Optimal and dynamic planning for Markov decision processes with co-safe LTL specifications.
Bruno LacerdaDavid ParkerNick HawesPublished in: IROS (2014)
Keyphrases
- markov decision processes
- dynamic programming
- average cost
- planning under uncertainty
- average reward
- decision theoretic planning
- finite horizon
- macro actions
- partially observable
- action sets
- optimal policy
- reinforcement learning
- state space
- policy iteration
- infinite horizon
- finite state
- discounted reward
- reachability analysis
- bounded model checking
- stationary policies
- total reward
- action space
- transition matrices
- partially observable markov decision processes
- expected reward
- state and action spaces
- probabilistic planning
- planning problems
- reinforcement learning algorithms
- model based reinforcement learning
- optimal control
- dynamic environments
- markov decision problems
- model checking
- optimality criterion
- linear program
- markov decision process
- optimal solution
- least squares
- multistage
- heuristic search
- machine learning
- linear temporal logic
- continuous state spaces
- temporal logic
- optimal planning
- decision theoretic
- long run
- classical planning
- initial state