This project, by Aadam Wiggers and Albin Jaldevik, is a PyTorch-based reproduction of the paper Deep Hedging by Hans Bühler, Lukas Gonon, Josef Teichmann, and Ben Wood. We'd like to extend our gratitude and give full credit for the original concept to these authors. The project is puerly for educational purposes and is not intended for commercial use.
We present a range of experiments demonstrating what can be achieved with our framework. For a detailed explanation and deeper insights, please refer to our report.
The following plot showcases the behavior of a trained agent:
We present the final Profit & Loss distribution for two distinct trading costs scenarios:
For a quick and easy start, you can use our Kaggle notebook. This allows you to run the code directly in your browser without any local setup.
To set up the project on your local machine, follow these steps:
-
Clone the repository and navigate to the root directory:
git clone https://github.com/appie-mathematics/Deep-Hedging cd Deep-Hedging -
Install the dependencies. To do so, run the following command:
conda env create -f environment.yml
Note: This requires Conda to be installed on your machine. The above command creates a conda environment named
deep-hedge.To activate the environment, use:
conda activate deep-hedge
-
Run the code by executing the following command:
python src/main.py
For any inquiries, feel free to open an issue on this GitHub repository or contact us directly.