应对突发大规模流行病的城市常规公交管控策略

The relatively small and confined space inside public transportation vehicles, which induces frequent and close contact between passengers, is a high-risk environment for the propagation of infectious disease. In the face of a sudden, large-scale epidemic such as the COVID-19, the strategy usually adopted by traffic managers is to shut down all bus lines, thus completely blocking disease transmission through the public transport system. However, for bus passengers, the lack of alternative means of public transport seriously reduces their mobility and even affects their basic living conditions. Here, we have explored the characteristics of individual passenger travel and contact among passengers, and have reconstructed dynamic contact networks based on data from bus IC cards in the city of Beijing. Experimental data show that the average contact time between any pair of bus passengers on a single trip is 17 min, and the average cumulative contact times between one commuter and other passengers during weekdays is 123 We used the Susceptible-Exposed-Infected (SEI) and Susceptible-Infected-Susceptible (SIS) models to simulate the dynamics of disease spread on such temporal networks, which led to the discovery of an efficient control strategy. Compared with a simulation on random networks, we found that the periodic occurrence of contact between bus passengers significantly promoted the rapid spread of disease. For conventional bus networks, we designed a partial shutdown scheme by aggregating the individual transmission influence into the bus lines, and used the average transfer distance between bus stations as an indicator of the reduced accessibility of the conventional bus system due to partial shutdown. The scale of the epidemic before and after the implementation of the partial shutdown scheme were compared. We found that disease outbreaks could be contained by stopping a small number of bus lines, and that the overall accessibility of the bus transportation system did not decrease significantly.