项目名称: 基于群体智能的人群疏散仿真模型及动态路径规划方法研究
项目编号: No.61472232
项目类型: 面上项目
立项/批准年度: 2015
项目学科: 自动化技术、计算机技术
项目作者: 刘弘
作者单位: 山东师范大学
项目金额: 83万元
中文摘要: 面向突发事件下大规模人群疏散仿真,结合群体智能的宏观模型与具有个体动力学特征的微观模型,提出综合两种模型优势的人群疏散仿真方法。该方法采用增加了群组吸引力的社会力模型及动态小生境分类算法,以家庭、同事、朋友等关系形成的群组为核心,将大规模人群划分为群组。按群组中个体的适应度值,选出引导者。在疏散过程中,群组内部为不区分个体差异的粒子。该模型把大规模人群疏散的问题,分解为各群组按块移动的问题。动态路径规划采用人工蜂群与微粒群结合的算法,有种群之间和种群内部两层交互机制:在选择引领蜂时,种群之间彼此交互,在全局范围内确定若干较优解,以防止算法陷入局部最优;确定目标后,引领蜂带领跟随蜂向出口移动,微粒群算法在种群内部进行,保证算法向着多个全局最优解收敛。方法综合考虑了人群宏观动力学特性和个体不确定行为的影响,弥补了宏观模型实用性不强及微观模型计算量过大的问题,使仿真更接近于人群疏散的真实场景。
中文关键词: 人群疏散;仿真模型;群体智能;路径规划;人工蜂群算法
英文摘要: For emergency evacuation simulation of large scale crowd, it combines the macro model of swarm intelligence and microscopic model with individual dynamics characteristics,and puts forward the approach of combining advantages of two models for crowd evacuation simulation method. The method adopts social force model with increased the group's attractiveness and dynamic niching classification algorithm, forms a group of family, colleagues, friends and other relations as the core, the large population are divided into groups. Then, according to the fitness values of the individuals in the group, select the guider. During the evacuation, a group is considered as not to distinguish individual differences of particles. The problem of large crowd evacuation is decomposed into several group blocks moving problem in this model. Dynamic path planning adopts artificial bee colony combined with particle swarm algorithm, and two layer interactive mechanism both between and within groups: when choosing guider bees, interaction among populations to determine the number of the optimal solution in the global scope, in order to prevent the algorithm falls into local optimum. After determining the goal, guider bees lead the onlooker bees to move to the exits, particle swarm optimization is executed within populations, to ensure that the algorithm converges toward multiple global optimal solution. The method considers the influence of population macro dynamics and uncertainty of individual behaviors, to make up for the macro model practicality is not strong and the microcosmic model with the problem of the large amount of calculation,, make the simulation more closer to the real scene in crowd evacuation.
英文关键词: Crowd evacuation;Simulation model;Swarm intelligence;Path planning;Artifical bee colony algorithm