Hitting time for Markov decision process

Abstract

We define the hitting time for a Markov decision process (MDP). We do not use the hitting time of the Markov process induced by the MDP because the induced chain may not have a stationary distribution. Even it has a stationary distribution, the stationary distribution may not coincide with the (normalized) occupancy measure of the MDP. We observe a relationship between the MDP and the PageRank. Using this observation, we construct an MP whose stationary distribution coincides with the normalized occupancy measure of the MDP and we define the hitting time of the MDP as the hitting time of the associated MP.

The first version of this paper pointed out some issues in the old version of "Cross-Domain Imitation Learning via Optimal Transport". The authors have then addressed these issues according to our suggestions in a new version. We therefore updated our paper, in which we removed contents related to these issues