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