文章编号： 1672-5328（2025）06-0071-09

孟丹青1，戴骏晨2，凌小静2
(1. 南京地铁建设有限责任公司，江苏南京210008；2.中咨城建设计有限公司江苏分公司，江苏南京210012)

摘要： 区域交通需求预测获得的通常是区县级以上的OD，而进行交通分配需要更细颗粒度的数 据，因此提出OD细分模型以解决已知大区OD求细分OD的问题。以南京都市圈区域出行为例，基 于手机信令等数据构建普通回归模型、混合模型、一体化标定模型、一体化标定混合模型，并对模 型效果进行对比。结果表明：已知大区OD求细分小区OD可采用指数函数表征小区间出行量，在 进行大区间出行量限制时转换为基于随机效用理论的离散选择模型形式；进行一体化标定可最大限 度提升模型解释能力，一体化标定模型的效果优于其他模型，且考虑混合效应能进一步提升效果。 此外，区域出行影响因素按影响程度从大到小分别为：出行成本、就业岗位规模、居住人口规模、 旅游类POI 数量和用地混合度，但中心地区吸引作用可能无法单独通过社会经济指标完全体现，考 虑混合效应更能反映现实。

关键词： 区域出行；出行需求；细分OD；出行分布；随机效用；混合模型

中图分类号： U491.1+4

文献标识码：A

OD Disaggregation Technologies for Regional Travel Demand Forecasting

MENG Danqing1, DAI Junchen2, LING Xiaojing2
(1. Nanjing Metro Construction Co., Ltd., Nanjing Jiangsu 210008, China; 2. CIECC Urban Construction Design Co., Ltd., Nanjing Jiangsu 210012, China)

Abstract： Regional travel demand forecasting typically yields origin-destination (OD) data at a coarse spatial resolution, such as the district or county level. However, finer-grained OD matrices are required for detailed traffic assignment. To address this gap, this paper proposes an OD disaggregation modeling approach to derive disaggregated OD matrices from known large-zone OD data. Taking regional travel within the Nanjing Metropolitan Area as a case study, the paper develops four models based on mobile signaling data, including a conventional regression model, a mixed model, an integrated calibration model, and an integrated calibration mixed model, and then compares the model outcomes. The results indicate that disaggregating large-zone OD flows into small-zone OD flows can be effectively represented using an exponential function to characterize the volumes of travel volumes between small zones, which can be transformed into a discrete choice model grounded in random utility theory when imposing constraints on travel volumes between large zones. Integrated calibration substantially enhances the explanatory power of the models, with the integrated calibration model outperforming the other alternatives, and the incorporation of mixed effects further improving model performance. In addition, the key factors influencing regional travel demand, ranked in descending order of impact, are travel cost, employment scale, residential population size, the number of tourism-related points of interest (POIs), and land-use mix. Notably, the attractiveness of central areas may not be fully captured by socioeconomic indicators alone; incorporating mixed effects provides a more realistic representation.

Keywords： regional travel; travel demand; disaggregated OD; trip distribution; random utility; mixed model