Parameter optimization of gravity density inversion based on correlation searching and the golden section algorithm

For density inversion of gravity anomaly data, once the inversion method is determined, the main factors affecting the inversion result are the inversion parameters and subdivision scheme. A set of reasonable inversion parameters and subdivision scheme can, not only improve the inversion process efficiency, but also ensure inversion result accuracy. The gravity inversion method based on correlation searching and the golden section algorithm is an effective potential field inversion method. It can be used to invert 2D and 3D physical properties with potential data observed on flat or rough surfaces. In this paper, we introduce in detail the density inversion principles based on correlation searching and the golden section algorithm. Considering that the gold section algorithm is not globally optimized. we present a heuristic method to ensure the inversion result is globally optimized. With a series of model tests, we systematically compare and analyze the inversion result efficiency and accuracy with different parameters. Based on the model test results, we conclude the selection principles for each inversion parameter with which the inversion accuracy can be obviously improved.

Using Genetic Algorithms to Improve the Search of the Weight Space in Cascade－Correlation Neural Network

In this paper, we use the global search characteristics of genetic algorithms to help search the weight space of the neurons in the cascade-correlation architecture. The cascade-correlation learning architecture is a technique of training and building neural networks that starts with a simple network of neurons and adds additional neurons as they are needed to suit a particular problem. In our approach, instead ofmodifying the genetic algorithm to account for convergence problems, we search the weight-space using the genetic algorithm and then apply the gradient technique of Quickprop to optimize the weights. This hybrid algorithm which is a combination of genetic algorithms and cascade-correlation is applied to the two spirals problem. We also use our algorithm in the prediction of the cyclic oxidation resistance of Ni- and Co-base superalloys.

Effects of initial states on the quantum correlations in the generalized Grover search algorithm

We investigate the correlations between two qubits in the Grover search algorithm with arbitrary initial states by numerical simulation.Using a set of suitable bases,we construct the reduced density matrix and give the numerical expression of correlations relating to the iterations.For different initial states,we obtain the concurrence and quantum discord compared with the success probability in the algorithm.The results show that the initial states affect the correlations and the limit point of the correlations in the searching process.However,the initial states do not influence the whole cyclical trend.

A NEW SYSTEM DYNAMIC EXTREMUM SELF-SEARCHING METHOD BASED ON CORRELATION ANALYSIS

Objective To propose a new dynamic extremum self searching method, which can be used in industrial processes extremum optimum control systems, to overcome the disadvantages of traditional method. Methods This algorithm is based on correlation analysis. A pseudo random binary signal m sequence u(t) is added as probe signal in system input, construct cross correlation function between system input and output, the next step hunting direction is judged by the differential sign. Results Compared with traditional algorithm such as step forward hunting method, the iterative efficient, hunting precision and anti interference ability of the correlation analysis method is obvious over the traditional algorithm. The computer simulation experimental given illustrate these viewpoints. Conclusion The correlation analysis method can settle the optimum state point of device operating process. It has the advantage of easy condition , simple calculate process.

Employment of predictive search algorithm in digital image correlation

A predictive search algorithm to estimate the size and direction of displacement vectors was presented.The algorithm decreased the time of calculating the displacement of each pixel.In addition,the updating reference image scheme was used to update the reference image and to decrease the computation time when the displacement was larger than a certain number.In this way,the search range and computational complexity were cut down,and less EMS memory was occupied.The capability of proposed search algorithm was then verified by the results of both computer simulation and experiments.The results showed that the algorithm could improve the efficiency of correlation method and satisfy the accuracy requirement for practical displacement measuring.

Composite Recommendation of Artworks in E-Commerce Based on User Keyword-Driven Correlation Graph Search

With the ever-increasing diversification of people’s interests and preferences,artwork has become one of the most popular commodities or investment goods in E-commerce,and it increasingly attracts the attention of the public.Currently,many real-world or virtual artworks can be found in E-commerce,and finding a means to recommend them to appropriate users has become a significant task to alleviate the heavy burden on artwork selection decisions by users.Existing research mainly studies the problem of single-artwork recommendation while neglecting the more practical but more complex composite recommendation of artworks in E-commerce,which considerably influences the quality of experience of potential users,especially when they need to select a set of artworks instead of a single artwork.Inspired by this limitation,we put forward a novel composite recommendation approach to artworks by a user keyword-driven correlation graph search named ART_(com-rec).Through ART_(com-rec),the recommender system can output a set of artworks(e.g.,an artwork composite solution)in E-commerce by considering the keywords typed by a user to indicate his or her personalized preferences.Finally,we validate the feasibility of the ART_(com-rec) approach by a set of simulated experiments on a real-world PW dataset.

A Correlation Tracking Algorithm Based on Template Partition Motion Estimation

A correlation tracking algorithm based on template partition motion estimation proposed for improving real time performance of the conventional correlation matching algorithms. The target trajectory fitted using the least square with equal space in whole interval and the target prediction point is found out. According to the requirements of block motion estimation(BME) algorithm,the template divided into some macro blocks. The searching process is conducted by using diamond search algorithm around the prediction point and the optimal motion vector of each block is calculated. A point corresponding to the motion vector with the best matching is taken as a rough matching point of the template. The relation of relative position between the block with matching point and the searching area determined to decide whether to conduct precise matching search or to construct a new search area in the gradient direction. The target tracking experiment results show that over 70% time cost can be reduced caompared with the conventional correlation matching algorithm based on full search method.

Improvement of a new rotation function for molecular replacement by designing new scoring functions and dynamic correlation coefficient

A previously published new rotation function has been improved by using a dynamic correlation coefficient as well as two new scoring functions of relative entropy and mean-square-residues to make the rotation function more robust and independent of a specific set of weights for scoring and ranking. The previously described new rotation function calculates the rotation function of molecular replacement by matching the search model directly with the Patterson vector map. The signal-to-noise ratio for the correct match was increased by averaging all the matching peaks. Several matching scores were employed to evaluate the goodness of matching. These matching scores were then combined into a single total score by optimizing a set of weights using the linear regression method. It was found that there exists an optimal set of weights that can be applied to the global rotation search and the correct solution can be ranked in the top 100 or less. However, this set of optimal weights in general is dependent on the search models and the crystal structures with different space groups and cell parameters. In this work, we try to solve this problem by designing a dynamic correlation coefficient. It is shown that the dynamic correlation coefficient works for a variety of space groups and cell parameters in the global search of rotation function. We also introduce two new matching scores： relative entropy and mean-square-residues. Last but not least, we discussed a valid method for the optimization of the adjustable parameters for matching vectors.

一种参数自适应VMD应用于轴承故障特征提取

针对传统的变分模态分解(VMD)需要预先设置模态个数和惩罚参数,提出了一种基于麻雀搜索算法(SSA)的参数自适应VMD方法。首先,引入一种新的测量指标-相关脉冲,该指标能反映出原始信号与分解模态之间的相关性,并且能有效突出包含丰富信息的模态。其次,基于相关脉冲指标,采用麻雀搜索算法选择最优VMD分解参数。最后,通过最大相关脉冲指标对模态分量进行分析,利用希尔伯特包络谱进行频谱分析。此外,将故障轴承放在轴承寿命试验台上进行仿真验证,实验结果表明该方法在轴承故障特征提取上具有可行性。

基于改进VMD-MCKD和深度残差网络的风机齿轮箱故障诊断

行星齿轮箱是风电机组传动系统中的重要部件,其运行工况复杂,背景噪声大,导致齿轮早期故障信号微弱且极易受背景噪声的影响。针对风电机组齿轮箱早期故障特征难以有效提取,齿轮故障难以识别的问题,提出一种风机齿轮箱故障诊断方法。首先,通过变分模态分解算法(variational mode decomposition,VMD)分解风机齿轮箱原始振动信号,获得振动信号故障的最优模态分量;接着,利用最大相关峭度解卷积算法(maximum correlated kurtosis decnvolution,MCKD)通过解卷积重构最优模态分量,削弱背景噪声增强故障冲击成分,获得故障特征;同时利用麻雀搜索算法(sparrow search algorithm,SSA)优化惩罚因子α、模态分解个数K、滤波器阶数L和反褶积周期T等参数,提升振动信号故障特征提取的准确度;最后,构建基于深度残差网络(deep residual network,ResNet)的齿轮箱故障诊断模型,建立齿轮箱故障特征与类别的非线性映射关系,实现风机齿轮箱故障分类识别。实验结果表明,所提风机齿轮箱故障诊断方法的准确率达到97.48%,相较其他方法在信号特征提取和故障诊断效率方面有明显提高。

基于随机游走麻雀搜索算法的多特征结构尺寸熔融沉积成型工艺参数优化

在熔融沉积成型过程中,打印参数对成型样件精度有着重要影响.为了提高整体尺寸精度,采用随机游走的麻雀算法获得最优实验方案.首先,以熔融沉积成型的分层厚度、喷头温度、打印速度和填充率为实验变量设计4因素4水平的正交试验;然后,以样件不同特征结构尺寸的相对误差为优化对象,使用田口-灰色关联法对实验数据进行处理;最后,通过随机游走的麻雀算法计算最优参数方案.结果表明,相比常用的田口-灰色关联法,采用优化后工艺参数成型样件的综合尺寸精度提高了20%,灰色关联度提高了27%.

基于GPU集群的空间VLBI射电源条纹搜索研究

探月工程四期将发射嫦娥七号“鹊桥二号”中继星,搭载4.2 m口径的抛物面望远镜,与地面射电望远镜构成首个月轨空间VLBI。空间VLBI受中继星轨道扰动和星载设备时延等因素影响,射电源的预报时延模型无法引导相关处理机正常工作,需通过基于实测信号条纹搜索的方法找到符合要求的高精度时延模型。为此,提出一种基于GPU集群的空间VLBI射电源条纹搜索算法,研究在GPU集群上负载均衡的任务分配方法,并行实现相关处理模块和残余值搜索模块。通过RadioAstron 1 min观测数据验证,相较于目前CPU集群平台,GPU集群处理速度提升了27.0倍。

基于相关性分析和SSA-BP神经网络的铝合金电阻点焊质量预测

基于电阻点焊过程中工艺信号特征,在不同间距、不同间隙和不同间距与间隙3种条件下,引入相关性分析方法分析工艺信号与熔核直径之间的相关性,并建立基于麻雀搜索算法-BP神经网络(sparrow search algorithmback propagation neural network,SSA-BP)的电阻点焊质量预测模型,将功率、焊接电流、焊接电压和动态电阻作为预测模型输入特征.结果表明,经麻雀搜索算法优化后的BP神经网络在测试集上的决定系数R2、均方误差(meansquare error,MSE)、均方根误差(root mean square error,RMSE)和平均绝对误差(mean absolute error,MAE)分别为0.95,1.55,1.24和0.90,均优于BP模型.获得了功率、焊接电流、焊接电压和动态电阻与熔核直径的映射关系,可为焊接的工艺参数设计提供依据.

基于GBDT特征提取与Tent-ASO-BP网络的铣刀磨损量预测

为了提高机械加工过程中小样本刀具磨损量监测的准确性,提出一种基于Pearson+GBDT特征提取、Tent混沌映射和原子搜索算法(ASO)优化BP神经网络(Tent-ASO-BP)的刀具磨损量预测模型。针对BP神经网络特征选择及参数选择难题,提出了基于Pearson+GBDT的双层过滤式特征筛选方式求取网络输入特征,并使用Tent混沌映射改进原子搜索算法(ASO)对BP神经网络最优权值和阈值进行求解。通过实验证明:Tent混沌映射改善了ASO,避免ASO陷入局部极值和过早收敛,即通过交叉验证证明Tent-ASO优化BP神经网络训练模型精度较ASO高。同时,验证了梯度提升决策树(GBDT)能够筛选出用于刀具磨损值映射的一组特征,且特征筛选能力强于同类算法Light GBM、Catboost、决策树、随机森林。

基于CRITIC和多策略秃鹰优化BiLSTM的水质预测研究

科学有效地预测水质对于水环境的可持续发展和人类健康具有重要意义,为此以固原市某黄河断面的水质监测数据为研究对象,提出了基于指标客观性的权重赋权(Criteria Importance Though Intercriteria Correlation,CRITIC)法和改进的秃鹰搜索(Improved Bald Eagle Search,IBES)算法优化双向长短时记忆网络(Bidirectional Long Short-Term Memory Network,BiLSTM)的组合水质等级预测模型。首先,采用CRITIC法确定各水质指标的权重,加权求和获得一项综合水质指标,从而提出一种改进的水质评价指标体系,以为BiLSTM提供更丰富、更可靠的水质特征信息。其次,在训练过程中引入Logistic映射和莱维飞行策略,并设计交叉共享及准反向搜索策略优化秃鹰搜索(Bald Eagle Search,BES)算法,以提升其种群多样性,增强寻优能力。最后,通过IBES算法迭代寻找BiLSTM的最佳学习率、隐藏层节点数以及正则化系数的超参数组合,进一步提高其预测水平。结果显示:与IBES-BiLSTM、BES-BiLSTM、GA-BiLSTM、PSO-BiLSTM和BiLSTM等模型相比,CRITIC-IBES-BiLSTM模型进行水质等级预测的准确率、精准率、召回率及F_(1)均最高,且具有更好的稳定性。

基于优化VMD-MCKD和谱峭度的滚动轴承复合故障诊断

针对滚动轴承振动信号中复合故障特征难以准确提取而导致故障诊断困难的问题,提出一种基于优化变分模态分解(VMD)和最大相关峭度解卷积(MCKD)结合快速谱峭度算法的滚动轴承复合故障诊断方法。利用改进麻雀搜索算法(ISSA)优化VMD和MCKD的参数,使用优化后的VMD对复合故障信号进行分解,并根据峭度准则筛选有效本征模态函数(IMF)进行信号重构,使用优化后的MCKD对重构信号进行解卷积与故障特征增强,并对解卷积信号进行包络谱分析提取故障特征频率。利用快速谱峭度算法对未提取出故障特征频率的解卷积信号进行处理,得到故障信息最丰富的频带参数并进行带通滤波处理。最后,对滤波后的信号进行包络谱分析,提取故障特征频率,从而实现故障诊断。仿真及实验结果表明:所提方法能有效分离复合故障并提取出故障特征频率,有效实现了复合故障诊断。

考虑物料关联的仓库区位分配优化研究

货位分配(storage location assignment problem,SLAP),即在存储区域为物料分配货位的过程。当仓库布局、拣货路径、订单组合等其他因素确定时,货位分配策略对订单拣货效率有很大影响。本文研究实际生产型仓库中的关联物料区位分配问题。生产中使用的相对稳定的BOM(bill of material)使得仓库中的物料具有稳定的相关性,因此,本文考虑将具有需求关联的物料存储在同一区域,以尽可能地减少在拣选物料时所需要的区域访问次数。此外,该仓库还存在两个重要特征,即存在两类不同尺寸货架构成的两类不同容量的区域及采用严格的重物下置原则。本文建立了以最小化区域访问次数为目标的数学规划模型,给出了求解该问题的一种聚类启发式方法与自适应大邻域搜索算法(adaptive large neighborhood search,ALNS),并设计了能够反映物料关联特征的小规模和大规模算例用于测试两种算法的性能。将两个算法结果与随机策略、CPLEX求解结果对比,结果显示聚类启发式方法与ALNS在大规模算例中表现明显优于随机策略和CPLEX的求解结果。

基于改进麻雀搜索算法和支持向量机的边坡稳定性

边坡失稳是由多种因素共同作用的结果,常规的数学模型难以准确预测。为提高边坡稳定性预测精度,采用多策略融合改进麻雀搜索算法(improved sparrow search algorithm,ISSA)优化支持向量机(support vector machine,SVM),进而建立边坡稳定性预测模型(ISSA-SVM模型)。将重度、黏聚力、内摩擦角、边坡角、边坡高、孔隙压力比6项因素作为输入特征,将边坡稳定性状态作为输出结果,进而预测边坡稳定性。选取中外工程实例建立边坡数据库,将ISSA-SVM模型与SSA-SVM模型进行对比分析,通过灰色关联度分析法(grey relation analysis,GRA)进行敏感性分析。结果表明:ISSA-SVM模型预测精度更高、泛化能力更强,黏聚力和内摩擦角是对边坡稳定性最为敏感的因子。所提ISSA-SVM模型不仅能够准确地预测边坡稳定状态,还可以为其他领域相关问题提供参考。

随机森林法对北京移动用户体验影响因素研究

目的/意义:移动信息技术是当今社会赖以需求的。提高客户满意度等是提升网络服务质量的最佳方法,也是共同推动移动网络高质量可持续发展的关键步骤之一。方法/过程:使用斯皮尔曼和随机森林对数据进行相关性和量化分析。对影响因素进行研究分析建立数学模型,进行调参处理。得到预测结果。结果/结论:结果表明影响移动用户体验的因素对客户打分影响程度的量化分析和结果。模型精确度89%,准确度强于其余模型。

一种M-ary扩频信号扩频序列集估计方法

针对M-ary扩频信号的扩频序列集估计问题,提出一种基于序贯相关检测的扩频码集估计方法。首先,对同步信号样点进行划分,并利用所分向量间的相关特性,通过迭代操作得到码集中某一个元素的估计;然后,将对应于该元素的样点向量从接收信号向量集合中剔除,并重复上述步骤,以序贯的方式得到扩频码集的一个估计;最后,利用原始信号对该码集的各个元素进行检验,剔除码集中的虚假元素从而得到最终估计结果。该方法可以实现M-ary扩频信号的扩频序列集估计,并且不受扩频码集类型的限制,由于避免了二维搜索以及K均值聚类的迭代操作,文中方法具有比文献方法更快的计算速度。