Representing and Solving Air Campaign Problems as Partially Observable Markov Decision Problems
Abstract
The original purpose of this project was to design algorithms and architectures for maintenance and deployment scheduling solutions to support large-scale strategic military airlift activities related to the needs of the Air Mobility Command (AMC), and secondarily to adapt these solutions to other military and civilian planning and scheduling problems. The research adopted a stochastic modeling framework and used novel techniques for planning in unpredictable dynamic environments with complex state and action spaces. Many of the original goals of the project were achieved in conjunction with the primary contract; however, several extensions were carried out by students receiving funding from the AASERT supplementary grant.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 30, 2001
- Accession Number
- ADA395052
Entities
People
- Thomas L. Dean
Organizations
- Brown University