CMSC 477/677: MAS Project
Your assignment is to implement some multi-agent system.
You will deliver the following things:
- Your well documented, well formatted source code. Also, include
a short document that tells how to compile/run/use your program.
- An 10 minute demonstration of your project.
In addition, graduate students must write up a report of experiments ran:
- A short document (2-3 pages of standard LaTeX format)
outlining any experiments or findings.
The effect of this 3rd requirement, is that undergrads may implement anything
related to MAS: from toy applications, full-blown simulations, experimental systems, etc.
Graduate students must select a topic in which some sort of result or finding
can be reached. That is, you are implementing an MAS system in order to prove something.
The first phase of the project is to select a general area or provide
a vague idea to the instructor. This way, teams of size 1-3 students may be selected so
students can work together on the same topic. If you would rather select someone to work
with explicitly, please do so.
Once teams have been selected, you will get together and formulate a short proposal
that outlines what you will be doing or the project. This proposal is subject to
the instructor's
modification or suggestions.
All deliverables other than the proposal are due the day of the final exam slot.
The project grade will be based on the following:
- Proposal (5%)
- Implementation (85%/60%)
- Presentation (10%)
- Experiments (0%/25%)
The topic can either be an original idea or a re-implementation of a previous experiment.
Some example projects or project ideas:
- Comparing auction types by running auction simulations
- Compare the efficiency of distributed election algorithms
- An extended version of the game-theory project that compares lots of strategies
or implements one (or a couple) interesting strategies (such as master/slave)
- Implement team-forming agents that work together to solve tasks
- Domain specific MAS, such as military operations, soccer teams, real-time strategy game AI,
etc.