Nonstationary Continuum-Armed Bandit Strategies for Automated Trading in a Simulated Financial Market

TL;DR

This paper introduces PRBO, a novel Bayesian optimization-based algorithm for automated trading that adapts to nonstationary market conditions, outperforming existing strategies in simulated financial markets.

Contribution

The paper presents PRBO, a new nonstationary bandit algorithm for automated trading that dynamically adjusts strategies using Bayesian optimization within a bandit-over-bandit framework.

Findings

PRBO outperforms the reference strategy PRSH in profit generation.

PRBO requires fewer hyperparameters to tune.

PRBO effectively adapts to changing market conditions in simulations.

Abstract

We approach the problem of designing an automated trading strategy that can consistently profit by adapting to changing market conditions. This challenge can be framed as a Nonstationary Continuum-Armed Bandit (NCAB) problem. To solve the NCAB problem, we propose PRBO, a novel trading algorithm that uses Bayesian optimization and a ``bandit-over-bandit'' framework to dynamically adjust strategy parameters in response to market conditions. We use Bristol Stock Exchange (BSE) to simulate financial markets containing heterogeneous populations of automated trading agents and compare PRBO with PRSH, a reference trading strategy that adapts strategy parameters through stochastic hill-climbing. Results show that PRBO generates significantly more profit than PRSH, despite having fewer hyperparameters to tune. The code for PRBO and performing experiments is available online open-source (https://github.com/HarmoniaLeo/PRZI-Bayesian-Optimisation).