Tests of revenue management performance under different demand correlation assumptions

This article summarizes results from Passenger Origin–Destination Simulator (PODS) research on how a revenue management (RM) system performs under various assumptions about demand correlation (between early‐ and late‐arriving customers). PODS typically assumes a relatively strong positive correlation; this article shows whether and how much the results change under weak positive correlation. First, we explore the revenue results for three different dynamic user influence (UI) strategies – unbiased, biased low and biased high – under varying demand levels, with and without hybrid forecasting (HF). In short, dynamic UI seeks to emulate RM analysts’ attempts to positively influence the RM system and was originally presented by Hao at a 2013 PODS meeting. Next, we explore the revenue results for three different Origin–Destination (O–D) strategies under varying demand levels, with and without HF for different correlated demand scenarios. In both studies, we use a large ‘international’ network with 572 O–D markets with four airlines competing for passengers, including a low‐cost carrier.