Robust dynamic distribution of security assets in transit systems

Abstract

A robust, mixed-integer, multistage program is presented: it seeks to secure a transit system in which risk is considered to be dynamic and varies over time. A time-varying risk measure reflects the unique nature of transit systems: accumulation of passengers at transfer facilities, stations, and transit vehicles is dynamic and increases the vulnerability of the transit users and system to adverse events. The model is robust under uncertainty and matches security assets at stations better in the face of time-varying risk by redistributing them. The volume-dependent risk measure and subsequent deployment of security assets were developed for the transit system in Washington, D. C., to demonstrate the variable nature of risk and response. The value of considering a robust solution was demonstrated by a comparison of the strategies developed from a robust approach with those from an expected value approach. Five scenarios, designed on recent events on the system, replicate the operational conditions of the transit system for the morning peak hour period and show the effectiveness of the developed deployment strategies.