Managing hotel reservations with uncertain arrivals.

Based on our interactions with managers at two large hotels, we present a realistic model of the hotel reservation problem. Unlike traditional models, ours does not assume that all customers arrive simultaneously on the targeted booking date. We explain why this assumption may not be appropriate for the hotel industry and develop a model of reservation booking which explicitly includes the room allocation decisions which are made on the targeted booking date. Based on observations of how the problem is solved in practice as well as the insights gained from this analysis, we develop simple heuristic procedures for accepting reservations. Computational results demonstrate that these heuristics perform well relative to an upper bound that is based on perfect information about reservations requests and customer arrivals. [ABSTRACT FROM AUTHOR]