In recent years a huge number of analyses and models have focused on comprehending and simulating the phenomenon of airport choice. In the last decade many researchers have employed models based on random utility theory, developing simple structures such as Multinomial Logit models (MNL), or more complex ones allowing for correlation between different alternatives (airports) and accommodating the fact that passenger behaviour varies across different groups of travellers. Hierarchical Logit models (HL) and Cross-Ne¬sted Logit (CNL) models have been proposed to take into account the influence of other dimensions of travel behaviour on a passenger’s choice of airport. Mixed Multinomial Logit models (MMNL) have been proposed to si¬mulate whether and how passenger behaviour varies randomly within individual groups of travellers. Combinations of two (airport and airline; airport and access mode) or three choice dimensions (airport, airline and access mode) have been investigated, and models have been proposed for different user classes, different trip purposes. Most of the models proposed in the literature have allowed broad understanding of the phenomenon, the influence of other choice dimensions, the re¬levance of specific attributes and their reciprocal weights. Much has been do¬ne, nevertheless some considerations should be made. Much work has been developed starting from aggregate data (demand flow and aggregate level of service attributes) and it is mainly based on MNL models (logistic functions). Such an approach should be interpreted as an effective regressive method, but it does not allow for user characteristics, trip characteristics or the incidence of relevant attributes such as air-fare, travel time and waiting time. Contributions based on disaggregate data use mainly revealed preference data (RP), whereas there are very few contributions based on stated preferences (SP) (Algers and Beser, 1997; Hess et al., 2007). Although a wide variety of models can be found in the literature, the main drawback of RP data is the difficulty estimating level of service attributes for the alternative chosen and especially for the alternatives not chosen. Most research using RP data, cannot rely on suitable detailed information on level-of-service (air fares) and, in particular, it interpret flight availability or capacity problems through frequency attributes, airline dummy variables or through the air fare itself. Among the implemented models, it is necessary to distinguish models that simulate airport choice only, and the models that simulate the combination of different choice dimensions such as airport-airline, airport-flight, airport- airline-flight. As regards the latter models, it should be no¬ted that the best-performing ones present complex utility functions, which are not easy to apply: they require a large amount of information, which if available in the survey, might not be easily known by the analyst and/or might not be easily forecasted in operational scenarios (air fares, frequency). As a consequence, the models cannot be applied to strategic planning scenarios, particularly when it is not possible to have any information on services offered. Mo-reover, more complex models applied to more complex choice sets (CNL and MML) do not seem to outperform the MNL (or HL) model sharply. In many cases the differences are negligible and no elasticity analysis is carried out, while trade-off analysis highlight small differences. Finally, most of the existing applications are carried out on the same area (San Francisco or London), where the choice process is among airports belon¬ging to a multi-airport region which is well defined and consolidated in the user’s mind, and where the low-cost phenomenon is also consolidated. In this paper a set of models simulating airport choice behaviours is proposed. The models are based on random utility theory, are easy to implement, address the main issues discussed above, investigate the incidence of trip type and cope with a choice-set, constituted by airports of different type (intercontinental airport, regional airport and city airport) that compete with one another on medium/short haul trips at a European scale. The models are calibrated on disaggregate data obtained by a SP survey. In all, 600 users from Campania (a region in southern Italy) were asked to face realist choice scenarios, built from real data taken from the main web-sellers. Each respondent had to choose the preferred solution to fly towards a given European capital city. The trip purpose analyzed was leisure and the choice set was defined by analyzing the services offered by the airports of Naples, Rome Fiumicino and Rome Ciampino. The quoted airports belong to different Italian regions (Campania and Lazio), they represent a realistic choice set for the users interviewed and present different characteristics with respect to accessibility, flight frequencies and services supplied. The choice context is interesting since it allows us to interpret and simulate competition among larger yet congested airports, city airports and regional airports typically used by low-cost airlines. Such a scenario is growing rapidly in Italy (the low-cost phenomenon is quite new and increasing rapidly) and in many European regions. Hence the results may be transferred to similar choice contexts. Different time horizons were investigated (departure in week 1 and trip duration of 3 days, departure in 3 months and trip duration 7 days), as well as the incidence of users’ past experience, were analyzed by estimating specific models or through specific attributes in the systematic utility functions. Moreover, non-linearity in the attributes of the systematic utilities was investigated and interesting results were achieved.
Analysis of competition between airports: travellers airport choice models
DE LUCA, STEFANO
2008-01-01
Abstract
In recent years a huge number of analyses and models have focused on comprehending and simulating the phenomenon of airport choice. In the last decade many researchers have employed models based on random utility theory, developing simple structures such as Multinomial Logit models (MNL), or more complex ones allowing for correlation between different alternatives (airports) and accommodating the fact that passenger behaviour varies across different groups of travellers. Hierarchical Logit models (HL) and Cross-Ne¬sted Logit (CNL) models have been proposed to take into account the influence of other dimensions of travel behaviour on a passenger’s choice of airport. Mixed Multinomial Logit models (MMNL) have been proposed to si¬mulate whether and how passenger behaviour varies randomly within individual groups of travellers. Combinations of two (airport and airline; airport and access mode) or three choice dimensions (airport, airline and access mode) have been investigated, and models have been proposed for different user classes, different trip purposes. Most of the models proposed in the literature have allowed broad understanding of the phenomenon, the influence of other choice dimensions, the re¬levance of specific attributes and their reciprocal weights. Much has been do¬ne, nevertheless some considerations should be made. Much work has been developed starting from aggregate data (demand flow and aggregate level of service attributes) and it is mainly based on MNL models (logistic functions). Such an approach should be interpreted as an effective regressive method, but it does not allow for user characteristics, trip characteristics or the incidence of relevant attributes such as air-fare, travel time and waiting time. Contributions based on disaggregate data use mainly revealed preference data (RP), whereas there are very few contributions based on stated preferences (SP) (Algers and Beser, 1997; Hess et al., 2007). Although a wide variety of models can be found in the literature, the main drawback of RP data is the difficulty estimating level of service attributes for the alternative chosen and especially for the alternatives not chosen. Most research using RP data, cannot rely on suitable detailed information on level-of-service (air fares) and, in particular, it interpret flight availability or capacity problems through frequency attributes, airline dummy variables or through the air fare itself. Among the implemented models, it is necessary to distinguish models that simulate airport choice only, and the models that simulate the combination of different choice dimensions such as airport-airline, airport-flight, airport- airline-flight. As regards the latter models, it should be no¬ted that the best-performing ones present complex utility functions, which are not easy to apply: they require a large amount of information, which if available in the survey, might not be easily known by the analyst and/or might not be easily forecasted in operational scenarios (air fares, frequency). As a consequence, the models cannot be applied to strategic planning scenarios, particularly when it is not possible to have any information on services offered. Mo-reover, more complex models applied to more complex choice sets (CNL and MML) do not seem to outperform the MNL (or HL) model sharply. In many cases the differences are negligible and no elasticity analysis is carried out, while trade-off analysis highlight small differences. Finally, most of the existing applications are carried out on the same area (San Francisco or London), where the choice process is among airports belon¬ging to a multi-airport region which is well defined and consolidated in the user’s mind, and where the low-cost phenomenon is also consolidated. In this paper a set of models simulating airport choice behaviours is proposed. The models are based on random utility theory, are easy to implement, address the main issues discussed above, investigate the incidence of trip type and cope with a choice-set, constituted by airports of different type (intercontinental airport, regional airport and city airport) that compete with one another on medium/short haul trips at a European scale. The models are calibrated on disaggregate data obtained by a SP survey. In all, 600 users from Campania (a region in southern Italy) were asked to face realist choice scenarios, built from real data taken from the main web-sellers. Each respondent had to choose the preferred solution to fly towards a given European capital city. The trip purpose analyzed was leisure and the choice set was defined by analyzing the services offered by the airports of Naples, Rome Fiumicino and Rome Ciampino. The quoted airports belong to different Italian regions (Campania and Lazio), they represent a realistic choice set for the users interviewed and present different characteristics with respect to accessibility, flight frequencies and services supplied. The choice context is interesting since it allows us to interpret and simulate competition among larger yet congested airports, city airports and regional airports typically used by low-cost airlines. Such a scenario is growing rapidly in Italy (the low-cost phenomenon is quite new and increasing rapidly) and in many European regions. Hence the results may be transferred to similar choice contexts. Different time horizons were investigated (departure in week 1 and trip duration of 3 days, departure in 3 months and trip duration 7 days), as well as the incidence of users’ past experience, were analyzed by estimating specific models or through specific attributes in the systematic utility functions. Moreover, non-linearity in the attributes of the systematic utilities was investigated and interesting results were achieved.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
