Imperial College London
Portrait Picture

Dr Mark H W Workman

Faculty of Engineering

 
 
 
//

Contact

 

mark.workman07 Website

 
 
//

Location

 

Centre for Environmental PolicyWeeks BuildingSouth Kensington Campus

//

Summary

 

Publications

Citation

BibTex format

@article{Radovic:2021,
author = {Radovic, D and Kruitwagen, L and Witt, CSD and Caldecott, B and Tomlinson, S and Workman, M},
journal = {Joule},
title = {Revealing robust oil and gas company macro-strategies using deep multi-agent reinforcement learning},
url = {http://arxiv.org/abs/2211.11043v1},
year = {2021}
}
Download

RIS format (EndNote, RefMan)

TY  - JOUR
AB  - The energy transition potentially poses an existential risk for majorinternational oil companies (IOCs) if they fail to adapt to low-carbon businessmodels. Projections of energy futures, however, are met with divergingassumptions on its scale and pace, causing disagreement among IOCdecision-makers and their stakeholders over what the business model of anincumbent fossil fuel company should be. In this work, we used deep multi-agentreinforcement learning to solve an energy systems wargame wherein playerssimulate IOC decision-making, including hydrocarbon and low-carbon investmentsdecisions, dividend policies, and capital structure measures, through anuncertain energy transition to explore critical and non-linear governancequestions, from leveraged transitions to reserve replacements. Adversarial playfacilitated by state-of-the-art algorithms revealed decision-making strategiesrobust to energy transition uncertainty and against multiple IOCs. In allgames, robust strategies emerged in the form of low-carbon business models as aresult of early transition-oriented movement. IOCs adopting such strategiesoutperformed business-as-usual and delayed transition strategies regardless ofhydrocarbon demand projections. In addition to maximizing value, thesestrategies benefit greater society by contributing substantial amounts ofcapital necessary to accelerate the global low-carbon energy transition. Ourfindings point towards the need for lenders and investors to effectivelymobilize transition-oriented finance and engage with IOCs to ensure responsiblereallocation of capital towards low-carbon business models that would enablethe emergence of fossil fuel incumbents as future low-carbon leaders.
AU  - Radovic,D
AU  - Kruitwagen,L
AU  - Witt,CSD
AU  - Caldecott,B
AU  - Tomlinson,S
AU  - Workman,M
PY  - 2021///
SN  - 2542-4351
TI  - Revealing robust oil and gas company macro-strategies using deep multi-agent reinforcement learning
T2  - Joule
UR  - http://arxiv.org/abs/2211.11043v1
UR  - http://hdl.handle.net/10044/1/101094
ER  -
Download