A discrete-time dynamic programming approach for network airline revenue management

Abstract

A discrete-time dynamic programming (DP) model is proposed to determine the optimal bid price for each booking request under both single and multiple booking scenarios. The optimal revenue function and opportunity cost are analysed and an asymptotical property is presented. Several approximation schemes to solve the DP model are proposed, including the network decomposition method and the function approximation technique. The solution methodology is discussed and the performance of the different methods is evaluated by their computational efficiencies and revenue impacts. The results show that the proposed approximation algorithms represent a major step forward in terms of revenue growth and computational ability.