Implementation of two deep reinforcement learning algorithms from
Hedging using reinforcement learning: Contextual k-Armed Bandit versus Q-learning
Loris Cannelli, Giuseppe Nuti, Marzio Sala, Oleg Szehr
and
Dynamic Replication and Hedging: A Reinforcement Learning Approach
P. N. Kolm and G. Ritter
We will also provide an approach based on classical deep learning, wich does not rely on reinforcement lerning.
The code provided is written in Python 3.8, and relies on the following libraries: