1. Improved multiple discreet-continuous extreme value model
- Author
-
Sindi, Alaa Rashad, Easa, Said M., and El Halim, Abd El Halim Omar Abd
- Subjects
Mathematical models -- Usage ,Transportation planning ,Cities and towns ,Engineering and manufacturing industries - Abstract
One of the advanced methods of modeling activity duration is the multiple discrete-continuous extreme value (MDCEV) model. The translating satiation parameter of the model is intended to capture the constant marginal utility effect, but model testing shows that it does not do so. In this paper, the structure of the translating satiation parameter was modified to incorporate an additional parameter that enables the model to capture this effect. Two datasets from cities in Saudi Arabia and Germany were used to test the applicability of the modified model for any dataset. The results showed that the proposed model structure increased the accuracy of activity duration prediction by up to 74%. The modified model represents an improvement to the conventional MDCEV model in terms of accuracy and flexibility, and as such should be a valuable tool in transportation planning and management.Key words: activity duration model, translating satiation parameter, constant marginal utility.Une des methodes avancees de modelisation de duree d'activite est le modele de valeurs extremes multiples discretes continues (<< MDCEV >>). Le parametre de traduction de satiete du modele est destine a capturer l'effet de l'utilite marginale constante, mais les essais du modele montrent qu'il ne fait pas ainsi. Dans le cadre de cette etude, la structure du parametre de traduction de satiete a ete modifiee afin d'incorporer un parametre supplementaire qui permet au modele de reproduire cet effet. Deux ensembles de donnees de villes en Arabie Saoudite et en Allemagne ont ete utilises afin de verifier l'applicabilite du modele modifie dans le contexte de n'importe quel ensemble de donnees. Les resultats ont revele que la structure du modele propose a donne lieu a une augmentation de l'exactitude de prediction de duree d'activite par au-dela de 74%. Le modele modifie represente une amelioration du modele de MDCEV traditionnel en matiere d'exactitude et de flexibilite et comme tel devrait etre un outil precieux dans la planification et la gestion du transport. [Traduit par la Redaction]Mots-cles: modele de duree d'activite, parametre de traduction de satiete, utilite marginale constante., IntroductionThe multiple discrete-continuous extreme value (MDCEV) model proposed by Bhat (2005) is an advanced model that uses random utility maximization theory and the Kuhn-Tucker demand system model to estimate activity [...]
- Published
- 2018
- Full Text
- View/download PDF