Agile Control of Military Operations (JFACC)

Abstract : This research focused on the problem of providing military commanders with real-time, optimal control of military air-to-ground operations for a 24-hour segment of a Joint Air Operations (JAO) campaign. In particular, we focused on developing control algorithms that anticipate possible air-to-ground mission modifications due to uncertain future events, thereby generating missions that can be readily adapted in the presence of contingencies. The primary benefit of this technology is agile and stable control of distributed and dynamic military operations conducted in inherently uncertain, hostile, and rapidly changing environments. The control methodology developed combines Approximate Dynamic Programming (ADP) and statistical hybrid state modeling techniques. Accordingly, a novel hybrid, multi-rate control architecture that tailors the control strategy for different battlespace situations was developed. For this JAO problem, a key concern was the scalability of the control methodology to larger scenarios. As a result, we investigated a broad spectrum of ADP control strategies. The solution quality and computation performance of these algorithms was tested and verified in a JAO simulator. It was shown through experimentation that the ADP strategies were able to produce operationally consistent control strategies that anticipated likely contingencies and positioned assets for opportunities of recourse in near real-time.