Modelling departure time, destination and travel mode choices by using generalized nested logit model: Discretionary trips

Despite traditional four-step model is the most prominent model in majority of travel demand analysis, it does not represent the potential correlations within different travel dimensions. As a result, some researches have suggested the use of choice modelling instead. However, most of them have represented travel dimensions individually rather than jointly. This research aims to fill this gap through employing the Generalized Nested Logit model for jointly representing three major travel dimensions; destination, departure time and travel mode. The suggested research methodology depends mainly on agglomerating alternatives that have similar error term's variances within specific gaps under common nests without any imposed restrictions. Moreover, different variance gaps lead to overlapped nesting system which can enable analysers modelling inner and inter-correlation. The proposed approach has been examined through modelling individuals' choices among the main shopping destinations in Eskisehir city, Turkey. In the light of estimation results, the proposed model attains a relatively good over-all goodness of fit which reflects a more prominent predictability power. Moreover, individuals in Eskisehir have been found perceiving more interest to the cost rather than time. From another hand, a behaviour of trading-off between performing such trips at peak periods by using transit or making them at off-peak by private car has been detected.