1 paper
Nicolas Gast, Bruno Gaujal, Chen Yan
We propose a new policy, called the LP-update policy, to solve finite horizon weakly-coupled Markov decision processes. The latter can be seen as multi-constraint multi-action band…