Modeling oil production based on symbolic regression

Guangfei Yang, Xianneng Li, Jianliang Wang, Lian Lian, Tieju Ma

Research output: Contribution to journalArticle

19 Citations (Scopus)

Abstract

Numerous models have been proposed to forecast the future trends of oil production and almost all of them are based on some predefined assumptions with various uncertainties. In this study, we propose a novel data-driven approach that uses symbolic regression to model oil production. We validate our approach on both synthetic and real data, and the results prove that symbolic regression could effectively identify the true models beneath the oil production data and also make reliable predictions. Symbolic regression indicates that world oil production will peak in 2021, which broadly agrees with other techniques used by researchers. Our results also show that the rate of decline after the peak is almost half the rate of increase before the peak, and it takes nearly 12 years to drop 4% from the peak. These predictions are more optimistic than those in several other reports, and the smoother decline will provide the world, especially the developing countries, with more time to orchestrate mitigation plans.

Original languageEnglish
Pages (from-to)48-61
Number of pages14
JournalEnergy Policy
Volume82
Issue number1
DOIs
Publication statusPublished - 2015

Keywords

  • Hubbert theory
  • Oil production
  • Symbolic regression

ASJC Scopus subject areas

  • Energy(all)
  • Management, Monitoring, Policy and Law

Fingerprint Dive into the research topics of 'Modeling oil production based on symbolic regression'. Together they form a unique fingerprint.

  • Cite this

    Yang, G., Li, X., Wang, J., Lian, L., & Ma, T. (2015). Modeling oil production based on symbolic regression. Energy Policy, 82(1), 48-61. https://doi.org/10.1016/j.enpol.2015.02.016