低速率移动自组织网络安全路由协议

由于移动Ad Hoc网络具有无需基础设施,可以快速部署,很难完全被摧毁等优点,被作为数字化战场通信的优先选择,设计实现了一种适用于低速率移动Ad Hoc网络的安全路由协议——LSRP,它适用于分层架构的低速移动电台自组网,可以在网络资源有限的情况下,防御来自外部和内部两方面的攻击。LSRP协议利用分布式门限RSA签名方案来阻止非授权接入,防范来自网络外部的攻击;利用邻节点信任度评估机制,进行恶意节点检测和剪除,防范来自网络内部的攻击,增强了网络的安全性。     

基于复杂网络的维修保障力量体系建模方法

针对传统的层级式维修保障力量体系在维修保障过程中暴露出的缺陷,从维修保障力量体系与复杂网络的相似性角度出发,提出用复杂网络理论来研究维修保障力量体系。在对维修保障力量活动过程进行网络化抽象的基础上,建立了维修保障力量网络的拓扑模型,提出了描述维修保障力量网络的特征量。以师维修保障力量体系作为研究对象建立了网络拓扑模型,并对该网络结构作了适当改进。通过一系列对比发现改进后的网络更有利于维修保障,这为维修保障力量体系建设提供了一种新思路和新方法。     

放大转发双向中继网络的移动级联信道统计特征分析

针对放大转发协议下双向中继网络移动级联信道,为了分析其统计特征,结合用户完成信息交互的过程,将移动级联信道分解为传输部分和干扰部分两个时间过程。依据它们的双高斯乘积特点,分别给出了对应的一阶概率密度分布、二阶自相关函数和多普勒功率谱的表达式。与此同时,通过采用Gamma分布对级联信道进行了概率近似,得到了描述移动级联信道时变特性的电平通过率和平均衰落时间等指标的近似计算式。借助数值仿真验证了上述表达式的正确性。     

Financial versus human resources in the Greek‐Turkish arms race: A forecasting investigation using artificial neural networks

This paper aims at forecasting the burden on the Greek economy resulting from the arms race against Turkey and at concentrating on the leading determinants of this burden. The military debt and the defence share of GDP are employed alternatively in order to approximate the measurement of the arms race pressure on Greece, and the method used is that of artificial neural networks. The use of a wide variety of explanatory variables in combination with the promising results derived, suggest that the impact on the Greek economy resulting from this arms race is determined, to a large, extent, by demographic factors which strongly favour the Turkish side. Prediction on both miltary debt and defence expenditure exhibited highly satisfactory accuracy, while the estimation of input significance, indicates that variables describing the Turkish side are often dominant over the corresponding Greek ones.     

基于复杂网络理论的作战系统结构拓扑模型研究

基于复杂网络理论,分析了信息化作战系统网络结构,构建了传统作战系统和信息化作战系统拓扑模型的生成算法,并通过对传统作战系统和信息化作战系统结构拓扑模型度量性质的比较,说明了信息化战争下的作战系统结构特点,为进一步研究网络化战争提供了良好的模型基础．     

The Trust Paradox in Nuclear Smuggling

ABSTRACT

This article explores the paradox of trust in the largest nuclear smuggling operation involving highly enriched uranium (HEU) discussed in open source literature. In the first effort to understand the type, extent, and role of trust in nuclear smuggling enterprises, it draws from literature on trust development in legitimate businesses as well as criminal enterprises. Observed behavioral patterns in this case challenge traditional notions of the internal dynamics of temporary groups engaged in nuclear smuggling and operational realities of such activities. The article seeks to explain why individuals agree (and continue) to operate in this high-risk environment, unbound by close personal ties, without any effort to verify the background, motives, or qualifications of the fellow conspirators. It offers ways to advance current nonproliferation efforts in non-state actor interdiction by exploiting the environment of shallow trust in temporary groups.     

具有时滞的Cohen—Grossberg神经网络的Hopf分支全局存在性研究

摘要：利用泛函微分方程的度理论,研究一类具有时滞的Cohen-Grossberg神经网络的全局分支的存在性,研究结果为该类神经网络的应用设计提供理论基础．     

Strategic intelligence during coin detention operations – relational data and understanding latent terror networks

One aspect of the global “War on Terror” that has received limited coverage in the academic literature is the problem of detained persons as it relates to intelligence. This is a surprising oversight, given the number of detainees that the USA has been responsible for (over 25,000 were in custody in Iraq alone at one time during its peak). The detention environment offers a prime strategic intelligence opportunity for the US intelligence community to study the tactics and organizations of individuals who have been removed from the overall conflict. In this article an easily implemented collection program is recommended to be deployed in US/coalition detention centers. The primary recommendation is to gather relational data on detainee communication, both authorized and illicit, and to use these data to perform network analyses of terrorist groups and their individual members.     

一种基于模糊竞争网络的主动声纳目标分类器

提出了一种基于模糊竞争网络的主动声纳目标分类器。该分类器结构简单,学习速度快,仅用少量样本进行训练便可获得良好的推广性能．便于实时学习和应用。通过对模式属于各类别的隶属度的分析,可以知道判决的可靠性,具有较好的应用价值。     

未知非线性系统的神经网络建模与控制仿真研究

针对一类未知非线性系统的建模与控制问题,采用基于神经网络的多步预测控制算法进行了仿真研究,仿真对象选取控制工程界常用的单摆试验装置.该算法的实现步骤为:获取系统开环试验数据;辨识神经网络正向动态模型;设计非线性优化控制律.仿真研究结果表明,所提出的建模与控制方法是有效的.