Evacuation Strategy Optimization Study Based on System Theory

Emergency evacuation is to transfer people from dangerous places to safe areas, so as to reduce or even avoid the potential harm to people. It is inherently a comprehensive system composed of evacuation managers, evacuees, road networks, shelters, etc. Security is one of the important indicators of such system. Moreover, in order to ensure the normal and efficient operation of evacuation system, each component should cooperate well with each other, thus making stability another important index of the evacuation system. In order to optimize evacuation safety, some residential areas may be arranged to stay much longer which is hard to be accepted, namely, the stability of evacuation system is low. In this paper, a system-based evacuation CSO model at residential level is proposed which compromises the security and stability of evacuation systems. The CSO model is a bi-level network optimization model, the upper level aims at minimizing the total risk of evacuation subject to the residential tolerance level and the lower level conveys a cell transmission-based dynamic traffic assignment problem. Using our model, we also study the impact of the number of shelters, the organizational form of road intersections, the uncertainty of evacuation demand and risk distribution on evacuation system.