QQ客服 官方微博 反馈留言

题名	大型公共环境人员疏散路径规划研究
作者	汪婕
学位类别	博士
答辩日期	2009-01-14
授予单位	中国科学院软件研究所
授予地点	软件研究所
关键词	人员疏散 路径规划 分布估计算法 贝叶斯网络
其他题名	Study on Path Planning of Crowd Evacuation from Large Public Environment
中文摘要	大型公共环境人员疏散路径规划研究对指导安全设施建设、应急预案的制定、实施和推演都有重要意义。近年来进化计算领域兴起的分布估计算法为路径规划问题的解决提供了新的优化工具。本论文将一种典型的分布估计算法——贝叶斯优化算法，应用到大型公共环境人员疏散路径规划问题解决过程中，针对单路径人员疏散、多路径人员疏散和多目标点人员疏散三类具体的疏散路径规划问题，分别设计和实现了相应的基于贝叶斯优化的人员疏散路径规划算法，并在仿真实验中根据大型公共环境空间类型复杂和人员相对密集等特点，面向不同的规划要求和目的，求解出了满足约束条件的最优疏散路径集合。 本文提出的基于贝叶斯优化的人员疏散路径规划算法，以大型公共环境的基本信息已知为前提，基于二维环境下的拓扑空间法，按照应用系统的规划要求，通过引入疏散性能，路径安全性和易通性等参数，对疏散空间建立全局环境模型；在系统分析分布估计算法基本框架和特点的基础上，引入其中一类典型算法——贝叶斯优化算法，设计出一系列适于不同疏散路径规划问题的基于贝叶斯优化的路径规划算法。该算法用贝叶斯网络对优选路径集合建立概率模型，并由建立起的贝叶斯网络产生新的路径集合，与原有路径进行竞争，可以在找到一条最优的路径集合的同时，利用中间结果，快速找出多条较优的路径。仿真实验标明，该算法可以有效的满足大型公共环境下人员疏散的任务要求。
英文摘要	Study on path planning of crowd evacuation from large public environment is very important for the instructions of both safety facilities construction and the establishment, implement, processing of emergency prediction. In recent years, estimation of distribution algorithm, which is originated from the field of Evolutionary Computation, provides a new optimization tool for path planning. In this paper, Bayesian optimization algorithm, a typical estimation of distribution algorithm, is applied into the solution of path planning of crowd evacuation from large public environment. We design and implement corresponding path planning algorithms separately based on Bayesian optimization against three specific evacuation path planning problems: single-path crowd evacuation, multi-path crowd evacuation and multi-object crowd evacuation. Furthermore, in our simulation experiments, we give optimal evacuation paths of constraint satisfaction under different planning requirements and purposes, according to the characteristics such as complex types of large public environment, concentration of crowds, etc. Supposed that the information of the large public environment is known, and according to the objective of path planning of crowd evacuation, the global environmental model for the evacuation space, based on the two dimensional topological space, is established with the introduction of evacuation efficiency, safety and possibility of the path in our algorithm. After systematic analysis to the framework and the features of estimation of distribution algorithms, a series of path planning algorithms of crowd evacuation, based on the idea of Bayesian optimization algorithm, are proposed for different evacuation path planning problems. Bayesian network is selected to model the probability distribution of the optimal path sets, and to generate new path sets, which compete with original path sets. It can present several additional optimal paths quickly, with the help of intermediate result generated in query of the most optimal path. Simulation experiments show that our algorithm effectively meets the requirements of crowd evacuation from large public environment.
语种	中文
公开日期	2011-03-17
页码	69
内容类型	学位论文
源URL	[http://124.16.136.157/handle/311060/7274]
专题	软件研究所_综合信息系统技术国家级重点实验室 _学位论文
推荐引用方式 GB/T 7714	汪婕. 大型公共环境人员疏散路径规划研究[D]. 软件研究所. 中国科学院软件研究所. 2009.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。