Efficiently Breaking the Curse of Horizon: Double Reinforcement Learning in Infinite-Horizon Processes.
Nathan KallusMasatoshi UeharaPublished in: CoRR (2019)
Keyphrases
- infinite horizon
- optimal policy
- reinforcement learning
- optimal control
- markov decision processes
- finite horizon
- markov decision process
- partially observable
- dynamic programming
- state space
- long run
- stochastic demand
- production planning
- single item
- policy iteration
- decision problems
- average cost
- temporal difference
- function approximation
- lost sales
- machine learning
- reinforcement learning algorithms
- finite state
- multistage
- action space
- multi agent