Research On Evacuation Model Based On Mutil-Agent

In recent years,due to lack of effective evacuation measures for person in the dense places,a lot of tragedies has been occurred which posing a serious threat to the safety of the people.Therefore,it is important that how to evacuate all the personnel of the gathering place safely and effectively in an emergency situation.As the traditional evacuation exercises not only consume a lot of manpower and resources,but also exist some kinds of hidden dangers.So,it has become the focus of academic attention through a variety of means to simulate and study the various ways of the entire evacuation process in order to develop contingency plans for emergency situations.In view of this problem,according to the existing research and simulation of the evacuation model,combined with the characteristics of the evacuation process,the model which based on Multi-Agent and makes an intensive study of the variety of technologies which the model involved is proposed,it can be simulate evacuation process more realistic.Firstly,this paper studies the concept of Agent,Multi-Agent and the modeling method based on Multi-Agent.The paper draws up the idea of the bottom-up,divides the evacuation model,establishes the physical scene and carries on the geometric abstraction as the continuous space scene;on account of the reason above,this paper establishes the perception model based on the Multi-Agent to simulate the interaction between the individual and the environment;based on the model above,this paper not only incorporates the individual's path decision strategy,collision resolution strategy and panic degree,but also proposes the evacuation model based on Multi-Agent's perceptual model as the framework of the whole evacuation model.Secondly,this paper optimizes the evacuation model by improving the particle swarm algorithm and social force model,taking into account the individual panic factors.By introducing the particle swarm algorithm as the path planning strategy of the individual,the particle swarm algorithm is improved and verified which is closer to the actual situation of evacuation;this paper introduce the social force model and modified it as the individual collision solution strategy in order to simulate the interaction between individuals,individuals and obstacles;this paper shows the impact of one of the emotion of the individual by introducing the mathematical expression of panic degree to make the model close to the reality.Finally,based on the previous research,this paper designs and implements a simulation system based on Multi-Agent evacuation model.The paper testifies the model is more close to reality which investigated by the multiple experiments and the reality scence which means it can really simulate the evacuation process.