考量不確定需求之貨櫃船運艙位配置模式

Abstract

imilar to the airline industry, the container shipping industry has a very high fixed cost and is thus a potential candidate for the application of revenue management (RM) or yield management (YM). However, in the effort of better utilizing the limited vessel capacity, booking requests are still usually managed by experienced employees without too much decision support in the container shipping industry. Some slot allocation models have been developed to determine the sales quota for a specific market (origin-destination pair). However, demand uncertainty is rarely considered in these decision models. This study develops a linear mathematical programming (LP) model to generate the booking limit for each origin-destination pair, with an uncertain demand represented by a random variable. In addition, the concept of overbooking is further incorporated to develop a modified LP model with over allocation by taking into account the loss due to unused capacity. An iterative approach is designed to successively estimate the opportunity cost of the capacity and re-solve the modified LP model to generate the inflated booking limits. A simulation experiment is performed by generating the booking requests and applying the booking limits. It is found the revenue obtained by the control decision based on the modified model is significantly higher than the one from the basic model or from the first-come-first-served (FCFS) policy.

Similar to the airline industry, the container shipping industry has a very high fixed cost and is thus a potential candidate for the application of revenue management (RM) or yield management (YM). However, in the effort of better utilizing the limited vessel capacity, booking requests are still usually managed by experienced employees without too much decision support in the container shipping industry. Some slot allocation models have been developed to determine the sales quota for a specific market (origin-destination pair). However, demand uncertainty is rarely considered in these decision models. This study develops a linear mathematical programming (LP) model to generate the booking limit for each origin-destination pair, with an uncertain demand represented by a random variable. In addition, the concept of overbooking is further incorporated to develop a modified LP model with over allocation by taking into account the loss due to unused capacity. An iterative approach is designed to successively estimate the opportunity cost of the capacity and re-solve the modified LP model to generate the inflated booking limits. A simulation experiment is performed by generating the booking requests and applying the booking limits. It is found the revenue obtained by the control decision based on the modified model is significantly higher than the one from the basic model or from the first-come-first-served (FCFS) policy.