Emergency Management Oriented Research Of Individual Behavior Modeling Of Macro Travelling And Micro Movement

A person,as an automatic individual,plays an important role in the occurrence and development of emergency.Groups composed by some individuals act as bearing body against emergency.If we want to do research of what role individual plays in the occurrence and development of the emergency and what influence the emergency will impose on individuals,it is very important to take care of modeling of individual behaviors.This paper mainly focuses on two aspects of individual behaviors: macro traveling behavior and micro movement behavior,both of which poses the power to influence the whole society.For example,macro traveling behavior will cause epidemic disseminate widely and micro movement behavior is of great significance in the process of evacuation.This paper had do some research to model both behavior in preparing for setting up relative artificial society.Our mainly work is as following:(1)The establishment of model of individual macro traveling behaviorFirstly,we designed a hierarchical mechanism to configure individual behavior for the convenience to use updated data to model individual behavior,thus promoting the validity of the model.Secondly,an algorithm for generating activity chain was proposed and was convinced to produce soundly heterogeneity among individuals.Thirdly,a message retreat method was designed based on general discrete event simulation executive to meet the needs of event retreat in areas of emergency management.(2)Artificial stationWe model the environment and normal-state crowds of Changsha railway station square based on the data we collected by field investigation.Much attention was paid to entrance modeling and a queue model was proposed to model the flow of crowds through entrances,which not only guarantee the validity of the model but also do something good for the performance of the model.As for the modeling of crowds,a method to determine the position for each individual before simulation was proposed and influence of hot district was taken into account.OD matrix was used to model the process of choosing destination for each individual and social fore was introduced to model individual locomotion.Some special techniques were used to model the group behavior composed of several individuals.An extended influence model was proposed to make path planning,which is capable of managing some dynamic obstacle in the model.(3)In the setting of sarin attack,BDI-based agents were used in the evacuation simulation.Firstly,the Gaussian puff model was used to model the dispersion of sarin gas.We analyzed quantitively how sarin influence the health of individuals and defined the notation of health index.Three kinds of model of individual sensing outer environment were proposed and the attribute anxiety was used to quantify psychological stress.And a model for changing intention based on anxiety was also proposed.Social force model was also used to model locomotion of individual in emergent situation.Memory and conformity was considered in the model of choosing exit.(4)Experiments,testing sarin releasing position,sarin releasing quantity,sarin dispersion rate,sarin releasing duration,were conducted in available artificial society.Firstly,we adopted a method to validate model from both macro and micro point of view.Then we proposed some metrics(health index value of expectation,0.95 fractile,total influenced number of individual,deadly-harmed number,dead number)to assess crowds state of health.An integrated matric were used to get a full understanding of crowds' state of health.The results showed that releasing rate and releasing position is in relationship with the harm imposed by sarin.