联系我们： 手动添加方式: 微信>添加朋友>企业微信联系人>13262280223 或者 QQ: 1483266981

MSIN0116: Decision Making and Analytics
Mario Vanhoucke
Individual coursework – 2023 (BiMBA)
Deadline: June 29, 2023
Upload your report in PDF to Moodle (No Word, Excel, or zipped files are allowed)
Assignment. Problem analysis (maximum 10 pages, Font Arial, Size 10)
Describe and analyse a (business) problem of reasonable size (minimum 20 decision variables) using the
methodologies discussed in the course module. Any problem is good. It can be a simplified version of a problem you
encountered in your business, or it can be a problem you invented or have seen somewhere (in a company, your free
time, internet, …) It doesn’t have to be a real one, but it should be realistic, make sense, and interesting. Just be
creative! You can use your own experience of the business, or you can just make up a problem yourself using your
own imagination and experience. As long as the problem is clear and challenging, and amendable for modelling, it’s
good!
The individual coursework consists of an assignment with 4 analyses. For each analysis, you should do the following:
Be concise and to-the-point, and include details about the assumptions and calculations, but also include
summary graphs with results.
Don’t forget to draw conclusion for each analysis (calculations and graphs without conclusions are not very
useful).
Indicate the advantages and disadvantages of each analysis from a management perspective.
Highlight possible improvements.
The use of the Solver tool (in MS Excel) is allowed, but you can use any other software tool if you want.
The written reports will be assessed on the following aspects:
Quality of the argument (clarity of problem description, clarity of assumptions, motivation of methodology used).
Demonstrating how calculations have been done (clear presentation of data, easiness to follow calculations,
clarity of the model).
Supporting arguments with data (interpretation of result, showing shortcomings, and drawing clear
conclusions).
Presentation (presenting results in a clear manner to management).
In the next paragraph, you find a summary of what each analysis should contain, but you should use this only as a
guideline. That means that you can skip some parts, and/or extend others. Beware that the written report must be
targeted to a company stakeholder, so it should be (i) crystal clear, (ii) easy and attractive to read, (iii) 100% correct,
and (iv) results oriented. As a decision maker, you should come up with not only a sound and correct analysis but
also with an ingenious and innovative solution. Focus on the problem and the solution, not only on the model. Try to
be original! Remember: you are a data scientist, so originality is in your nature!
Good luck!
Mario Vanhoucke
Analysis 1. Project description (maximum 1 page)
Give a short problem description (descriptive) and highlight the most important aspects (the nature of the problem,
the sector (healthcare, IT, construction, tourism, …), the client, the company, the challenge, …). Make this description
as attractive as possible, and include graphics whenever desirable.
Analysis 2. Mathematical Programming (maximum 4 pages)
Individual coursework MSIN0116 – Mario Vanhoucke
Analyse the problem and build (a) mathematical model(s) (either linear programming (LP), or integer programming
(IP))., which must include the following items:
Define the problem with data, and use assumptions to clearly define the scope of the problem.
Construct a LP/IP model and clearly show how you have defined the decision variables.
Solve the LP/IP model (under different assumptions, if necessary).
Analyse the solution(s) using sensitivity analysis (LP only) and draw conclusions for the management team.
Hint: Make sure that you clearly show that the problem could only be solved using the model, which means that –
without this model – people quickly come up with solutions that might be good (or not) but are certainly not optimal.
Convince people that the model leads to significant improvements for the company. Improvement the company
would otherwise not have found. You are a data scientist, don’t forget! Modelling is the art of doing better!
Analysis 3. Decision Tree Analysis (maximum 4 pages)
Extend the problem and model(s) with uncertainty in the project data, and carry out a decision tree analysis to find a
“robust solution”, including the following items:
Define uncertainty scenarios for some key parameters (data) of the problem.
Construct a decision tree and solve it.
Discuss the solution and perform a sensitivity analysis (on the probabilities).
(Optional) Discuss the use of risk aversion or risk seeking behaviour for your analysis, and discuss the
implications for the company.
Draw conclusions and highlight suggestions for improvements.
Hint: Make sure that the third analysis transforms the solution of the problem (found in analysis 2) a bit more
realistic by adding uncertainty in the data. You should show the company that the solution of analysis 2 is still a good
one, however, your proposed solutions (of analysis 3) might differ a little bit since the data is uncertain (and as you
know, data is always uncertain when it is used for the future decisions). Show the company people that adding
uncertainty might lead to better decisions.
Analysis 4. Executive summary (maximum 1 page)
Write a single page executive summary indicating the highlights of the previous analyses, and make sure it contains
the details that you want to communicate to your stakeholders.
Individual coursework MSIN0116 – Mario Vanhoucke