Development of Trip Generation and Trip Distribution Models for Car-sharing Considering Land Use and Transport Characteristics

Abstract

This study presents models for trip generation and distribution of car-sharing services based on So-car’s operation history data, GIS building integrated information, and public transportation information for 5 days during weekdays in Seoul metropolitan areas. Based on the trip generation model, the numbers of workers and students, total areas of commerce, business, education, and cultural facilities, as well as the density of subway stations showed positive (+) correlations, but the density of bus stops showed a negative (−) correlation. It was confirmed that the use rate of car sharing increased as accessibility to public transportation was lower in the case of outer areas of the city without subway stations. The adjusted R2 of the trip generation model for car-sharing was calculated as 0.587–0.637, thus confirming that the explanatory power of the model is high. By constructing the trip distribution model for car sharing by time zone based on the existing gravity model, the explanatory power of the model (adjusted R2) was calculated to be as high as 0.869–0.966, and the modified trip distribution model for car sharing reflected the generalization costs of taxi and public transportation. It was confirmed that as the generalization costs of taxis and public transportation increased, the utilization rate of car sharing increased.