Combined multinomial logit modal split and paired combinatorial logit traffic assignment model

Abstract

ABSTRACTTo better address the route overlap problem of the multinomial logit model used in combined modal split and traffic assignment models in the literature, this study proposes a combined multinomial logit modal split and paired combinatorial logit traffic assignment (MNL–PCL) model. The PCL model can account for the route overlap problem using a similarity index for each pair of routes in the network. It requires significantly fewer parameters to be calibrated using real-world data. Thereby, it circumvents parameter estimation issues associated with a cross-nested logit model. An equivalent mathematical programming problem is developed for the MNL–PCL model. Further, an analytical model is developed for sensitivity analysis of the MNL–PCL model. Several applications of the proposed MNL–PCL model are demonstrated using a numerical example by leveraging the results of sensitivity analysis. The study insights can assist decision-makers to design more effective strategies to promote ‘go-green’ travel modes and reduce network congestion.