Activity capacity-based urban shrinkage trend prediction model and response strategy comparison approach

Many countries are facing escalating urban shrinkage, with vast swathes of urban areas becoming desolate. Urban managers urgently need strategies to mitigate land and infrastructure wastage. Although many studies have developed trend prediction models based on single-source data, these models cannot analyze the causes, evolution, and impacts of urban shrinkage using multiple data sources and residents’ behavioral insights. Urban shrinkage significantly affects activity and travel flows, if future trends in these flows can be predicted, urban managers can identify facilities likely to experience reduced flow and develop targeted responses. Traffic network capacity is instrumental in assessing the ability to accommodate travel flow, but the origin–destination (O-D) demand-oriented approach falls short in capturing the nuances of travel times, modes, and purposes from a travel motivation standpoint. It also fails to provide demand information related to activities, such as activity locations, activity times, and activity sequences. This paper introduces a novel concept: activity capacity, which provides two key pieces of information: (1) the maximum activity flows an activity-travel network can accommodate under shrinkage; (2) the corresponding distribution of activity and travel flows. We establish a bi-level programming model. The upper level, the Urban Shrinkage-oriented Activity Capacity (USAC) model, seeks to maximize activity demand within the constraints of land use, urn shrinkage, and activity demand structure. The lower level, an Activity Capacity-oriented Activity-Travel Assignment (AC-ATA) model, particularly accounts for online-activity utility and travelers’ perceptual errors regarding activity node flows. A tailored Sensitivity Analysis-Based (SAB) method is employed to solve the USAC problem. Numerical examples demonstrate the USAC model’s effectiveness in predicting activity capacity and flow distributions under urban shrinkage and in evaluating response strategies, providing planners with critical and valuable insights. Additionally, the model’s sensitivity to parameters related to online activity, land use constraints, and travel costs is analyzed.