You should implement an agent-based method of your choice and test it in a domain of your choice. You should design rigorous experiments to evaluate your implementation, to show how it improves performance compared to some other method, as well as to gain insight about things like scalability, sensitivity to parameters, variations of a protocol, etc. You should write-up your project in a conference-paper-like format (probably on the order of 10 pages), including an interpretation of your results. In your introduction, please explain why you chose your particular project (i.e. what is the problem-solving/decision-making challenge in this domain?), justify the relevance of your work to an intelligent agents class (i.e. how is it an agent-based method?), and describe its relationship to existing methods or theoretical analyses in the literature. Also, please attach a printout of your source code.