A surrogate-based optimization method with RBF neural network enhanced by linear interpolation and hybrid infill strategy

In engineering, it is computationally prohibitive to directly employ costly models in optimization. Therefore, surrogate-based optimization is developed to replace the accurate models with cheap surrogates during optimization for efficiency. The two key issues of surrogate-based optimization are how to improve the surrogate accuracy by making the most of the available training samples, and how to sequentially augment the training set with certain infill strategy so as to gradually improve the surrogate accuracy and guarantee the convergence to the real global optimum of the accurate model. To address these two issues, a radial basis function neural network (RBFNN) based optimization method is proposed in this paper. First, a linear interpolation (LI) based RBFNN modelling method, LI-RBFNN, is developed, which can enhance the RBFNN accuracy by enforcing the gradient match between the surrogate and the trend observed from the training samples. Second, a hybrid infill strategy is proposed, which uses the surrogate prediction error based surrogate lower bound as the optimization objective to locate the promising region and meanwhile employs a linear interpolation-based sequential sampling approach to improve the surrogate accuracy globally. Finally, extensive tests are investigated and the effectiveness and efficiency of the proposed methods are demonstrated.     

一种有效的复杂地质体剖面图自动生成方法*

提出了一种基于平面隐函数的实体切割思想和区域自动识别技术的地质体剖面图生成方法。其主要步骤包括：采用平面隐函数实现平面和地质实体的求交运算,生成离散的交线;采用基于KD树的空间索引方法,确定交线的邻接关系,生成地质体轮廓线;封闭的轮廓线将平面划分为多个复杂的区域(内含孔、岛),采用图论的多叉树形式化表达区域存储和组织;最后对多叉树表示的区域进行矢量图案的填充。实验结果表明,该算法有较好的计算效率,是一种比较实用的地质体剖切方法。     

Sex-identifying urine and molt signals in lobster (Homarus americanus)

During courtship, premolt female lobsters,Homarus americanus, choose a male and initiate a pair bond by repeated approaches to his shelter. The male allows such a female to share his shelter for about one week. This knowledge formed the basis to search for quantitative evidence for lobster sex pheromone(s) used in courtship: male cues to allow premolt females to identify a preferred male, and female cues to allow males to identify a premolt mature female. In each of four 1500-liter naturalistic aquaria, the behavioral responses of one female and two male lobsters to male and female lobster urine (0.5 ml) and body odor (20 ml) stimuli were observed. These stimuli were injected once or twice per day into a continuously flowing delivery tube attached to lobster shelters. Habituation to stimulus introduction—a serious problem in earlier experiments—was apparently avoided in the more natural social and physical environment we employed in these experiments. We demonstrated that male and female molt body odors contain different chemical substances: females responded to male molt body odor and males responded to female molt body odor but not vice versa. In general, male and female intermolt urine caused strong responses; however, females responded only weakly to male urine. This suggests that male and female urine are chemically different. Female urine and molt body odor caused a typically male high-on-legs response. These results show that molt body odors and intermolt urine contain sex-specific substances, which may be used in lobster courtship as well as other social interactions.     

基于RetinaNet-CPN网络的视频人体关键点检测

为解决基于视频流的人体关键点检测效果不佳及视频流切片后可能会发生运动模糊的问题,提出了一种改进的RetinaNet-CPN网络对人体关键点进行检测,有效解决切片后运动模糊图像的干扰并提高了人体关键点的检测准确率.视频流切片后,先用改进的RetinaNet网络检测出图片中的所有人并对每个目标框做模糊检测,对大于阈值的目标框做去模糊处理,最后用引入注意力机制的CPN网络提取关键点.将RetinaNet衡量预测框与真实框差异的IOU函数改成DIOU后,在仿真实验中目标检测AP提高了近3％;对于模糊的图片,利用匀速直线运动频谱特征估算出的模糊核与实际模糊核相差不大,对其做去模糊处理后基本能恢复出原清晰图片;同时引入注意力机制为各通道和特征层分配合理的权重,使得CPN检测AP提高近1％,AR提升0.5％.     

基于改进双跳无线体域网的身份认证方案

在无线体域网（WBAN）的身份认证中,针对原方案通◢信传感器节点可追踪和不具备匿名性的缺点,提出一种改进双跳身份认证方案,并保留了原方案的高效运算特性。引入了二级节点N的保密密钥和核心节点HN的身份验证参数两个独立的保密参数,并让二级节点N的保密密钥独立。并与原方案保密值组成三个保密值,通过三个保密值来确保◣参数的保密性和新鲜性。安全性分析和BAN逻辑表明所提方案具备不可追踪性和匿名性,且计算成本与原方案相近,通信成本更低,存储成本略有上升。所提方案是原方案的有效改进。     

6×××系汽车车身板铝合金的合金化及组织性能的研究

采用力学性能试验和光学金相分析法,研究了Mg和Mn等元素对6×××系汽车板铝合金组织及性能的影响。结果表明:Mg和Mn是提高6×××系汽车板铝合金时效初期硬化值的元素,但对时效初期硬化速度影响不大。提高Mg和Mn含量,合金经170℃30min时效后的强度增大,伸长率降低。合金固溶处理淬火后的晶粒尺寸和形态不随Mg含量而改变,但随Mn含量增加而显著细化。     

无线体域网高能效竞争接入机制研究

低能耗是无线体域网WBAN通信最重要的指标.构建了IEEE 802.15.6无线体域网CSMA/CA竞争接入的功耗模型,分析了用户优先级、MAC帧负荷尺寸和数据到达量对信道接入能耗的影响,提出了一种通过数据汇聚传输提高能量效率的增强型信道接入方法.仿真结果表明,在低业务量及混合业务场景下,所提方法能够显著减少信道接入的能量消耗.     

一种改进的XML访问控制模型

在XML访问控制方面,改进传统的XML控制技术在现有模型中授权主体的定义范畴,同时将请求评估模块和决策执行模块显式分开,增加定制安全策略和安全规则的灵活性和通用性。     

Continuous collision detection for adaptive simulation of articulated bodies

We perform continuous collision detection (CCD) for articulated bodies where motion is governed by an adaptive dynamics simulation. Our algorithm is based on a novel hierarchical set of transforms that represent the kinematics of an articulated body recursively, as described by an assembly tree. The performance of our CCD algorithm significantly improves as the number of active degrees of freedom in the simulation decreases.