Supportive Tactical Aide Foresight Framework (STAFF)

The Military Decision-Making Process (MDMP) can be difficult and will greatly benefit from a machine aide to help staff officers. Operations for today’s Army can be complex, widespread, and proceed at high OPTEMPO. However, the MDMP is still largely centralized and manual. Decentralized interactions are hampered by high volumes of data and by difficult communications. Units can use help to remain synchronized and coordinated, even as they are overwhelmed by the large volume of intelligence that may or may not be relevant to each mission. During the combat planning process, Brigades often leave Battalions limited time to plan before the official OPORD is published before execution. Depending on the complexity of the mission and the experience of the staff, the full MDMP, especially Course of Action (CoA) generation, is shortened at the lowest levels. The CoA development process that enable an effective wargaming analysis, are of often short-circuited due to the time limitation. Traditional software tools for this type of analysis are not available because the bottlenecks are mostly cognitive.Artificial Intelligence (AI) and Machine Learning (ML) tools can mitigate some of the limitations. The most recent AI/ML technologies, with their quick information processing, swift communications, and cognitive capacities, can operate well beyond the capabilities of traditional software aides. These machine agents hold the promise of decentralizing and automating the planning process. However, existing Large Language Models (LLMs) are inadequate for CoA generation, because they do not incorporate the proper knowledge of military Tactics, Techniques and Procedures (TTP) and Standard Operating Procedures (SOP). Also, LLMs are text processors and require augmentation to understand other types of intel, especially spatial and temporal data, such as the Modified Combined Obstacle Overlay (MCOO). The proposed effort will resolve these shortcomings