Learning Optimal Policies in Potential Mean Field Games: Smoothed Policy Iteration Algorithms.
Qing TangJiahao SongPublished in: SIAM J. Control. Optim. (2024)
Keyphrases
- policy iteration
- optimal policy
- markov decision processes
- reinforcement learning
- model free
- fixed point
- learning algorithm
- policy evaluation
- average reward
- infinite horizon
- sample path
- finite state
- markov decision process
- convergence rate
- average reward reinforcement learning
- finite horizon
- least squares
- markov decision problems
- state space
- optimal control
- dynamic programming
- partially observable
- partially observable markov decision processes
- temporal difference
- markov random field
- reward function
- supervised learning
- control policies
- sufficient conditions