An Elementary Proof that Q-learning Converges Almost Surely.
Matthew T. RegehrAlex AyoubPublished in: CoRR (2021)
Keyphrases
- reinforcement learning
- cooperative
- function approximation
- learning algorithm
- multi agent
- theorem proving
- stochastic approximation
- optimal solution
- state space
- dynamic programming
- theorem prover
- model free
- temporal difference learning
- optimal policy
- potential field
- real time
- interactive theorem proving
- action selection
- learning rate
- artificial neural networks
- expert systems
- machine learning