Selecting Emergency Evacuation Routes in Urban Dynamic Traffic Network

Abstract

We examined the selection of emergency evacuation routes in an urban traffic network with dynamic information. A method of selecting emergency evacuation routes was proposed based on the uncertainty characteristics of urban traffic networks and the demand of urban disaster emergency management for an evacuation route selection and given the developments in advanced traveler information subsystems. The approach involves planning evacuation route schemes in an offline module and updating them in an online module with real-time traffic information. In the offline module, a model of evacuation route planning based on time-dependent networks was established by using prior knowledge of traffic network statistics. A synthetic algorithm using the successive shortest path flow and shortest path in the non-first-in-first-out time-dependent network was adopted in the concrete optimization course to obtain a solution. In the online module, an emergency evacuation route model based on real-time traffic information was established with forecast road impedance by nonparametric regression. The Floyd algorithm was adopted in the concrete optimization course, and an updated scheme of evacuation routes was implemented. The results, which were analyzed by simulation, indicate that the principle and method are correct and feasible.