基于动态贝叶斯网络的复杂网络抗毁性分析

复杂网络抗毁性是复杂网络在节点或边遇敌攻击后能继续维持基本功能的能力,是衡量军事信息网络鲁棒性和敏捷性的重要指标。针对复杂网络及对作战体系支撑能力的多指标、复杂化和动态演绎特点,在静态分析方法基础上,提出基于动态贝叶斯网络的抗毁性分析方法。建立了复杂网络抗毁性指标体系。构建了基于动态贝叶斯网络的复杂网络抗毁性评估模型,提出确定评估模型参数的方法。仿真验证了方法的可行性和有效性。     

Remotely Piloted Aircraft,Risk, and Killing as Sacrifice: The Cost of Remote Warfare

In this paper I argue that a remotely piloted aircraft (RPA) pilot’s act of killing remotely, when it is done in the defense of another person, can be viewed as an act of sacrifice. This argument concludes from two premises. First, the RPA pilot faces psychological risk to self by carrying out such an action; second, the RPA pilot is motivated to some significant degree by something other than self-interest. Moreover, I challenge both the view that RPA represent merely an incremental development in technology yielding an incremental expansion of the distance between the war-fighter and the target and the view that RPA is revolutionary in some fundamental sense. I instead argue that, based upon the natural geometric boundary imposed by the earth’s shape, RPA have maximized the physical distance between the war-fighter and the target and, over time, this fact will change, not only the psychological effects on the RPA pilot, but also the way we view killing and distance in general. When taken together, these two broad claims, that other-defense can justify killing in war and that there is a psychological cost paid by remote killers, yield a conception of remote killing as an act of sacrifice that will inform the broader conversation about ethics, risk, killing, and RPA.     

The fertile ground of Hell’s carnival: Charles T. R. Bohannan and the US Army’s Counter Intelligence Corps’ investigations of war criminals,collaborators, and the Huk,in the Philippines 1945–1947

Insurgencies thrive in regions where government legitimacy is absent. In the post-war Philippines, Captain Charles T. R. Bohannan of the Army’s Counter Intelligence Corps became actively aware of this dynamic. Bohannan is best known for his later work with Edward Lansdale and Ramon Magsaysay in defeating the Huk Rebellion (1950–1954). Here the author examines Bohannan’s early investigative work against Japanese war criminals, wartime Filipino collaborators, and the rising threat of communist subversion most associated with the Huk. All of these experiences fed into what would be the successful campaign against the Huk, chronicled in his seminal work, Counter Guerrilla Operations: The Philippines Experience, and offers lessons on the investigative (as opposed to tactical or psychological) nature of effective counter-insurgency work, as it relates to both legitimacy in governance and the rise of insurgencies.     

濒海要地防空雷达目标威胁分析与对策

为有效应对未来濒海要地防空作战敌方雷达目标威胁,应按战术任务重点分析警戒引导、武器控制、侦察监视和航行保障等四类目标,准确判断敌雷达可能对我造成的威胁,并从跟踪目标动态、注重预警能力、坚持集优打弱三个方面进行应对。     

美国陆军数据战略目标及实施

为深入研究美陆军数据战略目标及实施策略,为我军数据战略研究提供借鉴,文章对美陆军2016年发布的最新《陆军数据战略》进行分析,指出其目标是实现数据的可视化、可获取、可信任、可理解和互操作,并对美陆军数据战略的具体实施策略进行了剖析,指出美国陆军数据战略推动的是美军军事作战理论转型,从"网络为中心"转变为以"数据为中心"作战理念,其核心意图在于以数据优势实现指挥决策优势。     

飞行器试验多测控目标发射零点应急处理方法

飞行器飞行试验任务中,发射零点T0是整个测控系统启动运行的基准点,是保证测控系统有条不紊正常工作的关键因素。特别是执行多个测控目标连射飞行试验任务,多个发射零点的正确处理更是关乎试验任务的成败。在分析发射零点形成机制基础上,对发射零点在指控系统实时测控软件中的处理方法进行了研究,提出了一种实时测控软件重启后重新获得多测控目标正确发射零点的应急处理方法,解决了试验过程中一旦实时测控软件因重大软件故障导致初始发射零点丢失的问题,保障了试验任务的顺利进行。     

基于AHP-PCR5的多传感器目标识别

目标识别系统中所获得的信息常常是高度冲突和不确定的。基于DSmT理论的多传感器目标识别,可以解决证据高度冲突情况下的信息融合问题。然而由于DSmT理论融合结果分类精细而不利于判决,需要将某些分类结果进行重新分配。层次分析法(AHP)包含不确定知识矩阵,生成基本信度分配函数。基于此提出AHP-PCR5方法进行证据高度冲突情况下多传感器综合目标识别,不仅提高识别精度,降低识别过程的不确定因素,同时引入折扣系数灵活处理识别过程中具有不同可信度的多传感器融合问题。     

静态初始条件下的歼击机中距攻防引导方法决策

在静态初始条件下,歼击机中距攻防引导方法的选择是一个定性与定量相结合的类别划分问题,因此,将粗糙集理论与概率神经网络相结合用于该问题的解决。首先,利用粗糙集理论实现专家知识约简、空战态势信息集压缩,得到最小决策信息集;其次,利用概率神经网络进行概率决策推理;最后,通过实例分析,结果表明决策推理正确,在不确定环境下仍然有效,提高了决策过程的自动化程度。     

一种基于动态循环筛选模型的指标体系建立方法

针对评估中指标体系构建问题,提出一种层次网指标体系模型,对其中底层指标集的确立问题提出一种基于动态循环筛选模型的指标集选取方法,该方法在每一轮的循环过程中,通过评价专家专业度对专家集进行了优化,通过数据信息熵评价数据有效度系数,进而得到指标的重要度。指标集的双重优化确保了专家集的专业性和指标集的合理性,实际算例分析证明了此方法的可行性和有效性。     

KK分布杂波下的距离扩展目标检测算法

针对服从KK分布的大拖尾雷达杂波背景下的扩展目标检测问题,利用球不变随机变量表示了KK分布雷达杂波模型。在假设目标回波幅度已知的情况下,研究了基于Neyman-Pearson准则的距离扩展目标最优积累检测器,并通过对目标幅度的最大似然估计,推导了广义最大似然比检验检测器模型。为了降低这两种检测器中因计算第二类修正的贝塞尔函数而引入的运算复杂度,提出了一种基于顺序统计量的广义似然比检测器。该检测器利用检测窗内幅度较大的距离单元回波作为目标回波进行判决。利用蒙特卡罗仿真对这三种算法的性能进行了验证与比较,虽然最优积累检测器与广义似然比检测器具有更好的检测性能,但实现困难,计算量大,而基于顺序统计量的广义似然比检测器则具有更高的实用性。