in millions of $. The logistics cost per mile per unit is 5 $.
Warehouse coordinates fixed cost in mn $ Supply cap
Warehouse 1 (0,0) 3 2000
Warehouse 2 (0,1) 2 3000
Warehouse 3 (0,2) 3 2000
Warehouse 4 (1,0) 1 1000
Warehouse 5 (1,1) 3 4000
Warehouse 6 (1,2) 3 2000
Warehouse 7 (2,0) 4 3000
Warehouse 8 (2,1) 3 2000
Warehouse 9 (2,2) 2 1000
a) Which warehouses will you establish and what will be your distribution strategy to supermarkets for the overall minimum cost?
b) If a warehouse is built at location 2, it cannot be built at location 5. What will be your strategy?
c) If a warehouse is built at location 2 and location 5, it cannot be built at location 8. What will be your strategy?
d) If a warehouse is built at location 2 and location 5, it must be built at location 7. What will be your strategy?
e) The warehouse 2 can be constructed only if warehouse 1 is constructed. What will be your strategy?
f) If a ware house is built at location 1, then it must be built at location 6 and location 9
6
Formulation for part a
Let the 9 candidate warehouse locations be denoted by i=1 to 9
Decision Variables: Let xij be the amount of units delivered from warehouse i to supermarket j (18 of them
Let yi be the binary decision variable which is 1 if the warehouse is established at location i, else 0 (9 of them
Parameters: If fi is the annual fixed cost for ith warehouse, and the delivery cost per unit per mile is c, the
distance between ith warehouse and jth supermarket is d ij, demand of jth supermarket us Demand j, and
Min total cost = delivery cost + annual fixed cost = sum of x ij.dij.c (for i=1,9 and j=1 to 2) + sum of fi.yi (for i=1
to 9)
Sum of xij (for j=1 to 2) <= yi*Supply_capi for for each warehouse i= 1 to 9
7
All xij >=0, and all yi are binary variables
Part b
Y2 + y5 <=1
Part c
If y2+y5=2 , y7 must be = 1
If y2+y5=0 or 1, y7 can be either 0 or 1
y7 >= y2 + y5 -1
Part e
y2<= y1
Part f
Solution:
Let there be m sources denoted by i=1 to m, and n destinations denoted by j=1 to n
Let xij be the amount of material transported from ith source to jth destination
Let aij be the variable cost of transport per unit quantity from ith source to jth destination
Objective is to minimise the total transportation cost = summation of xij.aij
St. for each of the sources i=1 to m, summation of xij for j = 1 to n <= Cap i
for each of the destinations j= 1 to n, summation of xij for i=1 to m >= Demand j
All xij >= 0
Transportation Problem b
Given the following cost matrix of transportation which captures per unit cost of transportation from 5
sources to 5 destinations, what would be the transportation schedule if the demand of all destinations
needs to be met at minimum cost? Assume a fixed cost of INR 5000 for each route
D1 D2 D3 D4 D5 Capacity
S1 10 12 8 13 6 150
S2 20 17 19 22 15 140
S3 16 13 18 14 10 100
S4 21 19 23 17 25 80
S5 12 20 22 15 18 130
Demand 110 140 70 130 150
Solution:
Let there be m sources denoted by i=1 to m, and n destinations denoted by j=1 to n
Let xij be the amount of material transported from ith source to jth destination
Let aij be the variable cost of transport per unit quantity from ith source to jth destination
Let yij be the binary variable whether I am using the route from ith source to jth destination
Objective is to minimise the total transportation cost = summation of xij.aij + yij.fij
St. for each of the sources i=1 to m, summation of xij for j = 1 to n <= Cap i
for each of the destinations j= 1 to n, summation of xij for i=1 to m >= Demand j
For all values of i and j, xij <= M * yij
All xij >= 0