Urban Activity Models Based on Spatiotemporal Inference
CHEN Xianlong,ZHANG Hua,CHEN Xiaohong,ZHANG Tianran,MA Yilin
DOI: https://doi.org/10.13813/j.cn11-5141/u.2023.0107
2023-01-01
Abstract:Based on reviewing the limitations of existing transportation models, such as over-dependence on mathematical optimization solutions and insufficient analysis of actual travel activities, as well as the lack of connection between current and future modeling situations, this paper proposes model improvements using big data methods. The paper presents an extended demographic spatiotemporal inference model, including a base-year population synthesis module, a demographic development module, and a job-housing transition module, which achieves the spatiotemporal continuation of individual attributes and the result of travel demand associated with urban characteristics. The activity simulator designed in the paper includes multiple sub-modules, such as travel chain generation, destination selection, individual activity organization, and family member activity coordination, to simulate travel activities based on individual attributes and urban activity patterns. Empirical studies demonstrate that introducing individual attributes and urban activity patterns based on big data enhances the casual inference ability of travel activities, avoids large-scale computation of mathematical optimization solutions, and generates consistent modeling results in line with urban realities. This approach also achieves the inheritance of the present in the future and reflects the continuity of urban development.