טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
Ph.D Thesis
Ph.D StudentFreund Feinstein Uzi
SubjectInvestigation of the Flight Product Choice Problem and the
Dynamic Adaptation of Supply
DepartmentDepartment of Civil and Environmental Engineering
Supervisor Professor Shlomo Bekhor
Full Thesis text - in Hebrew Full thesis text - Hebrew Version


Abstract

Customers' decision making in general, and choice of airline service in particular, was broadly researched. Although consumers can maximize utility by searching for itineraries using airlines and online websites, and monitoring changes in their offering, such behavior is not reflected in current literature. This study fills this gap, by assuming that the choice process is not a one-time event but a continuous one taking place during the choice period, when customers can make various decisions in compliance with their preferences to maximize their utility. They can choose from the initial itineraries offering, or defer choice up to the end of the sales period, or decline choice altogether during the sales period. This process varies due to personality characteristics, proposed alternatives, and duration of the choice period.

Conducting such choice investigation required a suitable methodology. A unique general theoretical model and its application to airline itinerary choice were formulated. This model of decision making in dynamic settings allows choice deferring and declining. Its novelty is that the choice problem is treated not as a one-time event but as a prolonged one over a certain period. Simultaneously, itineraries supply is treated as an on-going event with a composition of dynamically changing characteristics and static ones in each itinerary.

Empirical investigation was conducted among more than 10,000 persons who received a complex online stated preference questionnaire, which included several choice sets per experiment. Real itineraries data on medium- and long-haul flights, with dynamic and static variables were used to construct tourist and business-persons versions. Choice behavior was modeled using 4,141 observations for calibrating four discrete choice models. These models range from the simple multinomial logit, models that account for partial similarity such as the nested logit, models that account for full similarity such as the cross-nested logit, and models that account for both similarity among alternatives and correlation due to repeated observations such as the mixed cross-nested logit.

Findings indicate that the developed methodology is applicable for investigating itinerary choice behavior in dynamic settings, including the validity of the theoretic model formulated in this study, the ability to capture the tendency to defer choice and the performance of the various models. Further to these general conclusions it was found that difference in the length of choice deferring, although small, is visible between tourists and business-travelers. The proximity of the first itinerary search to the departure date affects the tendency to defer choice and instantaneous choice with increasing proportion of the latter as shorter sales periods are allocated to the consumer. Finally, the examination of estimation results indicates the effect of dynamic, static and demographic variables on decision making in dynamic settings.

The results indicate the research and application potentials of modeling choice in dynamic situation, for air transportation and other fields like tourism. Since this study used a limited sample and hypothetical choices, follow-up studies will apply the methodology to larger samples, with hypothetic and real choices in air transportation and other research fields, where choice is carried out in dynamic situations.