Research on the implementation and effectiveness evaluation of deep reinforcement learning algorithms for portfolio optimisation
Year: 2025
Type: article
Abstract: Portfolio optimisation is the process of continuously allocating money to different assets to maximize returns, and since the theory was introduced in 1952, many researchers have improved upon it by introducing constraints and other objectives to make the model more realistic and by using more powerful and intelligent optimisation algorithms to solve complex mathematical models. In recent years, Deep Reinforcement Learning (DRL) has become a popular branch of machine learning and has shown excellent performance in solving complex problems. Some recent studies have shown that using DRL methods to solve portfolio optimisation problems has very good potential. Therefore, this paper investigates the implementation and application effects of deep reinforcement learning algorithms in portfolio optimisation, applying the reinforcement learning frontier algorithms A2C (synchronous Advantage Actor-Critc) and PPO (Proximal Policy Optimisation Algorithm) and investigating their effectiveness on the A-share market and the constructed environment. The empirical results show that these two DRL models are effective in portfolio optimisation in the A-share market and can better capture the upward price trend.
Source: Discover Artificial Intelligence
Authors: Gao Yunxiang, Tang Bangying
Institution: PLA Academy of Military Science
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: gold
Estimated APC paid: $1,590
OpenAlex ID: W4415618713