Dueling Deep Reinforcement Learning for Financial Time Series
This work addresses financial trading optimization for investors, but it is incremental as it applies existing RL methods to a specific domain with practical constraints.
The researchers tackled the problem of optimizing financial trading strategies using reinforcement learning on historical SP500 data, and found that their agents, based on Double DQN and Dueling Networks, outperformed random strategies despite challenges like transaction costs and data complexity.
Reinforcement learning (RL) has emerged as a powerful paradigm for solving decision-making problems in dynamic environments. In this research, we explore the application of Double DQN (DDQN) and Dueling Network Architectures, to financial trading tasks using historical SP500 index data. Our focus is training agents capable of optimizing trading strategies while accounting for practical constraints such as transaction costs. The study evaluates the model performance across scenarios with and without commissions, highlighting the impact of cost-sensitive environments on reward dynamics. Despite computational limitations and the inherent complexity of financial time series data, the agent successfully learned meaningful trading policies. The findings confirm that RL agents, even when trained on limited datasets, can outperform random strategies by leveraging advanced architectures such as DDQN and Dueling Networks. However, significant challenges persist, particularly with a sub-optimal policy due to the complexity of data source.