文章基本信息

标题：A framework to generate virtual cities as sandboxes for land use-transport interaction models
本地全文：下载
作者：Rounaq Basu ; Roberto Ponce-Lopez ; Joseph Ferreira 等
期刊名称：Journal of Transport and Land Use
印刷版ISSN：1938-7849
出版年度：2021
卷号：14
期号：1
页码：303-323
DOI：10.5198/jtlu.2021.1791
语种：English
出版社：University of Minnesota * Department of Civil, Environmental, and Geo-Engineering
摘要：One of the major critiques of land use-transport interaction (LUTI) models over the ages has been their over-dependence on individualized soware and context. In an effort to address some of these concerns, this study proposes a framework to construct “virtual cities” that can act as sandboxes for testing different features of a LUTI model, as well as provide the capability to compare different LUTI models. We develop an approach to translate any prototypical transportation infrastructure network into a plausible land use zoning plan and synthetic population that are suitable for spatially detailed LUTI microsimulation of the virtual city. Disaggregate units of spatial geometry, like parcels and postcodes, are generated using geospatial techniques applied to the transportation network. Households and jobs are randomly sampled from an actual city, and allocated in the virtual city based on matching density gradients. Students are matched with schools and workers are matched with jobs to complete the calibration of a synthetic population for the virtual city. Following the adjustment of behavioral models to complement the reduced scale of the virtual city, we demonstrate the integration between the land use and transportation simulation components in our LUTI model, SimMobility. e benefits of faster convergence times and shorter simulation times are clearly demonstrated through this exercise. We hope that this study, and the open-source releases of the SimMobility soware with the virtual city database, can accelerate experimentation with LUTI models and aid the transition from individualized LUTI models to a common shared integrated urban modeling platform.