CASIA OpenIR  > 毕业生  > 博士学位论文
基于Artience的人工交通系统研究与设计
其他题名Design of Artificial Transportation System based on Artience
缪青海
学位类型工学博士
导师王飞跃
2007-06-07
学位授予单位中国科学院研究生院
学位授予地点中国科学院自动化研究所
学位专业控制理论与控制工程
关键词复杂系统 人工社会 对等计算 多agent系统 人工交通系统 Complex System Artificial Society P2p Computing Multi-agent System Artificial Transportation System
摘要复杂系统与复杂性科学以其整体论的指导思想和多学科交叉的研究方法,被称作21世纪的科学;而城市交通系统涉及社会生活的各个方面,是一个有规则约束的天然的开放复杂巨系统。因此,运用复杂系统的理论与方法研究城市交通问题具有重要的科学价值,这也正是本论文研究的内容。本论文的工作主要包括以下几方面: 首先,总结了当前复杂系统研究的理论与方法存在的问题,引入了Artience理论与方法体系,并分别从人工社会、计算实验和平行系统三个方面对Artience的原理做了详细介绍。 其次,针对传统交通需求预测存在的不足之处提出了人工人口模型。在宏观层面上利用社会学阶层分析的研究成果,给出了社会人口各阶层城市空间分异和出行方式分类;在微观层面上运用马斯洛层次需求动机原理,建立了描述个人活动状态转移概率的模糊规则,提出了人工人口模型的运行机制。 在人工交通系统实现过程中,结合JXTA的框架结构、管理机制和实际交通系统的特点,设计了基于P2P网络的人工交通系统分布式层次化体系结构;利用JXTA的Socket对等通讯功能,设计了各个Agent之间的通讯与交互机制。 仿真试验部分,在所开发的人工交通系统平台上,提出了一系列实验方案,分别验证了模型设计的合理性、网络通讯的有效性和综合大区域仿真的可行性。
其他摘要The complex systems and complexity science featured by holism and multidiscipline, becomes a hotspot that both its research and applications are developing widely throughout the world. It is called the science of 21st century. On the other side, the city transportation system is an inherent open giant complex system involving nearly all aspects of our society, which is a great challenge to traditional methodology. For this situation, it is valuable to apply theories and methods of complex systems and complexity science to study problems in city transportation system. And this is what this thesis focuses on. In this thesis, artificial population model and traveller model are proposed based on Artience theories and methods. Following in time, by integrating such models systematically through a P2P computing platform, the Artificial Transportation System (ATS) is implemented and some related experiments are carried out. Main works are as follows: Firstly, the progress of complex systems and complexity science is chiefly summarized, and the existing problems are addressed. Consequently, a new methodology - Artience is introduced and interpreted in three aspects: artificial society, computational experiments and parallel systems respectively. Secondly, in order to address the problems that are out of the ability of traditional four-stage prediction method, an artificial population model is proposed. This model can be interpreted at two levels. On the macroscopic view, by using social stratification theory in sociology, the daily activities, spacial distribution and trip modes of different population are analysized. On microscopic view, by utilizing Maslow's hierarchy of needs in Psychology together with Markov Chain in stochastic process, fuzzy rules that discribe the person activity transition probability is designed. Mechanisms that how this model takes a role in artificial transportation system is also proposed. At the same, the infrastructure models, traveller models are introduced by laying emphasis on there driving logic and behavior choice mechanism. Lastly, at the beginning of ATS implementation, the developments and applications of P2P computing are introduced with an emphasis on principles and features of JXTA. By taking advantages of JXTA's management mechanism, a distributed and erarchical structure of ATS is proposed; communication and interaction mechanism between Agents is also designed based on JxtaSocket technology. On the basis of this implemented ATS, three gradual simulation experiments are carried out to testify the rationality of the models, the validity of network communications and the feasibility of traffic simulation in large areas.
馆藏号XWLW1093
其他标识符200418014628061
语种中文
文献类型学位论文
条目标识符http://ir.ia.ac.cn/handle/173211/6006
专题毕业生_博士学位论文
推荐引用方式
GB/T 7714
缪青海. 基于Artience的人工交通系统研究与设计[D]. 中国科学院自动化研究所. 中国科学院研究生院,2007.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
CASIA_20041801462806(2080KB) 暂不开放CC BY-NC-SA请求全文
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[缪青海]的文章
百度学术
百度学术中相似的文章
[缪青海]的文章
必应学术
必应学术中相似的文章
[缪青海]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

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。