Final Project for Reinforcement Learning (ORCS 4529) at Columbia University
Authors: Davit Shadunts, Eliza Cohn, Neal Ma
Term: Fall 2024
To setup this repository with access to all data, you will need to pull the git submodules. To do this, run this command from the root of the directory:
git submodule update --init --recursive
Pre-trained agents can be accessed from this Google Drive link. The folder contains the following six .pkl files, representing different agent models and strategies:
| File Name | Model | Strategy |
|---|---|---|
meanshift_norm_conservative.pkl |
Mean-Shift | Conservative |
meanshift_norm_honest.pkl |
Mean-Shift | Honest |
meanshift_norm_risky.pkl |
Mean-Shift | Risky |
timevarying_norm_conservative.pkl |
Time-Varying | Conservative |
timevarying_norm_honest.pkl |
Time-Varying | Honest |
timevarying_norm_risky.pkl |
Time-Varying | Risky |
Download the pkl files into the /agents directory to use them in a simulation.