Code to reproduce experiment results for "Distributional Reinforcement Learning
for Risk-Sensitive Policies"

Requirement: 
 - Python >= 3.7
 - tensorflow >= 2.1.0

(all other packages such as numpy etc should be as recent as possible,
 but older versions should work fine)


Synthetic experiments:

To produce plots in Figure 1,
 1. Run cvar_dynamic_synthetic.py and cvar_static_synthetic.py to train models
    and save results (in a ./res folder)
    (gpu and/or cpu parallerization can help speed up)

 2. Run synthetic_plots.py to generate the plots

Option trading experiments:

To produce plots in Figure 2 & 3,
 1. Run cvar_dynamic_option.py and cvar_static_option.py to train models
    and save results (in a ./res folder)
    (gpu and/or cpu parallerization can help speed up)

 2. Run option_plots.py to generate Figure 2
 3. Run option_eval_on_real.py to evaluate models and plot results (Figure 3)

