Formal Multi-Objective Synthesis of Continuous-State MDPs.
Sofie HaesaertPetter NilssonSadegh SoudjaniPublished in: IEEE Control. Syst. Lett. (2021)
Keyphrases
- continuous state
- multi objective
- reinforcement learning
- continuous state and action spaces
- policy search
- action space
- markov decision processes
- continuous state spaces
- finite state
- evolutionary algorithm
- robot navigation
- partially observable markov decision processes
- state space
- control policies
- state dependent
- planning problems
- optimal policy
- objective function
- genetic algorithm
- continuous action
- function approximation
- real valued
- dynamic programming
- neural network
- hidden state
- model free
- reinforcement learning algorithms
- markov decision process
- temporal difference
- optimal control
- belief state
- average reward
- state action
- action selection
- markov decision problems
- learning algorithm
- heuristic search