Jeremy Baxter and Richard Hepplewhite
This paper describes our use of the SIM_AGENT toolkit for research into agents for battlefield simulations. We briefly describe our work and then try to classify the domain in which we are working with respect to the environment, agent’s goals and actions, the architecture of our agents and environmental representation. The facilities of the toolkit are outlined and discussed with respect to the ease with which they enabled us to apply the toolkit to our domain. We conclude by discussing some benchmarks we believe could help in categorising toolkits and describing some features we would like to see in future agent toolkits.