Assessing Survivability of the Beijing Subway System

The assessment of survivability is a common topic in critical network infrastructure research. In order to examine the critical components whose disruptions can cause huge system degradation, many measures have been approached to depict the characteristics of network systems. Serving more than ten million passengers a day, the Beijing subway system, which ranks third in the world for its length and annual ridership, raises survivability issues in the face of potential disruptions of network components along with its constantly increasing complexity. In this research, we provide an accessibility-based survivability measure with which to explore how potential outages of network components might affect the overall functionality of the Beijing subway system. System survivability is measured from two perspectives: [1] connectivity under various simulated failures of stations and [2] variations in passenger flows in response to a disruptive influence. Plausible scenarios are constructed using local demographic data and daily ridership reports from subway management companies. To assess the possible range of influences, we develop a weighted rank-based simulation algorithm to approximate the extreme combinatorial disruption instances. The range of the potential effect highlights the best and worst-case scenarios so as to identify the critical components and help to prepare corresponding contingency plans. This research will enable the more legitimate allocation of limited emergency response resources and highlight the way of improving the survivability of the system.