Reinforcement Learning as a Rehearsal for Planning in Air Battle Management
Abstract
Conduct fundamental research to advance Reinforcement Learning as a Rehearsal (RLaR) to Stratagem wargame, and disseminate the findings through publications in peer-reviewed venues, in collaboration/consultation with AFRL project team. Combine the results of research and a host of existing ideas to develop complete end-to-end deep learning RLaR based planners, one for each of the blue- and red-agents trained together as sparring partners, and deliver the code as well as the trained models to AFRL.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 11, 2021
- Source ID
- FA87502010105
Entities
People
- Bikramjit Banjerjee
Organizations
- Rome Laboratory
- United States Air Force
- University of Southern Mississippi