Skip to content

Latest commit

 

History

History
9 lines (7 loc) · 632 Bytes

README.md

File metadata and controls

9 lines (7 loc) · 632 Bytes

Reinforcement-Learning-Project

Algorithmic trading using RL - Project @ IIT-M

The project evaluates the efficiency of deep-q-networks in the task of automated trading. A MDP formulation of the trading task is proposed and the optimal action function is learnt using a neural network. Transaction costs and liquidity constraints are taken into account. The agents are trained and evaluated on real-life cryptocurrency instruments like Bitcoin and Ethereum. The results show that the proposed method does well on learning short term trends and are able to profit well though the interpretability of the strategies is in question.