TY  - RPRT
TI  - A Sliding-Window Algorithm for Markov Decision Processes with Arbitrarily Changing Rewards and Transitions
AU  - Pratik Gajane
AU  - Ronald Ortner
AU  - Peter Auer
PY  - 2018
UR  - https://arxiv.org/abs/1805.10066
ID  - 1805.10066
ER  -