Robust resource allocation under uncertainty

Abstract

Current methods for determining optimal capacity controls under uncertainty, such as stochastic optimization, often assume a known distribution for unknown parameters. This paper presents a novel approach using robust optimization to address the stochastic resource allocation problem in airline seat-inventory control. Our static formulations account for demand dependencies, offering a streamlined alternative to existing customer-choice models in revenue management literature. We analyze the structure of our proposed formulations, and provide insights on several robust counterparts of the seat-inventory control problem, considering various measures of robustness. We introduce algorithms based on these robust formulations to derive actionable results. Through extensive simulations focused on seat allocation problems within the revenue management domain, our proposed formulations demonstrate significantly improved worst-case performances. Notably, even under favorable scenarios, the performance of our solutions are comparable to those of the existing methods in the revenue management literature. By providing protection against forecasting errors in demand distribution parameters and offering improved booking limit controls when demand falls below expected value, our formulations demonstrate superior revenue retention compared to existing methods in our comparative analyses.