考虑城市建成环境非线性作用的通勤时间模型

Residents’ commuting time model under the nonlinear impact of urban built environment

为探索城市建成环境对居民通勤时间的非线性作用,以南京为例,构建了包含个人家庭属性、通勤出行特征和交通小区建成环境的梯度提升决策树模型。结果显示:三类属性中,居民通勤出行特征对通勤时间的累计影响程度最大(52%),建成环境影响次之(38.88%);建成环境指标对通勤时间的非线性作用显著,其中,区域用地混合度和居住用地占比都显现出显著的阈值效应。研究可为后续精准化城市区域规划和设计提供科学的理论与依据。

To investigate the nonlinear impact of built environment on residents’ commuting time,a case study was conducted in Nanjing by using a Gradient Boosting Decision Tree. The model has considered the impact of socio-demographics,trip characteristics,and built environment at traffic analysis zone. The result shows that trip characteristics have the largest cumulative contribution on commuting time as high as52%,followed by built environment factors(38.88%). Significant nonlinear impacts of built environment have been presented. The nonlinear thresholds of land-use mix and residential ratio are identified. This study expects to provide supportive scientific references for urban future planning and design at the microlevel.