Optimization by Estimation of Distribution with DEUM Framework Based on Markov Random Fields

This paper presents a Markov random field （MRP） approach to estimating and sampling the probability distribution in populations of solutions. The approach is used to define a class of algorithms under the general heading distribution estimation using Markov random fields （DEUM）. DEUM is a subclass of estimation of distribution algorithms （EDAs） where interaction between solution variables is represented as an undirected graph and the joint probability of a solution is factorized as a Gibbs distribution derived from the structure of the graph. The focus of this paper will be on describing the three main characteristics of DEUM framework, which distinguishes it from the traditional EDA. They are： 1） use of MRF models, 2） fitness modeling approach to estimating the parameter of the model and 3） Monte Carlo approach to sampling from the model.

On the horizontal distribution of algal-bloom in Chaohu Lake and its formation process

Based on the remote sensing images of algae, the present work analyzes the horizontal distribution characteristics of algal blooms in Chaohu Lake, China, which also reveals the frequency of algal blooms under different wind directions. Further, an unstructured-grid, three-dimensional finite-volume coastal ocean model （FVCOM） is applied to investigate the wind-induced currents and the transport pro- cess to explain the reason why algal blooms occur at the detected places. We first deduce the primary distribution of biomass from overlaid satellite images, and explain the formation mechanism by analyzing the pollution sources, and simulating the flow field and transportation process under prevailing wind over Chaohu Lake. And then, we consider the adjustment action of the wind on the corresponding day and develop a two-time scale approach to describe the whole formation process of algae horizontal distribution in Chaohu Lake. That is, on the longer time scale, i.e., during bloom season, prevailing wind determines the primary distribution of biomass by inducing the characteristic flow field; on the shorter time scale, i.e., on the day when bloom occurs, the wind force adjusts the primary distribution of biomass to form the final distribution of algal bloom.

Origin of diffraction fringes in two-dimensional photoelectron momentum distributions for single ionization of atoms in few-cycle intense laser pulses

We use the strong field approximation with a time window function controlling the release time of electrons to study the intra-cycle and inter-cycle interferences in few-cycle intense laser pulses impinging on He. The diffraction fringes, i.e., the vertical stripe-like structure, observed in the experimental two-dimensional photoelectron momentum distributions of Gopal et al. （2009 Phys. Rev. Lett. 103 053001） have been attributed to the interplay of the intra- and inter-cycle interferences. The pure numerical calculations by solving the time-dependent Schrrdinger equation are also performed and the results are compared with the experimental measurements directly. It has been found that the position of the stripe-like structure can be used to determine the duration of the laser pulses used in experiments.

The Limiting Distribution of the MLE for the Location Parameters in Nonregular Translation Distributions and Its Asymptotic Efficiency

In this paper, we consider the location model Y = θ + 6, where θ is an unknown parameter, and e is the error belonging to the interval [a,b]. We assume that θhas the following density function: Then we give the limiting distribution of MLE θn for 1 < min(α,β) < 2 and consider the Bahadur asymptotic estimator. Since the results depend only on α,β,C1,C2 and are independent of the concrete form of f(x), they have adaptability.

Analysis and Verification on the Chain-like Model with Normal Distribution of Magnetorheological Elastomer

Magnetorheological 弹性体(MRE ) 是聪明的材料的一种新类型， rheological 性质能被外部磁场很快控制。它主要由橡胶和微米大小的铁磁性的粒子组成，它形成像链的结构。因此它的机械，电，并且磁性能被应用磁场改变，它作为导致磁电机的效果被叫。但是这效果不为设计申请当前是足够显著的。材料准备优化参数提高导致磁电机的效果因此是重要的。在这个工作，基于像链的模型，影响 MRE 的导致磁电机的效果的一些因素被与链介绍的角度的正常分发使用偶极子方法理论上分析。因素包括了粒子链的倾斜的角度，磁场紧张，并且砍种类，等等。一些实验也被执行。

Detecting N-nitrosamines in water treatment plants and distribution systems in China using ultra-performance liquid chromatography-tandem mass spectrometry

N-nitrosodimethylamine （NDMA） and several other N-nitrosamines have been detected as disinfection by-products in drinking waters in many countries around the world. An ultra-performance liquid chromatography- tandem mass spectrometry method with solid phase extraction sample preparation was developed to study the occurrence of N-nitrosamines in several water treatment plants and distribution systems in China. Isotope labeled N-nitrosodi-n-propylamine-dl4 （NDPA-dl4） was selected as the internal standard for quantification. The solid phase extraction procedures including pH, enrichment process and MS/MS parameters including capillary voltage, cone gas flow, cone voltage, collision energy were optimized to give average recoveries of 26% to 112% for nine N- nitrosamine species. The instrument detection limits were estimated to range from 0.5 to 5μg.L-1 for the nine N- nitrosamine species. NDMA and several other N-nitrosa- mines were found at fairly high concentrations in several water treatment plants and distribution systems. NDMA was found in all locations, and the highest concentrations in cities B, G, T, and W were 3.0, 35.7, 21.3, and 19.7 ng. L 1, respectively. A wide range of N-nitrosamines concentrations and species were observed in different locations. Higher concentrations of N-nitrosamines were detected in distribution systems that were further away from the treatment plants, suggesting that the contact time between the residual disinfectant and natural organic matter may play an important role in the formation of these compounds.

Particle size distribution and shape control of Au nanoparticles used for particle gun

Au nanoparticles are expected for the media to transfer genes into plants. However, the control of particle size distribution （PSD） and shape of Au nanoparticles is too difficult to design and prepare particles with suitable quality for the gene supporting media. Reduction crystal- lization experiments were performed in aqueous solution in order to clarify the effect of feeding conditions such as feeding profile, feeding rate, and feeding amount on PSD and shape of Au nanopartieles. Ascorbic acid （AsA） was selected as a reducing agent because it is safe for plants. Au particles of 50nm, 50-200nm, and 150-400nm were obtained in batch operation, single-jet, and double-jet, respectively. Moreover, in single-jet and double-jet, the mean size of the obtained Au particles increases with the decrease of feeding rate or the increase of feeding amount. It is concluded that PSD of Au nanoparticles can be controlled in the range of 50-400 nm by changing feeding conditions of AsA and HAuCl4 aqueous solution.

Optimization of Grain Boundary Character Distribution in Fe-18Cr-18Mn-0.63N High-Nitrogen Austenitic Stainless Steel

Grain boundary engineering（GBE） is a practice of improving resistance to grain boundary failure of the material through increasing the proportion of low Σ coincidence site lattice（CSL） grain boundaries（special grain boundaries） in the grain boundary character distribution（GBCD）. The GBCD in a cold rolled and annealed Fe-18Cr-18Mn-0.63N high-nitrogen austenitic stainless steel was analyzed by electron back scatter difraction（EBSD）. The results show that the optimization process of GBE in the conventional austenitic stainless steel cannot be well applied to this high-nitrogen austenitic stainless steel. The percentage of low ΣCSL grain boundaries could increase from 47.3% for the solid solution treated high-nitrogen austenitic stainless steel specimen to 82.0% for the specimen after 5% cold rolling reduction and then annealing at 1423 K for 10 min.These special boundaries of high proportion efectively interrupt the connectivity of conventional high angle grain boundary network and thus achieve the GBCD optimization for the high-nitrogen austenitic stainless steel.

The probability distribution of fatigue damage and the statistical moment of fatigue life

The randomization of deterministic fatigue damage equation leads to the stochastic differential equation and the Fokker-Planck equation affected by random fluctuation. By means of the solution of equation, the probability distribution of fatigue damage with the change of time is obtained. Then the statistical moment of fatigue life in consideration of the stationary random fluctuation is derived. Finally, the damage probability distributions during the fatigue crack initiation and fatigue crack growth are

湖山露天铀矿多排孔大区块压渣爆破效果评价

为改善湖山铀矿爆破位移大、矿岩大块率高和前部铲装效率低等问题,湖山铀矿设计并开展了5个大区块压渣爆破试验,每个区块量为36~53万t,炮孔排数15~21排,自由面预留5~40 m宽的渣体,测定和统计了试验区块爆堆形态、装载效率、爆破块度以及底板标高偏差等指标,并与同期清渣爆破区块进行对比分析。结果表明:在爆炸冲击作用下,爆堆存在前扑及隆起现象,爆堆表面呈高低不等连续分布状态,爆堆不同部位最大增高变化范围为4.06~5.85 m;铲装效率方面,液压铲在压渣爆破区块的平均装载效率为2722 t/h,比同期清渣爆破区块平均值高4.69%;爆破块度方面,压渣爆破区块爆后块度分布优于清渣爆破区块,在控制大块率方面更有优势。此外,压渣爆破在底板标高控制方面有一定优势。该研究首次将压渣爆破技术推广至多排孔大区块(15~21排)爆破中,这既符合大型露天矿山生产的客观需求,也是设备大型化后的必然趋势。可以使得湖山铀矿爆破更为精细化的同时,也可以为国内外类似的矿山爆破提供重要的现场数据。

高频聚焦超声声场和温度场的仿真研究

为探究临床常用的7 MHz高频聚焦超声在多层生物组织中的声传播以及毫秒级时间内的生物传热规律问题,基于Westervelt方程和Pennes传热方程,使用有限元方法建立高频聚焦超声辐照多层组织的非线性热黏性声传播及传热模型。首先分析了线性模型和非线性模型之间的差异,然后在非线性模型下探究换能器的参数对声场和温度场的影响。仿真结果显示:在7 MHz频率下,当换能器输出声功率超过5 W时,声波传播的非线性效应不可忽视(p<0.05);当声功率从5 W增大到15 W时,非线性模型与线性模型预测的温度偏差从20%增加到34.703%;高频聚焦超声波的非线性行为比低频更加显著,基频能量向高次谐波转移的程度增大,声功率为10 W和15 W时4次谐波与基波之比分别达到7.33%和12.12%;高频换能器参数的改变对组织中声场和温度场分布的影响较大,换能器焦距从12 mm减小到11.2 mm,焦点处最高温度增加了77%。结果表明,7 MHz聚焦超声的非线性声传播需要考虑到4次谐波的影响。该文提出的多层组织非线性仿真模型可为高频聚焦超声换能器参数优化及制定安全、有效的术前治疗方案提供理论参考。

适应IIDG接入的就地型馈线自动化改进策略

随着逆变型分布式电源(IIDG)的渗透率不断提升,基于单向潮流设计的就地型馈线自动化(FA)可能无法应对IIDG接入的场景。针对这一问题,以目前应用较多的2种就地控制型FA模式,即电压-时间型以及电压-电流型FA为例,详细分析了IIDG接入对就地控制型FA各开关动作情况的影响。在此基础上,结合电流保护、电压序分量以及方向元件对FA系统进行改进,提出了适用于有源配电网的就地控制型FA方案。最后,搭建FA系统仿真模型进行验证。结果表明,该策略可有效缓解IIDG接入对出口断路器保护、无压分闸以及电压电流型开关过流检测的影响,提高有源配电网的故障处理水平。

解构中国经济韧性——即时监测、联动辨识与贡献分解

本文基于经济增长脆弱性的测算模型,对中国省级经济韧性进行动态测度。在此基础上,运用MAI-AR-SV模型,分别提取中国宏观及区域层面经济韧性的共同因子,旨在系统性监测经济韧性状态变迁的同时,辨识经济韧性间联动机制的演化过程。进一步,通过对省级经济韧性的内生驱动因素进行贡献分解,深入辨析经济韧性的区域协同与异质分化特征。研究发现:(1)中国宏观经济韧性的动态演化路径存在显著的“事件驱动”特征,在极端事件冲击下,经济韧性表现出明显的波动态势。近年来,尽管受到全球重大公共卫生事件冲击,但中国经济恢复较快,继而于2021年由低韧性攀升至高韧性状态,意味着中国经济富有强劲的增长韧性。(2)中国区域经济韧性的状态变迁路径在样本区间内呈现出显著的异质分化特征,但是,随着后疫情时期经济形势的相对稳定,四大区域的经济运行无一例外地步入高韧性增长路径,这表明,中国经济稳定向好的长期发展趋势仍未改变。此外,区域经济韧性间的联动辨识结果表明,区域经济韧性已然由传统的三元分布结构演化为相互促进、优势互补的协调发展格局,具体表现为东部率先引领,同时仿射其他三区的雁阵式联动网络。(3)中国省级经济韧性的贡献分解结果表明,不同区域内省级经济韧性的同步程度存在明显的异质性,其中,东中部地区呈现出显著的“俱乐部收敛”趋向,而西部和东北各省经济韧性则具有多元异化的走势特征,说明区域内部经济发展的不平衡与不充分依旧显见。

差分隐私结合GAN的图像数据保护方法仿真

图像数据的过度挖掘会对图像隐私性带来影响,使得数据被泄露或滥用的风险不断增加。为避免图像隐私数据泄露,提出差分隐私结合GAN的图像隐私数据保护算法研究。针对当前图像数据参杂的不确定性,通过对图像数据清洗,获取的图像重要隐私区域,有利于后续噪声的加入,便于实现图像隐私数据的保护;根据差分隐私与生成对抗网络(Generative Adversarial Networks, GAN)结合,构建图像数据差分隐私保护的GAN模型,依据拉普拉斯实现机制,对清洗后的图像数据扰动,达到隐私数据保护的目的;由于模型在迭代期间会对图像数据隐私性产生侵犯,所以进一步对图像隐私性侵犯程度度量,令上述算法在实际应用时有效兼顾图像隐藏数据的安全性及隐私性。实验结果表明,所提方法对图像的加密和解密时间短,且隐私性高。

地铁隧道壁挂式配电箱的活塞风作用时变特性

列车活塞风效应直接影响地铁隧道内各种附属设施的疲劳使用寿命.针对活塞风作用下地铁隧道内壁挂式配电箱所受气动荷载的时变过程,采用数值模拟方法研究了列车活塞风作用下隧道内不同安装方式配电箱的表面压力分布及所受气动荷载变化规律的影响.研究结果表明:在车头或车尾经过配电箱时,配电箱表面压力值达到最大且为负压;贴壁式安装与悬空式安装的配电箱在活塞风作用下均承受拉压交变气动力,前者配电箱所受横向气动力较大,而后者配电箱所受纵向气动力较大,且最大值均为拉力;在车速较低、活塞风未进入自模区的情况下,配电箱表面静压值与列车车速相关,且与速度作用水头值呈线性关系.配电箱表面所受压拉力变化与箱体表面正负压变化对应;同等条件下,贴壁式配电箱所受荷载约为悬空式配电箱5倍,悬空式配电箱安装方式更安全,但应注意箱体在隧道纵向方向的维护和加固.

基于GDACO算法的引导疏散路径规划研究

针对教学建筑内中人员疏散存在盲目、无序、效率低等问题,提出了双层蚁群路径规划算法(GDACO)引导人员疏散的路径规划。对环境栅格进行处理,增添危险栅格,使初始蚁群信息素呈阶梯式分布;提出双层蚁群算法,引导层蚁群通过全局搜索寻找引导路径,疏散层蚁群基于引导范围实现疏散路径寻优;进行路径优化处理,剔除冗余节点,对转折部分进行平滑处理。试验结果分析表明,所提GDACO算法能够减少路径转折点,规避风险路径,满足教学建筑内的引导疏散需求,规划路径符合实际疏散情况。

整体化铺装层加固空心板桥受力性能试验研究

为了改进装配式空心板桥横向受力性能,以采用整体化铺装层加固的装配式空心板桥为研究对象,进行一跨8 m足尺模型室内试验,开展了车辆荷载作用下空心板桥受力性能和破坏模式试验研究,并与加固前的空心板桥试验结果进行对比。试验结果表明:采用整体化铺装层进行加固后,空心板的开裂荷载从100 kN提高到125 kN,提高了25%,说明整体化铺装层可以明显提高空心板的开裂荷载和抗弯受力性能。加固前的空心板桥出现了明显的单板受力现象,但是,对于采用整体化铺装层加固后的空心板桥,由于整体化铺装层的作用使得即使出现铰缝结合面开裂也可以保证空心板间的荷载传递,因此即使在3.0倍公路-Ⅰ级荷载作用下空心板之间的挠度差始终相对较小。

昌黎海域青岛文昌鱼栖息地评价及其分布特征

利用2021年调查数据,对昌黎海域青岛文昌鱼栖息地环境及其分布特征开展综合研究。研究结果表明,保护区海域水质、沉积物质量状况良好。青岛文昌鱼主要分布在中砂至细砂类型的底质中,主要粒组为0.125~0.500 mm,平均栖息密度为70.4 ind./m2,平均生物量为4.71 g/m2,Ⅲ龄青岛文昌鱼占优势地位。青岛文昌鱼密度与COD和浮游植物密度呈显著相关(p<0.05),与浮游植物种类数呈极显著相关(p<0.01);生物量与COD、沉积物铅和浮游植物种类数呈显著相关(p<0.05)。

基于GPM卫星探测的夏季高原涡降水结构特征

利用GPM卫星的双频反演产品DPR_MS和ERA5再分析数据,对2016~2020年夏季青藏高原上空能被GPM卫星探测到的32例高原涡降水特征进行研究。结果表明:(1)高原涡在高原上降水强度偏小,绝大部分低于2 mm/h,雨顶高度主要在6~9 km。降水类型以深厚弱对流为主(占59.64%),其次是浅薄降水(占40.35%),深厚强对流极少。(2)高原西部降水涡绝大多数为深厚弱对流、平均雨顶高度明显高于东部涡,但东、西部涡的平均降水强度没有明显差异。浅薄降水频次呈“东高西低、北高南低”的空间分布特征,深厚弱对流降水则相反。(3)深厚弱对流降水受高原涡强上升运动影响显著,在6.5~7.5 km高度有大量有效半径为1.1 mm的雨滴粒子堆积,近地面较强降水(7 mm/h)的雨滴破碎过程明显。浅薄降水的反射率因子在各个高度持续增大,其地面降水主要由雨滴粒子碰并形成,地表雨强受雨滴浓度影响明显。

基于网络药理学和分子对接探讨山楂叶治疗2型糖尿病的作用机制

目的:基于网络药理学和分子对接技术分析山楂叶治疗2型糖尿病(type 2 diabetes mellitus, T2DM)的作用机制。方法:从中药系统药理学数据库与分析平台(traditional Chinese medicine systems pharmacology database and analysis platform, TCMSP)、BATMAN、HERB、SymMap、TCM-ID等数据库筛选山楂叶中的化学成分,构建山楂叶化学成分库并导入SwissADME数据库筛选山楂叶的活性成分,从Swiss Target Prediction数据库检索相关文献并获取山楂叶的潜在靶点。利用DrugBank、TTD、GeneCards、CTD、OMIM数据库检索T2DM相关靶点,绘制韦恩图得到两者的交集靶点。借助STRING数据库和Cytoscape 3.9.1软件构建蛋白质相互作用网络,进行拓扑分析从而筛选关键靶点和核心成分,并采用AutoDock 4.2.6软件进行分子对接验证。在DAVID 6.8数据库中进行基因本体(gene ontology, GO)功能富集分析和京都基因与基因组百科全书(Kyoto encyclopedia of genes and genomes, KEGG)通路富集分析,并筛选关键通路。利用BioGPS和The Human Protein Atlas数据库分析关键组织及高表达基因在不同细胞的分布情况。结果:山楂叶活性成分92个及潜在靶点622个,山楂叶与T2DM的交集靶点306个,进一步筛选得到16个关键靶点,包括ALB、GAPDH、IL-6、SRC、TNF、AKT1、EGFR、VEGFA、MAPK3、ESR1、STAT3、PTGS2、mTOR、HSP90AA1、NR3C1、PPARA。山楂叶核心成分28个,包括α-亚麻酸、柚皮素、桑黄素等。KEGG富集分析筛选得到170条通路,其中血糖调节相关通路包括AGE-RAGE信号通路在糖尿病并发症中的作用、PI3K-Akt信号通路、血管内皮生长因子信号通路、胰岛素抵抗、胰岛素信号通路等。分子对接结果显示,14个关键靶点与27个核心成分均具有良好的亲和力和结合活性。交集基因在甲状腺、肝脏、心脏和肺中的数量最多,主要分布在甲状腺腺细胞、肝细胞、肺泡细胞、平滑肌细胞等。结论:山楂叶治疗T2DM具有多成分、多靶点、多通路、多组织的特点,其可能通过柚皮素、桑黄素、木犀草素等多种活性成分作用于IL-6、AKT1、VEGFA、STAT3等关键靶点,调节AGE-RAGE信号通路、胰岛素抵抗等多条通路。