Reinforcement Learning for Optimal Execution when Liquidity is Time-Varying

TL;DR

This paper demonstrates that Double Deep Q-learning can effectively learn optimal trading strategies in markets with dynamic, latent liquidity, outperforming traditional methods and approximations.

Contribution

It introduces a reinforcement learning approach to optimal execution under time-varying, unobservable liquidity, addressing a gap in existing models.

Findings

Learns optimal policies matching analytical solutions.

Outperforms benchmark and approximate methods.

Handles stochastic liquidity dynamics effectively.

Abstract

Optimal execution is an important problem faced by any trader. Most solutions are based on the assumption of constant market impact, while liquidity is known to be dynamic. Moreover, models with time-varying liquidity typically assume that it is observable, despite the fact that, in reality, it is latent and hard to measure in real time. In this paper we show that the use of Double Deep Q-learning, a form of Reinforcement Learning based on neural networks, is able to learn optimal trading policies when liquidity is time-varying. Specifically, we consider an Almgren-Chriss framework with temporary and permanent impact parameters following several deterministic and stochastic dynamics. Using extensive numerical experiments, we show that the trained algorithm learns the optimal policy when the analytical solution is available, and overcomes benchmarks and approximated solutions when the solution is not available.