All learning is local: Multi-agent learning in global reward games , (2003)

by Y-H Chang, T Ho, L Kaebling
Venue:Advances in Neural Information Processing Systems, MIT Press, ISBN .