• Login
  • Register
  • Search

Research on Optimal Portfolio of DDPG Strategy Based on Deep Reinforcement Learning

Kai Jing


With the continuous development of artificial intelligence and large numbers, people are no longer satisfied with the traditional
investment portfolio method. This paper proposes a way of investment portfolio based on reinforcement learning, and researches stocks
through deep deterministic strategy gradient algorithm. We set up stock virtual account assets, set rewards and punishments, and charge certain transaction fees, so that the agent can carry out autonomous transactions to achieve the effect of decision optimization.


Deep Deterministic Strategy Gradient Algorithm; Agent; Decision Optimization

Full Text:


Included Database


[1] Shaokang D, Jiarui C, Yong L, et al. Reinforcement Learning from Algorithm Model to Industry Innovation Innovation: A Foundation Stone of Future Artificial Intelligence[J]. ZTE Communications,2019,17(03):31-41.

[2] Wang Kang, Bai Di. Research on Portfolio Management Based on Deep Reinforcement Learning [J]. Modern Computer,2021(01):3-11.

[3] Jiao Yuming. Stock portfolio Management and empirical research based on deep reinforcement learning [D]. Northwestern University, 2021.

[4] Chen Jia. Application of Deep reinforcement Learning in Stock Portfolio Management [D]. Huazhong University of Science and

Technology, 2022.

[5] Xu Jie, ZHU Yukun, Xing Chunxiao. Research on Financial transaction Algorithm based on deep reinforcement Learning [J]. Computer Engineering and Applications, 2022, 58(07): 276-285.

[6] Li Ming. Research on Stock Market Prediction Based on Deep Learning [D]. Nanjing University of Posts and Telecommunications,2019.

[7] Liu G. Research and application of stock market prediction model and evaluation method based on deep learning [D]. Beijing University of Posts and Telecommunications, 2020.

DOI: http://dx.doi.org/10.18686/fm.v9i2.12434