酒店收益管理外文文献翻译中英文2019.docx

外文文献翻译原文及译文标题：酒店收益管理中动态客房分配的解决方法中英文 2019文献出处：N. Aydin, S. I. Birbil[J]European Journal ofOperational Research, Volume 271, Issue 116 , November 2018,Pages 179-192译文字数： 4700多字原文Decomposition methods for dynamic room allocation in hotel revenue managementN.Aydin， S.I.BirbilAbstractLong-term stays are quite common in the hotel business. Consequently, it is crucial for the hotel managements to consider the allocation of available rooms to a stream of customers requesting to stay multiple days. This requirement leads to the solving of dynamic network revenue management problems that are computationally challenging. A remedy is to apply decomposition approaches so that an approximate solution can be obtained by solving many simpler problems. In this study, we investigate several room allocation policies in hotel revenue management. We work on various decomposition methods to find reservation policies for advance bookings and stay-over customers. We also devise solution algorithms to solve the resulting problems efficiently.Keywords：Revenue management，Hotel，Capacity control，Decomposition methodsIntroductionHistorically, the airline industry played the steering role in revenue management (RM). Today, however, there is a wide range of applications in different industries with volatile demand, requesting fixed and perishable capacity (Kimes, 1989). Although the hotel industry is one of the typical application areas of revenue management, the research in this particular area lags behind the work produced for other service industries. In their recent work, Ivanov and Zhechev (2012) and Ivanov (2014) present a review of the methods proposed in the hotel RM literature and point out the gaps.In general, well-known airline RM techniques, such as booking control and pricing, can be applied to hotel RM problems. However, it is important to consider several constraints that are unique to hotel reservation systems. First, multi-day stays in hotels are quite common. While a flight itinerary includes, on average fewer than three legs, the number of nights a typical customer spends in a hotel can be a week or even more (Zhang & Weatherford, 2017). Second, the demand process is different. Hotel customers may decide to stay longer and extend their reservation while they are staying in the hotel (Kimes, 1989). Third, airline customers generally make advance bookings but a number of hotel customers consist of walk-ins. Moreover, the early reservations in the booking interval are even allowed to cancel their bookings at no extra cost.In this paper, we focus on the room allocation decisions for a hotel. The optimal policy to accept or reject an arriving customer can be obtained by analyzing the stochastic nature of the customer arrival process. In hotel reservation systems, the customers are classified as the advance bookings, the stay-overs and the walk-ins. While the advance bookings make room reservations before they arrive at the hotel, the walk-ins show up without any reservation. The stay-overs are the customers who ask for an extension for their reservations during their stay in the hotel. Recently, hotel reservation systems have started offering extended stay as an option due to high customer demand (Tepper, 2015). For instance, Priceline (2017) and Hotwire (2017)present “add-a-night” and “add to your stay” options to their existing customers. The arrival process of the advance bookings and walk-ins are similar. The only difference is that the walk-in customers arrive after the reservation period ends. However, the stay-over requests depend on the accepted advance bookings. To simplify our notation, we ignore the walk-in customers and formulate our problem by considering the advance bookings and the stay-overs. Then, we explain how one can easily incorporate the walk-in customers to our proposed models. To the best of our knowledge, the dynamic model of stay-over customers in a network setting has not been previously studied in the literature.The research contributions in this paper come from the application and the analysis of two decomposition approaches. These are the day-based and the periodbased decompositions. Our day-based decomposition is similar to the one proposed by Kunnumkal and Topaloglu (2010). We simplify their decomposition method and show that our proposed model provides a lower bound to their model. We set forth a dynamic model for the advance bookings and formulate a linear program for the problem. The resulting model is then solved with the constraint generation method. We also propose alternate approximate models, which provide upper and lower bounds on the optimal expected revenue of the original model. To manage the stay- over requests, one needs to keep track of the number of reservations in each booking type. A day-based method, however, decomposes the network problem into independent days, and this decomposition approach causes loss of information on the number of custo。