Reinforcement Learning for Portfolio Management

1. Learns the optimal action, rather than models the market.
2. Adaptive to temporary changes of the market, due to its online training.
3. Optimizes the long-term (cumulative) reward, rather than the instantaneous benefit.

Exclusively Python 3 compatible, because of typings

• source scripts/setup.sh

• Interim Report: Introduction to qtrader
• Papers Review: Motivation, pros & cons of existing methods
• Resources: List of relevant resources
• Final Report: Master's Thesis
• Presentation: 15 minutes project presentation