Practical Research on Deep Reinforcement Learning in Financial Trading Strategy Optimization

DOI

10.2991/978-94-6463-742-7_49How to use a DOI?

Keywords

deep reinforcement learning; financial trading; strategy optimization; adaptive risk control; deep Q network; futures; foreign exchange

Abstract

This paper proposes an adaptive risk control deep Q network (ARC-DQN) model based on deep reinforcement learning, focusing on the optimization of futures and foreign exchange trading strategies. The model introduces risk compensation factors and adaptive weight adjustment mechanisms to incorporate the dynamic balance between returns and risks into the decision-making process, achieving robust trading in a highly volatile market environment based on the standard DQN framework. The model represents market data with a multi-dimensional state vector, including prices, technical indicators, and risk indicators; the reward function is adjusted through a risk compensation mechanism during the action selection process, thereby effectively controlling volatility and maximum drawdown. The experiment uses historical data from the futures and foreign exchange markets for simulation. The results show that the ARC-DQN model improves the cumulative rate of return by about 22% compared with the traditional DQN, the Sharpe ratio increases by about 18%, the maximum drawdown decreases by about 15%, and the system response time is less than 2 s.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Jiaqi Liang
AU  - Yanru Zhao
PY  - 2025
DA  - 2025/05/31
TI  - Practical Research on Deep Reinforcement Learning in Financial Trading Strategy Optimization
BT  - Proceedings of the 2025 4th International Conference on Bigdata Blockchain and Economy Management (ICBBEM 2025)
PB  - Atlantis Press
SP  - 512
EP  - 523
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6463-742-7_49
DO  - 10.2991/978-94-6463-742-7_49
ID  - Liang2025
ER  -