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

Bingde Liu; John P Cartlidge

doi:10.46354/i3m.2023.dhss.001

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

Bingde Liu
, John P Cartlidge^*

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

99 Downloads (Pure)

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).

Original language	English
Title of host publication	13th International Defence and Homeland Security Simulation Workshop, DHSS 2023
Editors	Agostino G. Bruzzone, Benjamin Goldberg, Francesco Longo
Publisher	Caltek
Number of pages	10
ISBN (Electronic)	9788885741942
DOIs	https://doi.org/10.46354/i3m.2023.dhss.001
Publication status	Published - 18 Sept 2023
Event	International Defense and Homeland Security Simulation Workshop - Athens, Greece Duration: 18 Sept 2023 → 20 Sept 2023

Publication series

Name	Proceedings of the International Defense and Homeland Security Simulation Workshop, DHSS
Volume	2023-September
ISSN (Print)	2724-0363

Workshop

Workshop	International Defense and Homeland Security Simulation Workshop
Abbreviated title	DHSS 2023
Country/Territory	Greece
City	Athens
Period	18/09/23 → 20/09/23

Bibliographical note

Publisher Copyright:
© 2023 The Authors.

Keywords

Multi-Armed Bandit
Market Simulation
Bayesian optimization
Financial Trading
Automated Trading
Trading Agents

Access to Document

10.46354/i3m.2023.dhss.001Licence: CC BY-NC-ND

Full-text PDF (final published version)
This is the final published version of the article (version of record). It first appeared online via Caltek at https://doi.org/10.46354/i3m.2023.dhss.001 . Please refer to any applicable terms of use of the publisher.
Final published version, 554 KBLicence: CC BY-NC-ND

Handle.net

Persistent link

Cite this

Liu, B., & Cartlidge, J. P. (2023). Nonstationary Continuum-Armed Bandit Strategies for Automated Trading in a Simulated Financial Market. In A. G. Bruzzone, B. Goldberg, & F. Longo (Eds.), 13th International Defence and Homeland Security Simulation Workshop, DHSS 2023 Article 001 (Proceedings of the International Defense and Homeland Security Simulation Workshop, DHSS; Vol. 2023-September). Caltek. https://doi.org/10.46354/i3m.2023.dhss.001

Liu, Bingde ; Cartlidge, John P. / Nonstationary Continuum-Armed Bandit Strategies for Automated Trading in a Simulated Financial Market. 13th International Defence and Homeland Security Simulation Workshop, DHSS 2023. editor / Agostino G. Bruzzone ; Benjamin Goldberg ; Francesco Longo. Caltek, 2023. (Proceedings of the International Defense and Homeland Security Simulation Workshop, DHSS).

@inproceedings{1e11b465fd3c4eeabbf4e7dbb5589c89,

title = "Nonstationary Continuum-Armed Bandit Strategies for Automated Trading in a Simulated Financial Market",

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).",

keywords = "Multi-Armed Bandit, Market Simulation, Bayesian optimization, Financial Trading, Automated Trading, Trading Agents",

author = "Bingde Liu and Cartlidge, \{John P\}",

note = "Publisher Copyright: {\textcopyright} 2023 The Authors.; International Defense and Homeland Security Simulation Workshop, DHSS 2023 ; Conference date: 18-09-2023 Through 20-09-2023",

year = "2023",

month = sep,

day = "18",

doi = "10.46354/i3m.2023.dhss.001",

language = "English",

series = "Proceedings of the International Defense and Homeland Security Simulation Workshop, DHSS",

publisher = "Caltek",

editor = "Bruzzone, \{Agostino G.\} and Benjamin Goldberg and Francesco Longo",

booktitle = "13th International Defence and Homeland Security Simulation Workshop, DHSS 2023",

}

Liu, B & Cartlidge, JP 2023, Nonstationary Continuum-Armed Bandit Strategies for Automated Trading in a Simulated Financial Market. in AG Bruzzone, B Goldberg & F Longo (eds), 13th International Defence and Homeland Security Simulation Workshop, DHSS 2023., 001, Proceedings of the International Defense and Homeland Security Simulation Workshop, DHSS, vol. 2023-September, Caltek, International Defense and Homeland Security Simulation Workshop, Athens, Greece, 18/09/23. https://doi.org/10.46354/i3m.2023.dhss.001

Nonstationary Continuum-Armed Bandit Strategies for Automated Trading in a Simulated Financial Market. / Liu, Bingde; Cartlidge, John P.
13th International Defence and Homeland Security Simulation Workshop, DHSS 2023. ed. / Agostino G. Bruzzone; Benjamin Goldberg; Francesco Longo. Caltek, 2023. 001 (Proceedings of the International Defense and Homeland Security Simulation Workshop, DHSS; Vol. 2023-September).

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

TY - GEN

T1 - Nonstationary Continuum-Armed Bandit Strategies for Automated Trading in a Simulated Financial Market

AU - Liu, Bingde

AU - Cartlidge, John P

PY - 2023/9/18

Y1 - 2023/9/18

N2 - 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).

AB - 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).

KW - Multi-Armed Bandit

KW - Market Simulation

KW - Bayesian optimization

KW - Financial Trading

KW - Automated Trading

KW - Trading Agents

U2 - 10.46354/i3m.2023.dhss.001

DO - 10.46354/i3m.2023.dhss.001

M3 - Conference Contribution (Conference Proceeding)

T3 - Proceedings of the International Defense and Homeland Security Simulation Workshop, DHSS

BT - 13th International Defence and Homeland Security Simulation Workshop, DHSS 2023

A2 - Bruzzone, Agostino G.

A2 - Goldberg, Benjamin

A2 - Longo, Francesco

PB - Caltek

T2 - International Defense and Homeland Security Simulation Workshop

Y2 - 18 September 2023 through 20 September 2023

ER -

Liu B, Cartlidge JP. Nonstationary Continuum-Armed Bandit Strategies for Automated Trading in a Simulated Financial Market. In Bruzzone AG, Goldberg B, Longo F, editors, 13th International Defence and Homeland Security Simulation Workshop, DHSS 2023. Caltek. 2023. 001. (Proceedings of the International Defense and Homeland Security Simulation Workshop, DHSS). doi: 10.46354/i3m.2023.dhss.001

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

Abstract

Publication series

Workshop

Bibliographical note

Keywords

Access to Document

Handle.net

Fingerprint

Cite this