COMP 567
Case Study Information Winter 2012
Students will gain experience by working in teams on a substantial case
study of an application of integer programming. This study will involve
the solution of a realistically sized integer programming problem on a
reasonably
large set or sets of data. There are four
parts to the case study, the first three parts are oral, the fourth
written:
1. Initial description of problems,
2. Proposals by consultant teams,
3. Preliminary findings and progress report,
4. Final written case
study
report.
The initial presentations are by each
individual student. After
this
each student joins both a consultant's team of 3-4 people for
one of the
problems
presented, and a management
team for a different problem.
1. Initial
description of problems(5%): due
February 1 by emailing a
pdf file to Gary
Project
descriptions
Each student plays the role of a manager of a business that
needs help
with some kind of optimization problem. The problem is stated in
general terms using in writing using a maximum
of
2
pages. Indicate the data available, the
objective desired, and whatever relevant constraints are to be
considered. No mathematical
formulation is allowed. The problem should
be solvable by integer programming techniques. You must invent
this problem, not just copy it from a textbook or from the web.
You submit this as a pdf file
to Gary at grouma@cs.mcgill.ca
All descriptions will be posted on a web page. Students
will form consultant teams: 4 groups of 3 and 1 group of 4.
You may wish to use webCT
to set up teams.
Each team works on a different problem. Teams register with Conor,
first come first served.
Once the consultant teams are fixed we will assing management teams.
2. Proposals by Consultants (10%): February
13,15 Projects chosen, teams and schedule are
here
Each consultant team gives a proposal to solve all or part of the
problem they have chosen. The proposal should describe the following:
an overview of the problem, the specifications of the input data
required, the decision variables to be considered, the objective
function and the constraints to be handled. Specific deliverables
should be described. It is best to describe a small prototype that
should
be easily solvable, a reasonable sized problem that might take
considerable computer time (CPLEX), and possible extensions if time
permits. Teams of three have up to 15 minutes, and teams of four have
up
to 20 minutes. Each member of the
team must present part of the
proposal. The management team for the project must provide
constructive criticism and comments after the proposal. They will be
required to supply the relevant input data within a week.
3. Preliminary
findings and progress report (15%): April 2,4 (Note: new dates)
Each consultant team presents the preliminary results of their project
using the
same time and ground rules as the proposal. Then the management team
questions the consultants for a 10 minute period, and there is open
discussion. It is important to participate in these discussions.
Consultants must be able to demonstrate they have successfully solved
at least a prototype version of the original problem, and computational results must be presented.
If the slides are
self-explanatory, no written report is required.
4. Final Written Report (10%): April 27
Ground rules: Consultants may discuss
informally with their management team at any time, but should not
collaborate with other consultant teams.
5. Pointers for a good presentation
Formalities :
Introducing oneself, finishing in the alloted time
Clarity of Explanation:
Structured the presentation well
Conveyed main ideas well
Interest/Motivation :
Speaker was clearly interested & motivated
Audience was interested (asked questions?)
Mechanics:
Good use of overheads, blackboards, handouts (& other
visuals)
Established good contact with audience (voice level, eye
contact)
Pertinence:
Choice of details to present
Speaker related the topic to course material