Multinomial Logit Model of Home to Work Morning Peak Period Trip Mode

1438 Words3 Pages

A Multinomial Logit (MNL) Model of morning peak-period home-to-work (HW) trip mode choice is the proposed model for the particular study area to analyse explanatory variables influencing travel mode choice. This method is essential due to the fact that mode choice is not only affected by spatial constrains but also by socio-demographic constraints. To evaluate these variables, a Multinomial Logit Model is utilized and five mode choice utility functions are drafted in order to determine morning peak-period home-to-work trips. In order to conclude whether or not our proposed model can be suitably implemented, data from the regional travel survey is inputted to the model to yield real world results. The results of the model can predict with 50.7 percent certainty for the morning peak-period trips for this particular area of study. By means of determining an accurate morning peak-period home-to-work trip mode choice model, policymakers can utilize the results of this model to aide in decision-making for policies regarding transportation, sustainability, and future planning for this particular area of study.

A dataset was provided by a regional travel survey that was used for modelling purposes and included information regarding home-to-work (HW) morning peak-period trip mode explanatory variables. As seen below in Figure 1, a decision tree is presented outlining the HW morning peak-period trips that were modelled. A Multinomial Logit Model was determined to be the preferable model for this assignment because this particular model is capable of analysing multi-variable choices. In comparison, a model that is not suitable is a Binary Logit Model which would be used for models that have less than two variable choices.

Given are...

... middle of paper ...

...ors, arterial/collector/and local roads; optimal layout of transit; and streetscape design.

In order to obtain favourable and accurate results, an additional number of iterations of the dataset should be processed if time permits. Likewise, if time permits, it is preferable to begin with all the variables in each utility function and then work backwards by means of the complete backwards elimination process. Additionally, in order to obtain refined results, the utilization of advanced modelling software such as PTV Visum Software could be advantageous.

Works Cited

Bierlaire, M. (2003). BIOGEME: A free package for the estimation of discrete choice models, Proceedings of the 3rd Swiss Transportation Research Conference. Ascona, Switzerland.

Idris, O.A. (2014). Transportation Planning and Design. [Lecture notes]. Retrieved from https://connect.ubc.ca/

Open Document