Towards Real-time Emergency Evacuation By Smart Phones With Dynamic Game Theory

The emergency under the condition of throng is usually accompanied by extreme crowding and chaos,panic mood,highly dense crowd and a large amount of uncertain event information in the surrounding environment,which may make the crowd rush to a certain established safety exit and lead to serious public safety accidents.How to evacuate the crowd in the crowded area quickly and safely after the public emergency has become a hot research topic in various fields.However,in the absence of timely and unobstructed information transmission channels,the chaotic environment under the condition of large amount of crowd gathering poses challenges to the safety evacuation technology and becomes a long-term bottleneck.The popularity of mobile Internet has made up for the shortcoming of this problem.With the popularity of smart phones and other electronic devices,4G cellular network has been mature and even 5G mobile network has come into practical application,providing a new channel for information transmission.People can sense the location information through intelligent mobile location services in the environment,which can be collected by public institutions too.These congestion information will be used to judge the situation of target site and cooperate with the user’s smartphone,helping each user to calculate the best evacuation route guidance and pass personal information to the users in real time.This way should be expected to become new evacuation guidance technology.However,the complex interaction between a large number of individual users’ behaviors and the complexity of real-time evacuation path algorithm under dynamic conditions bring challenges to the path evacuation guidance technology of intelligent devices.In order to solve this problem,combined with the relevant ideas of game theory,this paper studies from three aspects: crowd interaction movement analysis under congestion environment,evacuation path selection game and equilibrium,and user evacuation system simulation based on smart phones,The specific work and contribution are as follows:(1)Analysis of crowd movement in a congested environment.Crowd congestion models are analyzed and studied to ensure that the anisotropic interactions in the model are handled in an internally consistent manner.In this paper we focus on two classes of models,one based on inter-crowd interactions(the speed of the pedestrian depends on the angle between the two directions of the crowd)and another based on intra-crowd interactions(the speed of the pedestrian depends on the density of the crowd).For both cases,this paper proves sufficient conditions under which the optimal controls are determined uniquely almost everywhere in the domain.Under this condition,the direction of motion can be uniquely determined at almost all positions in the domain.For the inter-crowd models,this paper also proves sufficient conditions under which the Nash Equilibrium is unique almost everywhere in the domain.(2)Game selection of evacuation route.When deciding how to choose the evacuation route,it should be considered that users want to get the least time.In this paper,crowdsourcing technology is applied to help users choose the evacuation path,so that all users can reach the exit in the shortest time.In this paper,the problem is modeled as an evacuation route selection(EPS)game,which is proved to be equivalent to a symmetric network congestion game.On this basis,an evacuation route selection game algorithm is designed,and at least one Nash equilibrium is found in the distributed system.(3)An evacuation path selection guidance system containing Android application and server is designed,and the simulation experiment of distributed path selection game algorithm is carried out.At the same time,a framework of emergency evacuation guidance system is designed,and a simulation platform of emergency evacuation system is built.The algorithm proposed in this paper is verified by comparing with the traditional evacuation path selection algorithm.The experimental results show that the emergency evacuation guidance algorithm designed in this paper can effectively reduce the overall evacuation time of all users and achieve a good effect.