An immediate-return reinforcement learning for the atypical Markov decision processes.
Zebang PanGuilin WenZhao TanShan YinXiaoyan HuPublished in: Frontiers Neurorobotics (2022)
Keyphrases
- markov decision processes
- reinforcement learning
- reinforcement learning algorithms
- state space
- optimal policy
- finite state
- policy iteration
- action space
- state and action spaces
- transition matrices
- partially observable
- decision processes
- model based reinforcement learning
- function approximation
- model free
- dynamic programming
- average cost
- decision theoretic planning
- reward function
- planning under uncertainty
- finite horizon
- markov decision process
- infinite horizon
- multi agent
- machine learning
- average reward
- state abstraction
- temporal difference
- reachability analysis
- learning algorithm
- approximate dynamic programming
- decentralized control
- stochastic games
- control problems
- markov decision problems
- policy evaluation
- discounted reward