Distributed Cooperative Multi-Agent Planning Over Information Space
Abstract
The proposed work aims at developing a new theory to address multi-agent planning over information space. The theory particularly extends the state-of-the-art of the potential game theory to effectively address non-separable rewards. Key goals of the proposed methodology including the following: (a) the developed algorithm should run in a polynomial time, (b) the proposed game-theoretic distribute planning would perform better than a centralized sequential greedy scheme, and (c) convergence and performance analysis is attached to the theory. In the long run, the theoretical framework in this work will be useful to better understand and design cooperative operations of multiple autonomous systems, interacting with dynamic uncertain environment.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Dec 05, 2016
- Source ID
- FA23861614064
Entities
People
- Han-Lim Choi
Organizations
- Air Force Office of Scientific Research
- KAIST
- United States Air Force