Mixed Logit Model Application in Mode Choice: Case of Mashhad Work Trips

Modal split models, as the third step in the four-step transportation modeling framework, determine the share of different travel modes. Choice models as probability models have been used in recent decades and have faced significant progress. The mixed logit model has been known for many years, but has only become fully applicable since the advent of computer and simulation technology. This model can approximate various random utility models according to the accuracy required, through adopting appropriate distributions for attributes coefficients in the utility function. The purpose of this research is to present a mixed logit model structure for mode choice, in order to describe the taste variation among individuals and the source of the variation in response to the various attributes that influence the mode choice. The required data is from Mashhad O-D survey in 1387 and model calibration is executed in Biogeme software. Results of mixed logit model indicates among passengers a taste variation in choosing between a personal car and motorcycle, based on car and motorcycle ownership. The source of this taste variation is modeled and captured through random coefficient analysis. Finally, it is shown that mixed logit models are superior to multinomial logit model with a confidence level of 99 percent. The superiority is however small, partly due to the inadequacy of the aggregate data.