PhiBE: A PDE-based Bellman Equation for Continuous Time Policy Evaluation.
Yuhua ZhuPublished in: CoRR (2024)
Keyphrases
- policy evaluation
- matrix inversion
- least squares
- monte carlo
- reinforcement learning
- markov chain
- temporal difference
- model free
- markov decision processes
- policy iteration
- variance reduction
- state space
- partial differential equations
- function approximation
- optimal control
- linear program
- dynamical systems
- optimal policy
- optical flow
- semi parametric
- state action
- statistical inference
- stochastic processes
- partially observable markov decision processes
- action selection
- infinite horizon
- transition probabilities
- gaussian process
- learning algorithm