Initial Experimental Results of Applying Continuity-guided Regeneration to Reactive Replanning and Rescheduling

We describe a reactive replanning technique developed to enable transportation planning in a highly dynamic, complex and critical domain: the military aeromedical evacuation of patients to medical treatment facilities. In early 1993, the Department of Defense tasked the U.S. Transportation Command (USTRANSCOM) to consolidate the command and control of medical regulation (assigning patients to medical treatment facilities) and aeromedical evacuation operations during peace, war and projected contingencies. The ensuing analysis led to TRAC2ES (TRANSCOM Regulating and Command and Control Evacuation System), a decision support system under development for USTRANSCOM (Johnson, 1994). Probably the most challenging aspect in planning and scheduling medical evacuation operations has to do with the dynamics of a domain in which requirements and constraints continuously change over time. Reactive replanning is a key capability of TRAC2ES. The key concern in reactive replanning is plan continuity -the new plan should take into account the current plan and should not introduce unnecessary disruptions. In this paper, we discuss this aspect of reactive replanning in depth, introduce an approach to reactive replanning called Continuity-Guided Regeneration (CGR), and describe the initial experimental results of using this approach. This paper appeared in Proceedings of the CONSTRAINTS-96 International Workshop on Constraint-Based Reasoning, Key West, FL, May 1996. This text may contain errors; please contact authors for an accurate copy.