Graphical Agent Development Environment (GRADE II)

Increasingly sophisticated models and simulations developed by DoD to support a range of analysis and training functions have triggered a strong demand for realistic representation of the key human decision makers that populate the battelspace. Developinghigh-fidelity and user-friendly human behavior representations (HBRs) has not been a trival exercise however, especially given the complexity of today's simulation environments. To address this need, we propose to develop a full-scope prototype HBR basedon SAMPLE, our agent-based cognitive architecture, which has already overcome the significant hurdle of being integrated in a high-fidelity air combat simulation. We propose a three-phase effort: 1) Improve the representational fidelity of SAMPLE byproviding for an explicit model of multi-tasking and a principled means of accounting for individual differences and cultural factors; 2) Enhance the usability of SAMPLE via the continued development of a Graphical Agent Development Environment (GRADE) tosupport improved SAMPLE development, verification and visualization; and 3) Integrate the SAMPLE/GRADE/HBR Toolkit into a working USAF air combat simulation environment for validation and demonstration of overall utility in a Test and Evaluation (T&E)exercise.