Life Is Random, Time Is Not: Markov Decision Processes with Window Objectives.

Thomas Brihaye Florent Delgrange Youssouf Oualhadj Mickael Randour

Published in: CONCUR (2019)

Keyphrases

markov decision processes
optimal policy
finite state
state space
policy iteration
dynamic programming
transition matrices
reinforcement learning

planning under uncertainty
reinforcement learning algorithms
average cost
finite horizon
decision theoretic planning
markov decision process
risk sensitive
reachability analysis

decision processes
average reward
partially observable
infinite horizon
state and action spaces
action sets
discounted reward
semi markov decision processes

action space
decision diagrams
continuous state spaces
factored mdps
multiple objectives
model based reinforcement learning
real valued
stochastic shortest path