Time series uni-mode or random utility multi-mode approach in national passenger models: the impact on Italian air demand forecast.

The increase of the air transport demand in the last decades, supported also by deregulation policy, has had as a relevant effect the increase of transport services offered by the different air carrier companies and, as a negative consequence, the increase of the congestion level both in the airways and at the airports. The air transportation demand directly affects the planning of the airport terminals in terms of design of the land services (check-in/check-out points, waiting areas, facilities as restaurants, shops and so on). In order to achieve the objective of a correct design of such areas, both the absolute demand value and its temporal distribution are needed. Multi-mode models proposed in literature to forecast travel demand at a national level can be used to obtain an estimate of the air travel demand, by using explicitly random utility models for simulating the mode choice. However, this kind of model does not analyse the temporal evolution of the demand nor the specificity of the most important hubs subject to the increase of specific demand segments. Uni-mode time series models allow us to overcome these difficulties but in literature only uni-mode models calibrated at inter-national level are proposed. In this work a uni-mode time series model with a first order correlation error term has been developed for the estimation of the air travel demand at national level with particular reference to specific attractive national basins. The explanatory variables used are activity variables (as population in the origin/destination areas, pro capita income) and level-of-services variables (as number of flights and the corresponding run time; direct or indirect connections; dimension of the aircraft; variability index with reference to the frequency). The variability index considers the temporal distribution of the flights for each origin/destination pair, and it is an indirect indicator of the temporal distribution of the air transport demand. For the estimation of the model parameters both cross-section and time-series data have been considered for the last decade. Furthermore, the elasticity values obtained by using both the multi-mode models and the uni-mode model are compared as well as different scenarios for different configurations of the flight offer; the corresponding variations of the air transport demand have been estimated in order to define the supply boundary with respect to the actual airport capacity.