Market Making via Reinforcement Learning in China Commodity Market

Abstract

Market makers play an essential role in financial markets. A successful market maker should control inventory and adverse selection risks and provide liquidity to the market. As an important methodology in control problems, Reinforcement Learning enjoys the advantage of data-driven and less rigid assumptions, receiving great attention in the market-making field since 2018. However, although the China Commodity market has the biggest trading volume on agricultural products, nonferrous metals, and some other sectors, the study of applying RL to Market Making in China market is still rare. In this thesis, we try to fill the gap. Our contribution is threefold: We develop the Automatic Trading System and verify the feasibility of applying Reinforcement Learning in the China Commodity market. Also, we probe the agent's behavior by analyzing how it reacts to different environmental conditions.