基于蚁群算法的人群疏散仿真研究

发布时间：2018-06-26 01:40

本文选题：人群疏散仿真 + 蚁群算法　；参考：《山东师范大学》2017年硕士论文

【摘要】：近几年随着我国软实力的增强,经济、文化、娱乐业大力发展,人口规模不断扩大,各种高楼大厦、大型体育场等比比皆是。这些人流密集的场所其实存在着安全隐患,在紧急情况发生之前能科学地预防,在紧急情况发生时能正确地疏导人群,是现在相关部门问题解决的根本。因此这类场所必须随时做好充分的应急疏散准备,防止当遇到这种危险情况时因为逃离现场方法不当或由于时间的延误等原因导致拥挤、踩踏等现象造成人员的伤亡。研究人群的逃生规律,可高效地提高人们在遇到紧急危险情况时的自救能力。针对现在的疏散仿真技术中存在的人群行为单一,模拟精度不高等问题,社会力蚁群模型在本文中被提出,该方法一方面控制了成本的输出,还提高了模型精度,对大型建筑内法的人群疏散有很好的指导作用。本文采用基于粒子群算法和蚁群算法作为微观模型指导微观路径规划,用社会力和蚁群算法结合的模型指导人群疏散过程中的个体行为,通过大量仿真实验的提出与实现,证明了其有效性。三维动画基于图形学蓬勃发展的大背景下在各个领域都占据着不可替代的位置,比如电视广告,游戏研发,建筑物模型的构建等等。本文基于拥有强大功能的Maya软件对建筑物和人物模型进行创建,利用Mel语言的优势,效果非常清晰逼真,同时缩短了研发的时间,提高了工作效率。本文主要工作及创新点概括如下:1.提出一种新的人群疏散模型,该模型将社会力模型与蚁群算法相结合,用社会力作为蚁群算法中信息素更新策略的度量值。建立社会力蚁群模型,该模型充分考虑了人群疏散中个体间的作用力关系,解决了算法中易于出现的停滞和早熟现象。该模型可以有效发挥蚁群算法在人群模拟中的优势,较传统算法具有更高的效率和运算速度,对大规模建筑内人群疏散有重要参考价值。2.提出一种路径规划方法,它是基于拓扑图的,将粒子群算法和蚁群算法相结合的粒子群蚁群算法(PSACO)。这种算法先用PSO算法生成一条初始路径,然后把这条初始路径转化为蚁群算法(ACO)的初始信息素分布,这样做很大程度上克服蚁群盲目搜素的缺点,在时间效率上起到了很好的效果。最后仿真实验中验证,基于拓扑图的这种结合算法在实际场景系统中的确比其他算法在路径规划上取得了小小进步,路径规划的性能被大大提高,为今后的工作奠定了重要的基础。3.基于Maya这一强大场景制作软件,结合Mel语言创建真实性的建筑物模型和人物模型。然后应用于实验室项目之中,以VS2012+OSG2.3.1为平台,生成仿真动画,渲染精美效果。结果显示本文所述方法实现了高质量的疏散仿真现象,效果非常清晰逼真,同时缩短了研发的时间,提高了工作效率,也为以后的疏散研究提供了思路,具有一定的使用价值。
[Abstract]:In recent years, with the strengthening of soft power, the development of economy, culture, entertainment industry, the scale of population is expanding, various tall buildings, large stadiums and so on are everywhere. In fact, there are hidden safety problems in these crowded places, which can be prevented scientifically before an emergency occurs, and can correctly divert the crowd when an emergency occurs, which is the fundamental to solve the problems of relevant departments now. Therefore, such places must be prepared for emergency evacuation at any time, so as to prevent people from being injured or injured due to improper methods of escaping from the scene or delays in time, such as stampede and so on. The study of crowd escape law can effectively improve the ability of self-rescue in emergency situations. In order to solve the problems of single crowd behavior and low simulation precision in the current evacuation simulation technology, the social force ant colony model is proposed in this paper. On the one hand, the method not only controls the output of cost, but also improves the precision of the model. It is a good guide to the evacuation of large buildings. In this paper, particle swarm optimization (PSO) and ant colony algorithm (ACA) are used as microscopic models to guide micro-path planning, and social forces and ant colony algorithm (ACA) are used to guide individual behavior in evacuation process. Its validity is proved. 3D animation occupies an irreplaceable position in all fields under the background of the vigorous development of graphics, such as TV advertising, game development, building models and so on. Based on Maya software which has powerful functions, this paper builds buildings and human models, makes use of the advantages of Mel language, the effect is very clear and lifelike, at the same time, it shortens the time of research and development, and improves the working efficiency. The main work and innovation of this paper are summarized as follows: 1. A new crowd evacuation model is proposed, which combines social force model with ant colony algorithm and uses social force as a measure of pheromone updating strategy in ant colony algorithm. A social force ant colony model is established, which fully takes into account the interaction between individuals in crowd evacuation, and solves the stagnation and precocity phenomena that are easy to occur in the algorithm. This model can give full play to the advantage of ant colony algorithm in crowd simulation, and has higher efficiency and speed than traditional algorithm. It has important reference value for evacuation of large scale buildings. This paper presents a path planning method, which is based on topological graph and combines particle swarm optimization algorithm with ant colony algorithm, particle swarm ant colony algorithm (PSACO). This algorithm uses PSO algorithm to generate an initial path, and then converts the initial path into the initial pheromone distribution of ant colony algorithm (ACO). In time efficiency has played a very good effect. Finally, the simulation results show that the algorithm based on topology graph has made a little progress in path planning, and the performance of path planning has been greatly improved. For the future work laid an important foundation. Based on Maya, a powerful scene making software, a real building model and a character model are created with Mel language. Then applied to the lab project, VS2012 OSG 2.3.1 as the platform, generate simulation animation, render beautiful effects. The results show that the method presented in this paper realizes the high quality evacuation simulation, and the effect is very clear and lifelike. At the same time, the research and development time is shortened, the work efficiency is improved, and the thought is provided for the later evacuation research, which has certain practical value.
【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：C912.6;TP18

【参考文献】