大延时网络中的主动队列管理机制

研究了网络延时对路由器主动队列管理机制的影响,分析了几种典型主动队列管理算法在大延时网络中的性能。在介绍了基于内模补偿的DC-AQM算法的优缺点之后,根据PID控制器延时补偿的Ziegler-Nichols设定方法,提出了ZNDC(Ziegler-Nichols delay compensation)AQM算法并进行了仿真实验验证,实验结果表明算法达到了预期的目标。     

气囊辅助RTM制备复合材料承力筒内腔尺寸控制技术

根据气囊充压压力与复合材料承力筒内径之间的变化关系,给出了成型复合材料承力筒内表面的气囊尺寸计算方程。分析了气囊充压压力增大对承力筒的壁厚、纤维含量和弯曲性能的影响。采用气囊辅助RTM工艺整体制备出复合材料承力筒。试验结果表明,气囊充压压力使复合材料承力筒的壁厚减薄,纤维体积含量增加,弯曲性能提高;选择适当的气囊充压压力可以制备出满足设计要求的复合材料承力筒。     

对象边框标注数据的弱监督图像语义分割

针对图像语义分割应用中像素级标注数据费时昂贵的问题,主要研究以对象边框标注数据为代表的弱监督模型下的图像语义分割方法。使用基于金字塔的密集采样全卷积网络提取图像的像素级特征,并用GrabCut算法转化对弱监督数据进行数据标记,通过将图像特征和标记数据进行联合训练,构建了基于金字塔密集采样全卷积网络的对象边框标注弱监督图像语义分割模型,并在公开数据集上进行了验证。实验结果表明,所构建的弱监督模型与DET3-Proposed模型、全矩形转化模型以及Bbox-Seg模型相比,达到了更好的分割效果。     

面向癫痫脑电的简化深度学习模型

针对脑电信号随机性强、动态变化迅速等特点,提出了一种简化深度学习模型研究癫痫脑电识别问题。提出的模型以一维卷积神经网络为基础,在结构方面简化了卷积层、池化层等以提高模型效率,在整体框架方面应用了Keras框架,在训练优化算法方面采用RMSProp算法作为模型优化算法,通过预定义的目标函数来进行损失估计,模型设计上加入了批标准化层和全局均值池化层。基于所提模型,从三个方面研究了癫痫脑电识别问题,即:利用经验模态分解,分别选取前三阶、前五阶、前七阶、前八阶的本征模态函数分量,在简化模型上进行对比分析;利用提出模型所具备的深度学习特点,直接识别原始脑电信号而无须特征提取环节;增加了三种不同方法分别提取7类特征,对相同的脑电数据进行对比分析。性能分析结果表明:对于五类不同的脑电信号,前三阶的本征模态函数分量的识别率达到92.1%,比其他几种处理方式识别率高;前八阶的本征模态分量识别率不及原始信号,表明人工数据处理时会给数据带来噪声; 所提出的简化深度学习模型能高效处理癫痫脑电识别问题,具备较高效率和较好性能。     

美国陆军网络装备作战评估及启示

作战试验能有效地推动装备体系的建设发展,本文通过分析美国陆军网络集成评估与联合作战评估的作战试验特点,提出军队开展装备作战试验的几个建议。首先,总结了美国陆军网络集成评估到联合作战评估的发展演变,主要包括网络集成评估的地位和作用、网络集成评估的实施流程以及13次网络集成评估和3次作战评估的有关情况;其次,分析了网络集成评估的一系列作战试验特点,诸如实际作战需求牵引、逼真场景构建、试训一体化模式、部队战斗力的直接生成以及作战能力的军民融合式发展等;最后,提出了网络集成评估带来的若干启示,从满足实战能力需求开展作战试验、突出装备体系的互操作能力评估、探索试训结合的作战试验模式、组建专门的作战试验部队及鼓励工业部门参与作战试验等方面推动军队作战试验的发展。     

Damage of multi-layer spaced metallic target plates impacted by radial layered PELE

Three different kinds of PELE (the penetrator with lateral efficiency) were launched by ballistic artillery to impact the multi-layer spaced metal target plates. The impact velocities of the projectiles were measured by the velocity measuring system. The damage degree and process of each layer of target plate impacted by the three kinds of projectiles were analyzed. The experimental results show that all the three kinds of projectiles have the effect of expanding holes on the multi-layer spaced metal target plates. For the normal structure PELE(without layered) with tungsten alloy jacket and the radial layered PELE with tungsten alloy jacket, the diameters of holes on the second layer of plates are 3.36 times and 3.76 times of the diameter of the projectile, respectively. For radial layered PELE with W/Zr-based amorphous composite jacket, due to the large number of tungsten wires dispersed after the impact, the diameter of the holes on the four-layer spaced plates can reach 2.4 times, 3.04 times, 5.36 times and 2.68 times of the diameter of the projectile. Besides, the normal structure PELE with tungsten alloy jacket and the radial layered PELE whit tungsten alloy jacket formed a large number of fragments impact marks on the third target plate. Although the number of fragments penetrating the third target plate is not as large as that of the normal structure PELE, the area of dispersion of fragments impact craters on the third target plate is larger by the radial layered PELE. The radial layered PELE with W/Zr-based amorphous composite jacket released a lot of heat energy due to the impact of the matrix material, and formed a large area of ablation marks on the last three target plates.     

Research on construction of operation architecture based on complex network

Operation architecture plays a more important role in Network Centric Warfare (NCW), which involves dynamic operation networks with complex properties. Thus, it is essential to investigate the operation architecture under the informatization condition within NCW and find a proper network construction method to efficiently coordinate various functional modules on a particular situation, i.e., the aerial combat. A new method integrating the physical level and functional level of NCW is proposed to establish the operation architecture, where the concept of network operation constraints unit and net constructing mechanisms are employed to avoid conflicts among different platforms. Meanwhile, we conduct simulations to assess the effectiveness and feasibility of the constructed operation architecture and analyze the influence of the network parameters.     

Maximizing deviation method for multiple attribute decision making under q-rung orthopair fuzzy environment

Because of the uncertainty and subjectivity of decision makers in the complex decision-making environment, the evaluation information of alternatives given by decision makers is often fuzzy and uncertain. As a generalization of intuitionistic fuzzy set (IFSs) and Pythagoras fuzzy set (PFSs), q-rung orthopair fuzzy set (q-ROFS) is more suitable for expressing fuzzy and uncertain information. But, in actual multiple attribute decision making (MADM) problems, the weights of DMs and attributes are always completely unknown or partly known, to date, the maximizing deviation method is a good tool to deal with such issues. Thus, combine the q-ROFS and conventional maximizing deviation method, we will study the maximizing deviation method under q-ROFSs and q-RIVOFSs in this paper. Firstly, we briefly introduce the basic concept of q-rung orthopair fuzzy sets (q-ROFSs) and q-rung interval-valued orthopair fuzzy sets (q-RIVOFSs). Then, combine the maximizing deviation method with q-rung orthopair fuzzy information, we establish two new decision making models. On this basis, the proposed models are applied to MADM problems with q-rung orthopair fuzzy information. Compared with existing methods, the effectiveness and superiority of the new model are analyzed. This method can effectively solve the MADM problem whose decision information is represented by q-rung orthopair fuzzy numbers (q-ROFNs) and whose attributes are incomplete.     

Interdiction models for delaying adversarial attacks against critical information technology infrastructure

Information technology (IT) infrastructure relies on a globalized supply chain that is vulnerable to numerous risks from adversarial attacks. It is important to protect IT infrastructure from these dynamic, persistent risks by delaying adversarial exploits. In this paper, we propose max‐min interdiction models for critical infrastructure protection that prioritizes cost‐effective security mitigations to maximally delay adversarial attacks. We consider attacks originating from multiple adversaries, each of which aims to find a “critical path” through the attack surface to complete the corresponding attack as soon as possible. Decision‐makers can deploy mitigations to delay attack exploits, however, mitigation effectiveness is sometimes uncertain. We propose a stochastic model variant to address this uncertainty by incorporating random delay times. The proposed models can be reformulated as a nested max‐max problem using dualization. We propose a Lagrangian heuristic approach that decomposes the max‐max problem into a number of smaller subproblems, and updates upper and lower bounds to the original problem via subgradient optimization. We evaluate the perfect information solution value as an alternative method for updating the upper bound. Computational results demonstrate that the Lagrangian heuristic identifies near‐optimal solutions efficiently, which outperforms a general purpose mixed‐integer programming solver on medium and large instances.     

基于大数据分析的网络管理智能化

军事通信网综合管理系统中,各专业网络运行情况的主要数据都可汇总到综合网管数据库中。这些网管数据中隐含了网络运行的几乎全部信息,但即使是最有经验的网络管理专家面对这些繁杂的数据也只能望洋兴叹。面对大数据的智能分析技术则有可能实现网络运行态势的高度可视化,充分理解和详细把握全网运行状态,像知识、经验、联想丰富的专家一样不知疲倦地对网络运行状况进行全面细致地分析,从海量的网管数据中发现网络故障的蛛丝马迹,发现隐含的问题,预知网络变化。论文介绍了利用大数据分析技术实现网络管理智能化的现实目标,基于性能的网络异常预测预报、基于性能和事件的故障分析诊断、网络运行知识的挖掘利用等方面的功能和技术途径。