Mathematical Optimization: Models, Methods and Applications
Autor: rasp11ab • April 5, 2016 • Research Paper • 4,563 Words (19 Pages) • 874 Views
[pic 1]
Mathematical Optimization: Models, Methods and Applications Final Assignment 06-11-2015 Rasmus Cand.merc.mat 10 pages / 13.137 characters (including spaces) |
Part 1
General about part 1
The purpose with this part is to analyze a Single-Sourcing Problem (SSP). A Single-Sourcing Problem of course both has benefits and risks, but I will discuss that furthermore through the assignment. During the assignment I will try to discuss and comment on everything that I do. My code and the answers I receive from www.neos-server.org can be seen in my appendices.
(i)
In the first question in part 1, I am asked to solve the SSP using the data in Figure 1. We have 4 facilities and 30 customers. In Figure 1 the demand of each customer is also given, and of course I will have to satisfy this. Therefore this will become one of my constraints. It is also known that each facility has a capacity, and of course this will become a constraint as well. Because it is a SSP problem, we are also given the information that each customer has to be served by exactly one facility. When a facility delivers one unit to a customer it faces a cost. The purpose with the first question is to minimize the cost that the facility faces delivering the units. I will now show what the problem looks like:
[pic 2]
[pic 3]
[pic 4]
[pic 5]
[pic 6]
Now I have formulated the problem, and I will now use a Mixed Integer Linear Programming solver from www.neos-server.org. As I mentioned earlier my code and the whole answer from the website will be in my appendices.
With the constraints from the first question I get a minimum cost of: [pic 7]
(ii)
In the second question a new constraint has been given. It says that each facility must satisfy at least of the customers. Therefore our problem will now like this:[pic 8]
[pic 9]
[pic 10]
[pic 11]
[pic 12]
[pic 13]
[pic 14]
This is the new problem. As you can see we got a new constraint, which says that each facility must satisfy at least 6 customers. There are 30 customers, and therefore 6 is the 20%. With the new constraint I get a minimum total cost of: [pic 15]
As we can see, the minimum total cost is a little bit higher in question 2 than in question 1. If we take a look at the appendices in the part about 1.(i), then we can see on the screenshot from the website, that facility 4 only had 5 customers. That was the optimal solution in that case, so therefore if we add another constraint that says every facility must have at least 6 customers, then of course the minimum total cost will be higher.
...