【作者】 叶华；

【导师】 李彬；

【作者基本信息】 武汉理工大学 ， 控制理论与控制工程， 2006， 硕士

【摘要】 近年来,随着我国信息化建设和人工智能研究的快速发展,知识的自动获取己成为制约其发展的“瓶颈”。数据挖掘是人工智能与数据库技术结合产生的新技术,研究如何从大量的数据中智能地、自动地抽取出有价值的知识和信息,能在一定程度上解决知识的自动获取问题。由于智能交通系统的交通流数据采集系统已经积累了大量交通流数据,利用数据挖掘技术,可以使这些数据不但能够为智能交通系统中的控制系统服务,也能为决策系统和诱导系统等提供支持。本文提出用基于泛布尔代数的数据挖掘技术来作为智能交通系统的数据分析和数据解释方法,以解决其建模和模型求解的问题。 本文首先对数据挖掘技术及其发展现状进行了简单的介绍,并针对其研究现状、面临的挑战以及应用中的总体需求,提出一种新的数据挖掘方法,即,对泛布尔代数应用于数据挖掘进行可行性研究。 对泛布尔代数进行总体介绍,在分析和研究它的理论体系后,概括出:它的特性决定了它能在一定程度上解决数据挖掘的现存问题,比如可以保证数据挖掘结果的可用性、确定性及可解释性,可以解决复杂概念表达困难、属性间的相互关系强调不够、重复检验等问题;能就某一系统建立统一的泛布尔模型;并促进新型决策支持系统的开发研究,并对泛布尔代数与数据挖掘相结合的原理展开论述。 本文重点介绍如何将结合泛布尔代数的数据挖掘技术应用到智能交通系统中。文中针对城市单个交叉口的交通流,利用这种新的数据挖掘技术的初步分析结果,选择采用多相位实时控制策略;对提取的决策规则,建立起决策系统的泛布尔模型;仿真分析作出的评估优化表明,通过这种新的数据挖掘技术得出的控制模型,对实时改善交通流的管理和控制,效果优于以往的交通流控制技术。 在应用这种新的数据挖掘技术进行反复的挖掘过程后,将数据挖掘的结果应用到模拟的交通信号控制决策系统,并在CPLD上实现了本文提出的单交叉口多相位实时控制模型的设计。更多还原

【Abstract】 In recent years, with the fast development of information construction and the artificial intelligence studying in our country, to obtain knowledge automatically have become the "bottleneck "of its development. Data Mining is a new technology that combines Artificial Intelligence with database technology, deal with how to obtain valuable knowledge and message among mass data intelligently and automatically, can solve the problem of obtaining knowledge automatically to a certain extent. For the data collecting system of the intellectual traffic system has already accumulated a large number of traffic flow data, utilize Data Mining technology, can make these data not only be the service of the system of control in intellectual traffic systems, also offer support for the decision system and system of inducing etc. This text proposes utilizing a new Data Mining technology based on Fan Boolean Algebra as the way of explaining the data and analyses data in the intellectual traffic system.This article has simply introduced the Data Mining technology and its development at first, and put forward the new feasible method to its current situation and development prospect and the wholly requirement in use. namely explored the feasibility of combining Fan Boolean Algebra with the Data Mining technology.Introduced Fan Boolean Algebra theory as a whole, after analyzing and studying of it, concluded that: the characteristic of Fan Boolean Algebra has determined that it can solve the extant problem of Data Mining to a certain extent, for example, it can guarantee the result after data mining to be usable , assured and construable;can solve the problems about expression difficulty of complicated concept, correlation of attributes emphasized incompletely and redundant examining;it can set up the unified model of Fan Boolean in a certain system;and it can promote the developmental research of new decision support system.This text introduced how to apply the data mining technology combining with Fan Boolean Algebra to the intellectual traffic system especially. In the article aim against the traffic flow of the single crossing of the city, utilize the preliminary analysis result of such new data mining technology;choose to adopt the real-time control strategy of multiphase;to the decision rule that is drawn, set up systematical decision model of Fan Boolean Algebra. The assessment and optimizing made in stimulation analysis indicated: such new data mining technology improves the management and controlling in real time, superior to the previous controlling technology of traffic flow.In the end, applied this new Data Mining technology to obtain the knowledge time and again, utilized the result of the data mining to the mimic decision system of controlling traffic signal, and eventually accomplished the designing of multiphase real-time controlling modeling of single interaction which proposed by this article in CPLD.更多还原