Gaussian Processes for Demand Unconstraining

TL;DR

This paper introduces a Gaussian process-based unconstraining method for demand data in revenue management, capable of handling complex, realistic demand patterns better than existing methods.

Contribution

The paper develops a novel non-stationary Gaussian process model with a new covariance function for demand unconstraining, improving performance on complex demand data.

Findings

GP method outperforms existing unconstraining techniques

Handles nonlinear demand trends and discontinuities effectively

Adapts to variations and periods of constraining in demand data

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 a novel 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.