STRONG MARKET-MAKING USING DEEP REINFORCEMENT LEARNING: BITCOIN MARKET ANALYSIS

Abstract

This study evaluates the use of deep reinforcement learning (DRL) in market-making, specifically in the Bitcoin market. DRL has shown promise in providing robust market-making capabilities, including enhanced market liquidity and risk management, which may lead to more efficient price discovery and lower volatility. The study also discusses the historical perspective of market-making techniques and explains how agents can use DL algorithms and RL principles to improve preset objectives in financial markets. It also reviews essential DRL algorithms like Proximal Policy Optimization (PPO), Advantage Actor-Critic (A2C), and Deep Q-Network (DQN) and their specialized applications and the possible effect of DRL-based market-making on market dynamics. This project uses data science and machine learning to study Bitcoin market data. It's important for financial market analysis, especially in volatile and speculative cryptocurrency markets. The models' effectiveness is evaluated with spread capture ratio, market impact, and profitability. The findings can help academics and financial institutions understand how DRL can improve market efficiency and stability.