context: see the ART Testbed official Web site
due date: beginning of the last class
project: this counts for the remaining 50% of your final mark! Program an agent according to the ART testbed rules (see references below).
- this is a group project: teams of 3-4 students
- a competition will be held between the teams, while the students comment on the behavior of their players. We will throw in a few other agents (some always honest, some always dishonest, some programmed by Parthy). This will take place on the due date.
deliverables: for each team
- commented source code
- a short paper (max 10 pages single spaced) + a few slides to describe the approach, the design and models documentation and a short user manual (to compile and run the agent). Be specific - trust metric representation, trust metric calculation, reasons for your choice of algorithm and references.
evaluation: the following criteria will be judged in decreasing order of importance:
- Approach (is it any interesting?). You can explore one or more of the following things:
Make every possible attempt at separating your strategy or representation from the rest of the code. One should be able to edit a config file to modify the strategy without touching the rest of the code.
- Quality (will I be able to reuse it?)
- of the presentation
- of the documentation
- of the code (no bugs or crashes)
- code reusability
- of the design
- Performance (is your team any good?). Please note that this is the least important part of your mark.
resources
- IMPORTANT: questions about project descriptions and deliverables should be addressed to Babak. Questions about the ART Testbed should be addressed to Parth Chandrasekaran.
- The ART website is of course the first place you want to visit [2]. There are also agents from previous competitions. These can act as good benchmarks for your agents.
- Parthy's ART page [1]. The sequence diagrams in Parthy's slides give you an overview of the different protocols and relevant functions that you need to implement.
- Here are some different trust calculation methods:
(last edited November 3, 2009)
Find Page by browsing or searching
TrustProject