In the early stages of major emerging infectious disease outbreaks, the transportation sector must identify bus stops with high relative accessibility to the risk area and interdict the accessible path, i.e., the shortest path, to regulate the spread of the outbreak within the transit system. Thus, this paper proposes a novel multi-sink shortest-path network interdiction model that incorporates accessibility measures in the reverse direction for the first time. The model consists of two steps: first, a detailed index of transit system accessibility is constructed; second, based on the accessibility definition, the accessibility path interdiction is formulated as a bi-level bi-objective programming problem. The upper-level planning aims to minimize the accessibility and traffic control cost of the transit system at bus stops in the epidemic risk area. In contrast, lower-level planning involves solving the shortest path search problem. An example network is applied to validate the proposed model and algorithm’s effectiveness, and the results show that the model is valid. Moreover, computational experiments evaluate the model’s performance in a large-scale network. The proposed model can compute the optimal decision in only 200 seconds in a real-case application.

Complete vehicle trajectory data is essential for urban traffic flow modeling studies. This study proposes a framework for filling vehicle trajectories in spatial and time for automatic vehicle identification (AVI) data. Based on the particle filter, the dynamic correction factor is innovatively used to improve algorithm accuracy. After four resamplings, such as traffic situation index and traffic event factor, spatial trajectory filling is completed. The Copula function fills the time trajectory by analyzing the correlation between upstream and downstream paths. Finally, the experiment was conducted in Xiaoshan District, Hangzhou, China. The results show that for spatial trajectory filling, the average accuracy exceeds 97% with 75% camera coverage. In time trajectory filling, the time trajectory filling error is reduced by 35% compared to the Hellinga algorithm.