National Technical University of Athens (NTUA)
Συμβολή στη διερεύνηση των παραγόντων διαμόρφωσης της ζήτησης και επιλογής μέσου πρόσβασης σε αεροδρόμια
Abstract
dc:descriptionIn many regions, capacity limitations on existing airports in combination with estimates for future traffic growth led airport authorities to move airports to new locations. Oslo, Hong Kong and Athens are typical examples. In such cases airport infrastructure development is accompanied by road and/or rail infrastructure needed for the access to the airport by passengers, cargo, visitors and airport employees. Access mode choice has a direct impact on the business plan of the relocated airport. It affects parking service provision, commercial activities, curbside management, and the determination of capacity requirements, capital expenditures, revenues, cashflows and pricing policies. It is also important for the development of access to the airport infrastructure (new rail links, metro extension, road network expansion, bus services). In this thesis a method is proposed that predicts behavioral changes in the relocated airport using current practice in the existing airport. The development of these forecasts cannot be based on benchmarking techniques and on mere transfer of experience from other airports. The access conditions observed at any airport are unique and depend on the nature and volume of airport traffic; the location and geographical setting of the airport and the urban areas it serves and the economic, social and political structure of the country in which it is situated. Assessment of world’s larger airports shows a significant diversity in solutions and discourages easy generalisations. First access modal split at the existing airport is determined via traveler classification and discrete choice modeling. Both are conducted via a passenger survey study specifically designed for this purpose. The survey is conducted at the operating airport with the goal to provide sufficient and accurate information to determine the model parameters and to secure that the derived results are reliable. The questions of the survey are structured into three groups in correspondence with attribute determination, traveler classification and mode choice. Questions in the first group request information about trip origin, car ownership, taxi fare if taxi were used, luggage items, occupation, age, time spent at the airport, data about the return-home flight. The last two items are used to determine parking costs. Questions in the second group seek information that differentiates travelers in their choice of access mode. Place of residence (city hosting the airport, other city, abroad), trip purpose (business, holiday/tourism, visit relatives/friends), are typical representatives. Questions in the third group request information about the mode used to access the airport (drive and park, car passenger with driver also flying, offered a lift to the airport by someone who either parked and accompanied the passenger to the terminal or dropped him off at the curbside, taxi, public bus, other bus, rental car). Traveler classification produces a partition of travelers into groups such that preferences within groups exhibit a certain degree of homogeneity. This homogeneity is manifested by a set of common characteristics in decisions of mode choice. Residence status, trip purpose and trip destination, provide typical segmentation variables. Segmentation variables are selected based on observation of real behavior and intuition. It is often useful that the resulting categories are further split or merged, depending on their ability to represent distinct behavioral patterns with respect to access choice. A hierarchical agglomerative clustering algorithm is employed to further enhance segmentation and to support the differentiation and cohesiveness of the travelers’ clusters. The access behavior of each group is governed by a discrete choice model. Typical access mode choices are: ‘drvp’ (drive and park), ‘lift’ (offered a ride by a third party; it involves short term parking), ‘drop’ (dropped off, no parking), taxi, public buses, tourist buses and rental cars. The attribute values that are required by each discrete choice model are determined by the so-called attribute encoder. Typical attributes are travel time, travel cost and car ownership. A linear in the parameters utility function for each passenger and each access mode is employed. Maximisation of the likelihood function over the data sample is used for the estimation of the parameters. In order to select an appropriate discrete choice model, both nested and non-nested structures are considered. Non-nested models include: heteroscedastic extreme value models, multinomial probit models and random parameter models. Nested forms based on various tree structures are also examined. The proposed methodology was applied in the new Athens international airport. It was found that Athens residents on domestic non-business trips show a somewhat similar behaviour with international business travelers, as expected from the classification analysis. The dependence on other drivers (‘drop’, ‘lift’) looses 9% in favour of autonomy (‘drive and park’, ‘taxi’). Airport relocation does not cause significant behavioural changes to Athens residents on international non-business trips. These travelers fly abroad mainly for tourism for longer duration. Their hesitation to park and leave the car sustains in the new airport. Athens residents on domestic business trips tend to prefer the ‘drive and park’ option. The short duration of these trips minimizes the generalized parking cost. This feature is evident in the access choice for the relocated airport. In relation to the ‘self-driver/other driver’ partition that separates ‘drvp’ from the remaining mode alternatives, across the board losses incurred in the ‘other driver’ share are gained by ‘drive and park’. The methodological elements described above (passenger classification, passenger and parking survey, access choice model and parking capacity estimation) can be combined into a flexible parking forecasting tool that enables the analysis and assessment of diverse access demand and supply characteristics.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2002
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Psaraki-Kalouptsidi, Voula
- Ψαράκη-Καλουπτσίδη, Παρασκευή
Subjects
dc:subject × 12Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/16596