Optimal Policy Generation for Partially Satisfiable Co-Safe LTL Specifications.
Bruno LacerdaDavid ParkerNick HawesPublished in: IJCAI (2015)
Keyphrases
- optimal policy
- bounded model checking
- markov decision processes
- finite horizon
- state space
- decision problems
- reinforcement learning
- dynamic programming
- infinite horizon
- long run
- state dependent
- finite state
- multistage
- sufficient conditions
- markov decision process
- bayesian reinforcement learning
- temporal logic
- model checking
- control policies
- linear temporal logic
- lost sales
- average reward
- initial state
- partially observable markov decision processes
- machine learning
- develop a mathematical model
- markov decision problems
- average cost
- sat problem
- reward function