Representing Operational Planning Knowledge

The operational military planning problem has begun to receive increased attention from industry, academia, and government as a domain suitable for the application of artificial intelligence (AI) technologies. The current paper describes the operational planning problem in terms familiar to the AI planning/problem-solving community, introducing the major measures of complexity in this domain. To do so, the key elements of a general AI planning model are identified, and then they are related to corresponding elements in the operational planning problem. Following this, several key concepts relating to knowledge representation in the operational planning domain are introduced, and several extensions to a standard object-based representation are discussed. Lastly, an implementation of these representational concepts being conducted as part of the ARES Project at the US Army's Communications/Automatic Data Processing Center is discussed.