Incorporating continuous representation of preferences for flight departure times into stated itinerary choice modeling

Abstract

The passenger volumes of cross-Strait air travel between Taiwan and China have grown rapidly, largely due to the gradual introduction of direct flights. This study captures passengers’ preferences for airlines and flight departure times in the Taiwan–China market. The model structures examined include the multinomial logit, nested logit, and mixed nested logit models. In particular, trigonometric functions are incorporated into choice modeling to identify passenger preferences regarding flight departure times. A stated preference survey was conducted at Taiwan Taoyuan International Airport. A mixed nested logit model, which captures substitution patterns among competitive alternatives as well as individual heterogeneity, outperforms the other models. The estimation results indicate high degrees of competition among airlines and flight-departure times, especially highly substitutable departure times within the same airline. The continuous departure-time specification, which uses a series of sine and cosine functions, adequately captured passenger preferences. From the qualitative aspect of airline services, a premium for service improvement is highest for cabin crew service, followed by onboard food and check-in service. Airlines can use these results to improve their service quality and adjust flight schedules.