Creating and Capturing Expertise in Mixed-Initiative Planning

Abstract : Compared with prior engagements, commanders today are exposed to a battlespace that is more dynamic and less predictable. With increasing frequency, commanders are confronted with an array of problems whose solution requires knowledge beyond their military training. In these novel situations, decision makers often rely on their past experiences, incorporating a process best described in research as analogy-based reasoning and/or recognition-primed decision making. While the relevancy of the experience is based on the individual, a key goal would be to capture and exchange relevant experiences between individual decision makers. The shocking events of 11 September 2001 may have been less shocking to anyone serving in the Pacific theater of operations toward the end of World War II and experienced Kamikaze warfare. This paper describes work in progress at the USAF Research Laboratory Information Directorate to capture, develop, and provide experience to commanders during mixed-initiative planning. The objective of this work is to provide a rich database of experiences for the commander to compare to the current situation. The research described in this paper is aimed at developing a computational representation for episodic models, and reasoning on those models for retrieval and experience extraction.