Prototype Architecture for an Automated Scenario Generation System for CombatSimulations

摘要

This research effort examines the problem of automating the scenario generationprocess and presents a prototype architecture for an automated scenario generation system. This architecture is designed using an object-oriented approach which leads to a modular and modifiable design. The architecture provides a mechanism for automatically generating scenario files from a textual US Army operation order. This translation process occurs in two phases. First, the text operation order is translated into an intermediate format called the mission file. In the second phase, the system reads the mission file, instantiates intelligent entities, and assigns missions to those entities. The intelligent entities emulate the subordinate leaders in planning missions. The thesis effort assumes that sufficiently sophisticated natural language processing algorithms can translate the operation order into the mission file. A prototype scenario generation system was implemented using NASA's CLIPS expert system development tool. The prototype system successfully translates mission files into scenario files for the BATTLESIM combat simulation model developed at the Air Force Institute of Technology. Automated scenario generation, CLIPS, Mission planning, Combat simulations, War-games, COOL, Artificial intelligence, Expert systems.