IMPROVEMENT AND REALIZATION FOR THE MAXIMUM WEIGHT MATCHING ALGORITHM

Some new concepts of effective incidence matrix,ascending order adjacency matrix andend-result vertex are introduced,and some improvements of the maximum weight matchingalgorithm are made.With this method a computer program in FORTRAN language is realized onthe computers FELIX C-512 and IBM-PC.Good results are obtained in practical operations.

Weighted Maximum Likelihood Technique for Logistic Regression

In this paper, a weighted maximum likelihood technique (WMLT) for the logistic regression model is presented. This method depended on a weight function that is continuously adaptable using Mahalanobis distances for predictor variables. Under the model, the asymptotic consistency of the suggested estimator is demonstrated and properties of finite-sample are also investigated via simulation. In simulation studies and real data sets, it is observed that the newly proposed technique demonstrated the greatest performance among all estimators compared.

Online Scheduling on Bounded Batch Machines to Minimize the Maximum Weighted Completion Time

We investigate the online scheduling problem on identical parallel-batch machines to minimize the maximum weighted completion time.In this problem,jobs arrive over time and the processing times(of the jobs)are identical,and the batch capacity is bounded.For this problem,we provide a best possible online algorithm with a competitive ratio of(√5+1)/2.Moreover,when restricted to dense-algorithms,we present a best possible dense-algorithm with a competitive ratio of 2.

Effects of Body Weight and Water Temperature on Maximum Food Consumption of JuvenileSebastodes fuscescens (Houttuyn)

Maximum rate of food consumption (C max) was determined for juvenile Sebastodes fuscescens (Houttuyn) at water temperature of 10, 15, 20 and 25℃. The relationships of C max to the body weight (W) at each temperature were described by a power equation: lnC max = a + b lnW. Covariance analysis revealed significant interaction of the temperature and body weight. The relationship of adjusted C max to water temperature (T) was described by a quadratic equation: C max =-0.369 + 0.456T - 0.0117T 2. The optimal feeding temperature calculated from this equation was 19.5℃. The coefficients of the multiple regression estimation relating C max to body weight (W) and water temperature (T) were given in the Table 2.

High-order maximum-principle-preserving and positivity-preserving weighted compact nonlinear schemes for hyperbolic conservation laws

In this paper,the maximum-principle-preserving(MPP)and positivitypreserving(PP)flux limiting technique will be generalized to a class of high-order weighted compact nonlinear schemes(WCNSs)for scalar conservation laws and the compressible Euler systems in both one and two dimensions.The main idea of the present method is to rewrite the scheme in a conservative form,and then define the local limiting parameters via case-by-case discussion.Smooth test problems are presented to demonstrate that the proposed MPP/PP WCNSs incorporating a third-order Runge-Kutta method can attain the desired order of accuracy.Other test problems with strong shocks and high pressure and density ratios are also conducted to testify the performance of the schemes.

改进YOLOv5的无人机航拍图像目标检测算法

针对无人机航拍图像目标检测中目标尺度多样、相似目标众多、目标聚集导致的目标漏检、误检问题,提出了改进YOLOv5的无人机航拍图像目标检测算法DA-YOLO。提出由特征图注意力生成器和动态权重学习模块组成的多尺度动态特征加权融合网络,特征图注意力生成器融合处理不同尺度目标更重要的特征,权重学习模块自适应地调节对不同尺度目标特征的学习,该网络可增强在目标尺度多样下的辨识度从而降低目标漏检。设计一种并行选择性注意力机制(PSAM)添加到特征提取网络中,该模块通过动态融合空间信息和通道信息,加强特征的表达获得更优质的特征图,提高网络对相似目标的区分能力以减少误检。使用Soft-NMS代替YOLOv5中采用的非极大值抑制(NMS)以改善目标聚集场景下的漏检、误检。实验结果表明,改进算法在VisDrone数据集上检测精度达到37.79%,相比于YOLOv5s算法精度提高了5.59个百分点,改进后的算法可以更好地应用于无人机航拍图像目标检测中。

一种求解高密度最大权重团的信息传播算法

在高密度复杂结构带权无向图中,由于信息传递复杂,传统信息传播算法求解最大权重团问题效率较低。利用最大权重独立集与最大权重团的映射关系提出一种针对高密度带权无向图求解最大权重团问题算法,该算法以最大积信息传播算法为框架,将最大权重团的约束条件与置信传播算法迭代方程相结合,设计信息传播算法势函数。同时,将高密度复杂结构带权无向图映射为因子图,并进行去环操作,利用信息传播迭代式进行特征收敛计算,通过迭代收敛后的最大后验概率计算最大权重团最优解。基于不同密度随机图进行实验对比分析,实验结果表明,该算法求解高密度复杂结构带权无向图最大权重问题时非常有效,求解总权值的准确度与求解速率均高于标准置信传播算法。

局部判别损失无监督域适应方法

在无监督域适应任务中,源域和目标域的分布不同,源域数据标签已知,但是目标域的数据标签未知。最大平均差异MMD是一种具有代表性的分布度量方法,广泛应用于源域与目标域之间的分布差异度量。然而,MMD度量及其变种方法通常忽略了样本的类内紧凑性和类间可分离性,降低了特征表达的可判别性。因此,提出局部判别损失无监督域适应方法,从2个方面提升域适应方法的判别能力:(1)重新设计MMD度量方法的权重,解决类别不均衡问题,使难对齐类别在域间分布上保持一致;(2)探索局部对比损失,平衡正样本对和负样本对之间的关系,从而学习到更好的判别性特征。结合域间损失和类间损失,可使同一类样本靠近,不同类样本之间远离。该方法简单有效,即插即用,可扩展至注意力机制的网络结构上。在多个域适应数据集上,该方法的有效性均得到了验证。

基于多应用场景的改进DV-Hop定位模型

针对距离矢量跳(DV-Hop)定位模型定位精度低、优化策略场景依赖性强的问题,提出一种基于函数分析和模拟定参的改进DV-Hop模型——函数修正距离矢量跳(FuncDV-Hop)定位模型。首先,分析DV-Hop模型的平均跳距、距离估计和最小二乘法中的误差原因,引入待定系数优化、阶跃函数分段实验、带等效点的权重函数策略和极大似然估计修正;其次,考虑多应用场景,用控制变量法,分别将总节点数、信标节点比例、通信半径、信标节点数和待测节点数作为变量,设计对照实验;最后,进行仿真定参和整合优化测试两阶段实验,最终的改进策略较原DV-Hop模型的定位精度提高了23.70%~75.76%,平均优化率57.23%。实验结果表明,FuncDV-Hop模型的优化率最高达到了50.73%,与基于遗传算法和神经动力学改进的DV-Hop模型相比,FuncDV-Hop模型的优化率提升了0.55%~18.77%。所提模型不引入其他参量,不增加无线传感器网络(WSN)的协议开销,且有效提高定位精度。

^(18)F-PSMA-1007 PET/CT与mp-MRI对前列腺癌的检测效能及与病理分级的相关性研究

目的探讨18氟标记前列腺特异性膜抗原-1007正电子发射计算机体层摄影(fluorine-18 prostate specific membrane antigen 1007-positron emission tomography/computed tomography,^(18)F-PSMA-1007 PET/CT)与多参数磁共振成像(multi-parameter magnetic resonance imaging,mp-MRI)单独及联合对前列腺癌(prostate cancer,PCa)的检测效能,并比较最大标准化摄取值(maximum standardized uptake value,SUV_(max))、表观弥散系数(apparent diffusion coefficient,ADC)、SUV_(max)/ADC、T1、T2、质子密度(proton density,PD)值与PCa病理分级的相关性。材料与方法回顾性分析我院2020年4月至2022年9月疑似PCa并拟行穿刺活检或手术的患者病例,最终50例患者被纳入研究,其中42例确诊PCa。按照国际泌尿病理协会(International Society of Urological Pathology,ISUP)分级共分为5组,包括中高级别组(分级≥4)25例和低级别组(分级1~3)17例,比较不同分组的SUV_(max)、ADC及SUV_(max)/ADC的差异。采用Spearman相关分析SUV_(max)、ADC、SUV_(max)/ADC、T1、T2、PD值之间的相关性及分别与ISUP分级的相关性。以病理为金标准,分析^(18)F-PSMA-1007 PET/CT与mp-MRI单独及联合对前列腺良恶性的检测效能。绘制受试者工作特征(receiver operating characteristic,ROC)并计算曲线下面积(area under the curve,AUC)、敏感度和特异度,评价SUV_(max)、ADC、SUV_(max)/ADC及联合参数诊断效能,通过DeLong检验比较AUC间的差异。结果PCa高级别组与低级别组的ADC值、SUV_(max)、SUV_(max)/ADC差异均有统计学意义(P均<0.001)。50例患者的ADC值与SUV_(max)呈负相关性(r=-0.516,P<0.05),42例确诊PCa患者的ADC值与ISUP分级呈负相关(r=-0.616,P<0.05),SUV_(max)、SUV_(max)/ADC与ISUP分级呈正相关(r=0.549、0.639,P均<0.05)。20例完成定量磁共振图像编译(magnetic resonance image compilation,MAGiC)序列患者T1、T2、PD值与ISUP分级无相关性(r=0.045、0.202、0.028,P均>0.05);T1、T2值与ADC值呈正相关(r=0.616,r=0.756,P均<0.05),PD值与ADC值呈负相关(r=-0.506,P<0.05)。SUV_(max)与T1、T2、PD值没有明显相关性(r=-0.132,r=-0.422,r=0.230,P均>0.05)。ROC曲线分析显示,SUV_(max)的AUC为0.940,差异有统计学意义(P<0.001),以7.80为临界值,诊断PCa的敏感度为83.33%,特异度为100.00%;ADC值的AUC为0.970,差异有统计学意义(P<0.001),以1.20×10^(-3)mm^(2)/s为临界值,诊断PCa的敏感度为95.24%,特异度为87.50%;SUV_(max)/ADC的AUC为0.970,差异有统计学意义(P<0.001),以6.43×10^(3)为临界值,诊断PCa的敏感度为90.48%,特异度为100%;SUV_(max)和ADC联合的AUC为0.976,差异有统计学意义(P<0.001),以0.85为临界值,诊断PCa的敏感度为90.48%,特异度为100.00%。结论联合^(18)F-PSMA-1007 PET/CT与mp-MRI可以提高对PCa的诊断效能;ADC值、SUV_(max)及SUV_(max)/ADC能够区分低危与中高危PCa。

Data-Based Filters for Non-Gaussian Dynamic Systems With Unknown Output Noise Covariance

This paper proposes linear and nonlinear filters for a non-Gaussian dynamic system with an unknown nominal covariance of the output noise.The challenge of designing a suitable filter in the presence of an unknown covariance matrix is addressed by focusing on the output data set of the system.Considering that data generated from a Gaussian distribution exhibit ellipsoidal scattering,we first propose the weighted sum of norms(SON)clustering method that prioritizes nearby points,reduces distant point influence,and lowers computational cost.Then,by introducing the weighted maximum likelihood,we propose a semi-definite program(SDP)to detect outliers and reduce their impacts on each cluster.Detecting these weights paves the way to obtain an appropriate covariance of the output noise.Next,two filtering approaches are presented:a cluster-based robust linear filter using the maximum a posterior(MAP)estimation and a clusterbased robust nonlinear filter assuming that output noise distribution stems from some Gaussian noise resources according to the ellipsoidal clusters.At last,simulation results demonstrate the effectiveness of our proposed filtering approaches.

未知信号传播速度的TDOA定位方法研究

研究了未知信号传播速度时的到达时间差(Time-Difference-Of-Arrival,TDOA)定位问题,提出两种联合估计信号传播速度和目标位置的定位方法。第一种方法为两步加权最小二乘方法。在该方法中,首先不考虑变量之间关系,得到一个关于未知变量的初始加权最小二乘估计。为改进第一步估计的性能,第二步考虑第一步估计中变量之间的关系,将其转换为一个标准的广义信赖域子问题,最终获得更高精度的估计性能。第二种方法为半正定松弛方法,通过构建非线性非凸加权最小二乘问题,然后利用半正定松弛技术将其松弛为凸的半正定规划问题,容易求解。仿真结果表明,所提两种方法均能够在高斯噪声下,且噪声不太大时达到克拉美-罗界。

熵权模糊改进法在宁夏黄河水质评价中的应用

为了解宁夏境内黄河流域的水质状况,选取银古公路桥、平罗黄河大桥、中卫香山湖、吴忠金沙湾、固原沟圈5个监测断面在2021年4月、6月、8月、12月4个时期的溶解氧、高锰酸盐指数、氨氮、总磷、总氮5项评价指标,共获得80组水样。针对熵权模糊评价法在权重趋于1时其微小变化会引起熵权成倍变化的病态问题和由于最大隶属度原则不能普遍适用而导致评价失效的问题,提出了一种熵权模糊改进的评价方法,通过改进熵权法确定权重和加权等级,建立改进的模糊综合评价模型,对水质样品进行水质分析和综合评价。结果表明,在权重分配方面,相比于隶属度法和传统的熵值法,改进熵权法分配的权重更为合理;在水质评价方面,加权等级法评定出的结果更符合研究区域的真实水质状况;宁夏境内黄河流域Ⅰ类和Ⅱ类水质占比超过90%,水质状况良好,5个断面4个时期的水质综合评价结果显示,12月份的水质最优,Ⅰ类水质占比85%;对水质的时空分析显示氨氮为全局性污染因子,溶解氧为区域性污染因子。改进的模糊综合评价模型对宁夏黄河水质的评价结果更为客观合理。

NSD随机阵列最大加权和的矩收敛性及应用

该文利用NSD随机阵列的极大值矩不等式研究了NSD随机阵列最大加权和的矩收敛性,相关定理改进且推广了已有关于AANA变量的结果 .此外,作为主要结果的应用,进一步研究了非参数回归模型加权估计量的矩相合性和弱相合性.

基于贝叶斯最大熵的黄河三角洲土壤含盐量空间分布预测

黄河三角洲土壤盐渍化问题是制约当地农业生产和生态稳定的关键因素。为了准确掌握盐渍土的空间分布,提高土壤含盐量的空间预测精度,本研究根据2022年5月黄河三角洲的193个采样点两个深度土壤含盐量分析数据,结合高程和Landsat 9遥感影像等数据,采用地理加权回归(Geographically weighted regression,GWR)模型构造区间型软数据,进而建立贝叶斯最大熵(Bayesian maximum entropy,BME)模型对研究区土壤含盐量的分布进行了预测,并同传统的地统计模型普通克里格(Ordinary kriging,OK)及GWR模型的预测结果进行了比较。结果表明:BME模型对土壤含盐量的预测精度高于另外两种模型。与OK模型相比,BME模型的RMSE在土壤表层(0~30 cm)和底层(90~100 cm)分别降低25%和21%,R^(2)分别提高0.543 2和0.352 7。BME模型作为本研究最佳土壤含盐量空间预测模型,展现了多源数据整合及非线性估计的优势。黄河三角洲表层土壤盐渍化率(88%)高于底层(68%),土壤含盐量大体呈现由西南向东北递增的趋势,沿海地区大于内陆地区,黄河三角洲北部是整个区域盐渍化最为严重的地区。

采用MPC-MTPA的新能源车永磁同步电机控制研究

针对传统永磁同步电机控制器结构复杂,负载突变时动态响应速度慢,导致电机抗干扰能力差、工作效率低的问题,提出了一种模型预测控制(model predictive control,MPC)融合最大转矩比控制(maximum torque per ampere,MTPA)的电机非级联控制方法。构建MPC和MTPA控制器,利用MPC中的代价函数同时完成转矩控制和MTPA控制,在代价函数中引入动态分配的权重系数优化非级联MPC-MTPA控制器。仿真和试验结果表明:相比传统电机控制方法,提出的变权重系数非级联MPC-MTPA控制方法的稳定性提高约26%,动态响应特性提高约50%。

An Efficient Proximal Point Algorithm for Unweighted Max-Min Dispersion Problem

In this paper, we first reformulate the max-min dispersion problem as a saddle-point problem. Specifically, we introduce an auxiliary problem whose optimum value gives an upper bound on that of the original problem. Then we propose the saddle-point problem to be solved by an adaptive custom proximal point algorithm. Numerical results show that the proposed algorithm is efficient.

基于MWST-CS-K2算法的贝叶斯网络结构学习

针对K2算法依赖最大父节点数和节点顺序的不足,提出了一种改进的贝叶斯网络结构学习算法(MWST-CS-K2)。该算法先通过构造最大支撑树(MWST)得到最大父节点数;再利用变量间的关联度和更新系数对加边、减边和反转边进行规则设定,通过改进的布谷鸟算法对鸟巢位置进行寻优,应用广度优先搜索策略搜索遍历得到节点顺序;最后将最大父节点数和节点顺序作为K2算法的输入搜索得到最终网络。实验表明,所提出的MWST-CS-K2算法在标准的ASIA、SACHS和CHILD网络数据测试中的平均正确边比率分别达到了97.3%、87.7%和95.6%,学习效果优于其他对比算法,获得的网络结构和标准的网络结构最为相似。

A Maximum Likelihood TOA Based Estimator For Localization in Heterogeneous Networks

In this paper, we exploit the concept of data fusion in hybrid localization systems by combining different TOA (Time of Arrival) observables coming from different RATs (Radio Access Technology) and characterized by different precisions in order to enhance the positioning accuracy. A new Maximum Likelihood estimator is developed to fuse different measured ranges with different variances. In order to evaluate this estimator, Monte Carlo simulations are carried out in a generic environment and Cramer Rao Lower Bounds (CRLB) are investigated. This algorithm shows enhanced positioning accuracy at reasonable noise levels comparing to the typical Weighted Least Square estimator. The CRLB reveals that the choice of the number, and the configuration of Anchor nodes, and the type of RAT may enhance positioning accuracy.