This repository contains the code for the paper:
Optimal-PhiBE: A PDE-based Model-free framework for Continuous-time Reinforcement Learning
arXiv:2506.05208
The following Jupyter notebooks reproduce the figures from the paper:
- Figure 5 (A–D):
comparison_1D_deterministic.ipynb - Figure 6 (A–D):
comparison_1D_stochastic.ipynb - Figure 7 (A–D):
comparison_2D_deterministic.ipynb - Figure 8 (A–D):
comparison_2D_stochastic.ipynb - Figure 9 (A–B):
dt_graph_real_data.ipynb - Figure 9 (C):
batch_size_graph.ipynb - Figure 10 (A–C):
comparison-merton.ipynb