Market Making with Deep Reinforcement Learning
Aidan Wong Weng Seng Supervisor: Dr. Huang Zhiyi FYP ID: fyp25009
Introduction
Market making is a critical function in financial markets, providing liquidity and facilitating
smoother trading by continuously buying and selling securities. Market makers earn profits
through high-frequency trading and by capturing the spread in the securities market in
which they operate. With the rapid advancement of technology, particularly in artificial
intelligence, deep reinforcement learning has emerged as a transformative tool in this
domain. Unlike traditional market making, which is rule-based on predefined market
conditions that are often criticized for containing strong, naïve assumptions, this research
aims to tackle these limitations by exploring the intersection of deep reinforcement
learning and market making. The goal is to enhance trading strategies, improve quoting
accuracy, and optimize inventory risk management.
Schedule
| Phase | Weeks | Key Milestones |
| Literature Review | 1~4 | Develop Theoretical Framework |
| LOB Simulation | 5~12 | Functional Environment with trainable model |
| Agent Development | 13~24 | Effective Q-function and agent decisions |
| Model Evaluation | 25~29 | Traditional agents and Benchmark analysis |
| Conclusion | 30~32 | Final Year Report |