【作者】 薛萍；

【导师】 金鸿章；

【作者基本信息】 哈尔滨工程大学 ， 控制理论与控制工程， 2008， 博士

【摘要】 近年来,随着通信网络规模迅速扩大,网络新业务的不断出现、网络拓扑结构的日益复杂化,系统环境的不确定性的增加,网络通信系统的信息脆性问题越来越引起相关专家学者的关注。本文以复杂系统作为研究对象,提出脆性同非线性、层次性、涌现性一样,是复杂系统的本质属性,其具体表现为系统在某种条件下的突然崩溃,把系统运行过过程中这种突然崩溃的风险称为脆性风险。本文将复杂系统这种脆性风险分析方法应用到网络通信系统中,对网络通信系统进行脆性风险研究,得出网络通信系统脆性风险分析模型。首先,在已建立的复杂系统脆性定义和脆性特点的基础上,将其引申到网络通信系统脆性分析中。复杂系统脆性特点是重点强调崩溃的连锁效应,即由于复杂系统中的某一子系统的崩溃,引发的连锁的崩溃效应,会使整个复杂系统崩溃。由此,以复杂系统脆性定义为基础,对网络通信系统进行脆性分析,根据信息论的信息率失真和信息传输理论可知信息率失真函数以及热力学中熵增大定理有着巧妙联系,在通信系统中建立了脆性测度函数的概念,脆性测度函数反映了信息在通信前后脆性变化情况,对导致通信系统崩溃起着重要作用。其次,复杂系统中信息脆性的度量既和系统自身相关,也和系统的外部环境相关。由此从复杂系统脆性风险的内因和外因两方面进行分析,建立复杂系统信息脆性风险模型,且定义和构造了脆性风险熵的概念。脆性风险熵可以衡量系统脆性风险的总体不确定性。复杂系统脆性问题的最大难点在于难以定量研究,本文从传统熵理论出发,分析脆性风险熵的一系列性质,为进一步定量研究复杂系统脆性问题打下基础。再次,在网络通信系统的脆性风险分析中由于系统风险的复杂性,存在着诸多不确定的影响因素,建立了基于多级模糊综合评判法的网络通信系统脆性分析模型。用Delphi法辨识出影响通信网络脆性的脆性因子,借助于层次分析法确定网络层次结构中脆性风险影响因素的崩溃系数,同时采用模糊综合评判的方法实现了系统信息脆性风险评价。最后,结合系统信息剩余度和信息极限熵的概念,在整个连续时间域上对一个具体通信网络脆性风险过程进行分析。分析结果指出影响通信系统脆性风险大小既和系统脆性状态(环境)关系紧密,又和系统脆性结构关系有关,从而找出相应的策略降低系统脆性风险。结果说明该方法不仅可以量化通信系统的脆性风险,而且为网络通信系统的脆性评估提供了可行性的分析方法。更多还原

【Abstract】 In recent years, with the rapid expansion of communication network, more and more new network services are provided, the network topology is complicated day by day and the uncertainties of system environment are increasing. The information brittleness of network communication system is being paid more attention to by the experts.Based on complex system, the dissertation present that, brittleness is an attribute of complex system, just like non-linearity layer and emitting of complex system. It can be showed by that the system may collapse suddenly, and we name the collapse of a running system brittleness risk. Brittleness risk model of communication system has been carried out, which analysis method of brittleness risk of complex system is applied for the network communication system,First, the definition and characters of complex system brittleness is studied. The definition of brittleness is expanded to the field of network communication system characters. Complex system brittleness emphasize on the chain affection of collapse, which means if one of the sub-system of the network communication system collapsed, the whole complex system will collapse like domino. Based on the definition of complex brittleness, network communication system brittleness is discussed. According to this, the theory framework of study on network communication system brittleness is established. From information ratio distortion and information transmission theory, of the information theory it can be found that there are ingenious relationships between information ratio distortion function and entropy accretion theorem of energetic. The definition of brittleness measurement function, which can express the variation of brittleness from beginning of the communicating of information to end, is established in the communication system, and the brittleness has an important effect in the collapse of network communication system.Next, measurement of information brittleness in complex system is related with not only the system itself but also the environment outer it. Analyzing from the internal and external cause, the model of complex system brittleness risk is established, and the concept of brittleness risk entropy is defined. Brittleness risk entropy can be used for measuring the uncertainty of the whole system brittleness risk. The most difficult question is to quantitatively determine the components of complex system. In this paper, beginning from the classical entropy theory, a series of characters of brittleness risk entropy are analyzed, which laid a foundation of studying on quantitatively determine the components of complex system brittleness.Once more, Because there are many uncertainties factors and complexity of system risk on the analysis brittleness risk of network communication system, analysis model of communication system brittleness is established, based on Multi-layer Fuzzy Comprehensive Evaluation method. Brittleness factors which can be affecting the communication network brittleness are identified through Delphi method. Analytic Hierarchy Process (AHP) method is applied to determining the collapse coefficients of brittleness risk affecting factors in hierarchy model, and Fuzzy Comprehensive Evaluation (FCE) measuring brittleness risk of system.Finally, we make an analysis to a practicality of communication network system, according to the concept of information surplus degree and the maximum entropy. Thus we make the measure to drop brittleness risk of system, by the result of application that brittleness risk size of system can been affected by brittleness status and brittleness structure. This result can be concluded that these methods can not only be used for quantifying the brittleness risk of network communication system, but also providing a workable way of evaluating brittleness risk of it effectively.更多还原