Correspondence: Richard M Ratliff, Sabre Holdings, 3150 Sabre Drive, Southlake, Texas 76092, USA. Tel: +1 682 605 1710; Fax: +1 682 605 7679; E-mail: richard.ratliff@sabre-holdings.com

¹Richard M. Ratliff is a senior research scientist in the Sabre Holdings research group. He leads advanced R&D efforts in airline pricing and revenue management including collaboration with universities.

²Beju V. Rao's efforts on this paper were performed while he was a Senior Operations Research Principal with Sabre Airline Solutions. Currently, he is a director and head of Analytical Center of Excellence within Corporate Credit Review Services at Capital One Financial Corporation.

³Chittur P. Narayan is a Principal in the Revenue Planning group at Sabre Airline Solutions since 1996. He has worked on a variety of decision support projects in the areas of network planning, scheduling, pricing, revenue management and airline cargo.

⁴Kartik Yellepeddi is a Senior Operations Research Analyst at Sabre Airline Solutions. Currently, he leads the Operations Research team in Bangalore supporting models in Airline Network Planning & Scheduling, Revenue Management and Pricing.

Received 1 October 2007; Revised 1 October 2007; Published online 11 April 2008.

Abstract

A new approach is described for airline revenue management (RM) demand unconstraining considering market recapture effects across multiple flights and fare classes. A good heuristic involving customer choice models is shown to provide a computationally efficient method for estimating demand, spill and recapture from historical bookings and availability data. By jointly estimating spill on closed alternatives and recapture onto open ones, the pervasive problem of double counted demands in RM system forecasts is avoided. Initial implementation experience shows the new methodology can be readily integrated into existing demand forecasting systems and should reduce forecast errors.