This paper presents our experiment in applying agent-based modeling to an evaluation scenario. (By “our” I mean myself, Van Parunak, Kirk Knestis, and Melanie Hwalek.) It is based on a real evaluation, although we had to add quite a bit to flesh out the details needed to build the executable model. It’s a small-scale exercise that we hope will provide a sense of how agent-based modeling differs from equation-based modeling, and why the agent-based approach provides unique knowledge that would otherwise be unavailable.
Evaluation Uncertainty 