Applying Deep Reinforcement Learning to Mean-Reversion Trading

Summary

The document describes applying deep reinforcement learning to the trading problem of mean reversion. It aims to make recent reinforcement learning methods more accessible to practitioners who might otherwise need to choose among many approaches or build agents from classical foundations. The proposed demonstration uses a reinforcement learning library originally developed for strategic games and applies it to financial decision-making.

The framework also incorporates economically motivated properties of functions, with the stated goal of producing a convergent and high-performing solution. However, the document gives no specific model architecture, reward design, trading rules, data, benchmark, or numerical results. It therefore outlines an approach and its motivation rather than providing enough evidence to assess profitability, robustness, or performance against other methods. The claims about convergence and performance are presented without supporting experimental detail in this excerpt.

Key ideas

• The work applies deep reinforcement learning to mean-reversion trading.
• It demonstrates using a library initially developed for strategic games in a trading problem.
• The framework incorporates economically motivated function properties into the learning method.
• The document claims a convergent, high-performing solution but provides no experimental details or results here.
• Profitability and robustness cannot be assessed from the information provided.

Tags

# Deep Reinforcement Learning with Function Properties in Mean Reversion Strategies


# Deep Reinforcement Learning with Function Properties in Mean Reversion Strategies









Over the past decades, researchers have been pushing the limits of Deep Reinforcement Learning (DRL). Although DRL has attracted substantial interest from practitioners, many are blocked by having to search through a plethora of available methodologies that are seemingly alike, while others are still building RL agents from scratch based on classical theories. To address the aforementioned gaps in adopting the latest DRL methods, I am particularly interested in testing out if any of the recent technology developed by the leads in the field can be readily applied to a class of optimal trading problems. Unsurprisingly, many prominent breakthroughs in DRL are investigated and tested on strategic games: from AlphaGo to AlphaStar and at about the same time, OpenAI Five. Thus, in this writing, I want to show precisely how to use a DRL library that is initially built for games in a fundamental trading problem; mean reversion. And by introducing a framework that incorporates economically-motivated function properties, I also demonstrate, through the library, a highly-performant and convergent DRL solution to decision-making financial problems in general.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.