Policy Synthesis and Reinforcement Learning for Discounted LTL.
Rajeev AlurOsbert BastaniKishor JothimuruganMateo PerezFabio SomenziAshutosh TrivediPublished in: CAV (1) (2023)
Keyphrases
- optimal policy
- reinforcement learning
- markov decision process
- markov decision processes
- infinite horizon
- average reward
- finite horizon
- total reward
- state space
- decision problems
- dynamic programming
- state and action spaces
- reinforcement learning algorithms
- reward function
- long run
- finite state
- action selection
- policy search
- temporal logic
- policy iteration
- state dependent
- markov decision problems
- average cost
- partially observable
- model free
- model checking
- optimal control
- function approximation
- control policies
- linear temporal logic
- reinforcement learning problems
- partially observable environments
- discounted reward
- initial state
- actor critic
- deterministic automata
- function approximators
- stationary policies
- sufficient conditions
- action space
- temporal difference
- continuous state spaces
- transition probabilities
- control policy
- learning algorithm