Stochastic Shortest Paths and Weight-Bounded Properties in Markov Decision Processes.
Christel BaierNathalie BertrandClemens DubslaffDaniel GburekOcan SankurPublished in: CoRR (2018)
Keyphrases
- markov decision processes
- shortest path
- finite state
- state space
- optimal policy
- transition matrices
- shortest path algorithm
- reinforcement learning
- policy iteration
- decision theoretic planning
- reachability analysis
- edge weights
- model based reinforcement learning
- dynamic programming
- strongly connected components
- average reward
- road network
- average cost
- infinite horizon
- factored mdps
- action sets
- planning under uncertainty
- continuous state spaces
- real time dynamic programming
- finding the shortest path
- flow graph
- action space
- finite horizon
- markov decision process
- path length
- travel time
- discounted reward
- geodesic distance
- monte carlo