# Overview
Code written using Tensorflow 2.0, JAX, FLAX and Python 3.0.

# File Structure
* [sddp_agent_main.py] (Main SDDP trainer file.)
* [sddp_agent.py] (SDDP algorithm.)
* [sddp_env.py] (Create application-specific problem formulation for SDDP.)
* [canonical_lp_converters.py] (Util functions for converting two case studies -- inventory and portfolio optimization to standard SDDP formulation.)
* [time_series_generator.py] (Util functions for time-series and scenario generation.)
* [stock_preprocess.ipynb] (Stock preprocessing colab.)
* [./piecewise_nn] (Neural network design dir.)
  - [cond_piecewise_nn.py] (Neural network that generates piecewise linear functions)
  - [ot_metrics.py] (Optimal transport objective for learning)
  - [main_nn.py] (Main demonstration of the usage of the neural networks)
