Gaussian processes for unconstraining demand

Ilan Price, Jaroslav Fowkes, Daniel Hopman

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

One of the key challenges in revenue management is unconstraining demand data. Existing state of the art single-class unconstraining methods make restrictive assumptions about the form of the underlying demand and can perform poorly when applied to data which breaks these assumptions. In this paper, we propose an unconstraining method that uses Gaussian process (GP) regression. We develop a novel GP model by constructing and implementing a new non-stationary covariance function for the GP which enables it to learn and extrapolate the underlying demand trend. We show that this method can cope with important features of realistic demand data, including nonlinear demand trends, variations in total demand, lengthy periods of constraining, non-exponential inter-arrival times, and discontinuities/changepoints in demand data. In all such circumstances, our results indicate that GPs outperform existing single-class unconstraining methods.

Original languageEnglish
Pages (from-to)621-634
Number of pages14
JournalEuropean Journal of Operational Research
Volume275
Issue number2
DOIs
Publication statusPublished - 1 Jun 2019

Fingerprint

Gaussian Process
Revenue Management
Extrapolate
Covariance Function
Arrival Time
Change Point
Gaussian Model
Demand
Gaussian process
Process Model
Discontinuity
Regression

Keywords

  • Demand unconstraining
  • Gaussian process regression
  • OR in airlines
  • Revenue management

Cite this

Price, Ilan ; Fowkes, Jaroslav ; Hopman, Daniel. / Gaussian processes for unconstraining demand. In: European Journal of Operational Research. 2019 ; Vol. 275, No. 2. pp. 621-634.
@article{9e03cba656454dfc9dc0caf846c32b69,
title = "Gaussian processes for unconstraining demand",
abstract = "One of the key challenges in revenue management is unconstraining demand data. Existing state of the art single-class unconstraining methods make restrictive assumptions about the form of the underlying demand and can perform poorly when applied to data which breaks these assumptions. In this paper, we propose an unconstraining method that uses Gaussian process (GP) regression. We develop a novel GP model by constructing and implementing a new non-stationary covariance function for the GP which enables it to learn and extrapolate the underlying demand trend. We show that this method can cope with important features of realistic demand data, including nonlinear demand trends, variations in total demand, lengthy periods of constraining, non-exponential inter-arrival times, and discontinuities/changepoints in demand data. In all such circumstances, our results indicate that GPs outperform existing single-class unconstraining methods.",
keywords = "Demand unconstraining, Gaussian process regression, OR in airlines, Revenue management",
author = "Ilan Price and Jaroslav Fowkes and Daniel Hopman",
year = "2019",
month = "6",
day = "1",
doi = "10.1016/j.ejor.2018.11.065",
language = "English",
volume = "275",
pages = "621--634",
journal = "European Journal of Operational Research",
issn = "0377-2217",
publisher = "Elsevier",
number = "2",

}

Gaussian processes for unconstraining demand. / Price, Ilan; Fowkes, Jaroslav; Hopman, Daniel.

In: European Journal of Operational Research, Vol. 275, No. 2, 01.06.2019, p. 621-634.

Research output: Contribution to JournalArticleAcademicpeer-review

TY - JOUR

T1 - Gaussian processes for unconstraining demand

AU - Price, Ilan

AU - Fowkes, Jaroslav

AU - Hopman, Daniel

PY - 2019/6/1

Y1 - 2019/6/1

N2 - One of the key challenges in revenue management is unconstraining demand data. Existing state of the art single-class unconstraining methods make restrictive assumptions about the form of the underlying demand and can perform poorly when applied to data which breaks these assumptions. In this paper, we propose an unconstraining method that uses Gaussian process (GP) regression. We develop a novel GP model by constructing and implementing a new non-stationary covariance function for the GP which enables it to learn and extrapolate the underlying demand trend. We show that this method can cope with important features of realistic demand data, including nonlinear demand trends, variations in total demand, lengthy periods of constraining, non-exponential inter-arrival times, and discontinuities/changepoints in demand data. In all such circumstances, our results indicate that GPs outperform existing single-class unconstraining methods.

AB - One of the key challenges in revenue management is unconstraining demand data. Existing state of the art single-class unconstraining methods make restrictive assumptions about the form of the underlying demand and can perform poorly when applied to data which breaks these assumptions. In this paper, we propose an unconstraining method that uses Gaussian process (GP) regression. We develop a novel GP model by constructing and implementing a new non-stationary covariance function for the GP which enables it to learn and extrapolate the underlying demand trend. We show that this method can cope with important features of realistic demand data, including nonlinear demand trends, variations in total demand, lengthy periods of constraining, non-exponential inter-arrival times, and discontinuities/changepoints in demand data. In all such circumstances, our results indicate that GPs outperform existing single-class unconstraining methods.

KW - Demand unconstraining

KW - Gaussian process regression

KW - OR in airlines

KW - Revenue management

UR - http://www.scopus.com/inward/record.url?scp=85059756209&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85059756209&partnerID=8YFLogxK

U2 - 10.1016/j.ejor.2018.11.065

DO - 10.1016/j.ejor.2018.11.065

M3 - Article

VL - 275

SP - 621

EP - 634

JO - European Journal of Operational Research

JF - European Journal of Operational Research

SN - 0377-2217

IS - 2

ER -