灰色理论在干旱预测中的应用

介绍灰色系统理论及其建模原理,利用沁河流域上游沁源气象站42年的实测降水量资料建立灰色预测GM(1,1)灾变模型,对干旱灾害进行预测,经残差、后验差检验等分析,模型精度较高平均达98.8%.并对实测资料进行检验,效果较理想,为沁河流域上游抗旱及供水提供必要的预测信息.     

基于? -SVR算法的隧道围岩位移演化规律预测

针对目前广泛应用的灰色理论、遗传算法(GA)和人工神经网络(ANN)等方法预测隧道稳定性的缺陷，提出应用稳健性能较好的? -SVR（support vector machine）算法对非对称连拱隧道围岩位移演化规律进行预测研究。应用加速混合遗传算法搜索? -SVR最优参数，以提高? -SVR的预测能力。将预测结果与灰色理论、BP神经网络预测结果进行比较，显示? -SVR算法学习和预测精度高。     

河流年径流量的R/S灰色预测

径流过程具有分形和灰色特征。基于此,将R/S分析与灰色系统理论相结合,提出了R/S灰色预测模型以预报河流年径流量。针对黑河正义峡水文站60a(1949~2011年)的年径流量资料,首先进行R/S分析,确定径流量序列的Hurst指数H和平均循环周期T;然后在一个周期内进行年径流量灰色预测。结果表明:R/S灰色预测结果的精度明显高于直接进行灰色预测的精度。该方法拓宽了分形和灰色理论在径流过程研究的应用范围,为河流径流量的科学预测提供了一种新方法。     

未确知测度BP神经网络模型在黑河流域水质预测中的应用

水质预测是水环境规划、评价和管理的重要依据,对促进水资源可持续利用及生态发展具有重要意义。针对水质预测中各项因子的不确定性,基于未确知测度理论(unascertained measure,UM),采用改变网络初值的方法,对BP神经网络加以改进,并利用黑河流域莺落峡水文站1998~2011年的水质监测资料进行分析和预测。以挥发酚为参考序列,用灰色关联方法分析参考序列与其他因子的关联度,并最终确定BP网络的输入节点为CODmn、DO、SO42-、Cr6+以及挥发酚,输出节点为挥发酚,从而建立UMBP模型。分析结果表明,UM-BP预测模型比标准的BP神经网络模型具有更高的预测精度。因此,该模型应用于黑河流域水质预测是可行的。     

基于因子分析的LM-BP网络的盾构施工地表沉降预测

盾构施工引起的地表沉降量的预测与评价具有很强的随机性和不确定性,用常规的数学模型难以准确表达。本文采用因子分析法对影响盾构施工引起地表沉降的9个因素进行了主因子的提取,得到4个主因子。以此数据集对BP神经网络进行训练,并用Levenberg-Marquardt(LM)算法对BP神经网络进行了优化,建立了基于因子分析的LM-BP神经网络预测模型。结合工程实例将预测结果与LM-BP神经网络模型的预测结果进行比较分析,表明本预测模型具有较高的准确性,符合实际工程的需要。     

基于贝叶斯正则化改进BP神经网络的页岩气有机碳含量预测模型

页岩气总有机碳(TOC)含量是评价岩性气藏的关键指标,受复杂地质及岩芯采集等多种因素的影响,常规室内测试分析获得的TOC含量的数据有限且结果有失准确。为合理准确预测页岩气TOC含量,本文首先通过对页岩气储层TOC含量测井资料综合分析选取8条测井曲线,并结合主成分分析法(Principal Component Analysis,PCA)提取四个主成分;其次基于贝叶斯正则化(Bayesian Regularization)改进的BP神经网络方法建立页岩气TOC含量预测的BR-BP模型;最后利用该模型对研究区A区页岩气TOC含量进行预测,并与常规的LM-BP神经网络模型的预测结果进行对比。结果表明：BR-BP模型有较强的非线性拟合能力,能够真实地反映出页岩气TOC含量与各测井参数之间的非线性关系,其模型预测结果与实际值基本吻合,与常规的LM-BP神经网络模型相比,其数据敏感性增强,预测精度有所提高,该研究方法具有一定的理论意义和参考价值,为我国TOC含量预测提供了一种新的技术方法和手段。     

基于LM-BP神经网络的黑河龙电渠流量推测研究

本文通过分析总结内陆流域黑河水系干流莺落峡龙电渠测验断面水文特点以及流量测验现存问题,提出应用matlab编程软件设计LM-BP神经网络预测模型的方法;采用有导师训练方式,并获取历史实测流量测验资料中点流速和平均流速数据做为样本集,对网络模型的设计、训练和测试做了研究和分析。根据网络仿真和测试结果可以得出,在误差允许范围之内,黑河龙电渠测验断面流量测验采用"五线一点"法的测验方案和LM-BP神经网络模型推测流量的方法,其推测值精度较高,且具有一定的推广应用价值。     

地面沉降的GM(2,1)模型预测研究

运用灰色理论建立地面沉降的GM(2,1)模型,其中非等时距位移序列采用拉格朗日插值函数转变为等时距序列。采用非等时距GM(2,1)及GM(1,1)模型对西安市地面沉降观测点进行安全预测。计算结果表明,非等时距GM(2,1)模型预测地面沉降精度总体较GM(1,1)模型高,预测结果与实际吻合较好。     

灰色新陈代谢GM(1,1)模型在滑坡变形预测中的应用

由于常规GM(1,1)模型用于预测时,精度较高的仅仅是最近的几个数据,越往未来发展,该模型的预测精度就越弱。针对常规GM(1,1)模型存在的不足,文章建立了灰色新陈代谢GM(1,1)滑坡预测模型,并利用该模型对向加坡滑坡和链子崖危岩体GA监测点位移变形进行了预测。结果表明,灰色新陈代谢GM(1,1)模型精度较高,预测误差较小,有很好的利用价值。     

利用人工神经网络对岩浆侵入范围的预测

以某矿三维地震勘探资料为依据,建立了岩浆侵入的正演模型,在应用灰色系统理论中的灰色关联分析方法进行属性优化的基础上,借助人工神经网络理论中的BP网络方法,建立了岩浆侵入范围预测的BP网络模型.通过误差分析证明,将灰色关联分析与BP网络结合起来进行岩浆侵入范围的预测是可行的.     

GA-BP神经网络模型在彬长矿区测井岩性识别中的应用

为提高测井岩性识别的自动化程度和地质解释精度,在分析遗传算法(Genetic Algorithm,简称GA)与误差反向传播算法(Back-Propagation,简称BP)各自特性的基础上,针对BP算法在反演中测井数据识别样本大以及BP算法本身存在的缺陷,提出了利用GA算法来同时优化BP神经网络的结构和连接权值的解决方案,建立了基于GA优化BP神经网络的测井数据岩性识别模型。该模型通过彬长矿区实际数据的检验,获得了较高的识别速度和准确率。      

Predicting the Shear Behavior of Cemented and Uncemented Carbonate Sands Using a Genetic Algorithm-Based Artificial Neural Network

Carbonate sands that are known as problematic soils, have some unusual features like particle crushability and compressibility that discriminate their behavior from other types of soil. Because of their vast diversity, they have a wide range of mechanical behavior. In recent decades, there have been many attempts to model the mechanical behavior of carbonate sands but all these efforts have been focused on experimental and case studies of some especial sands and there is still no unique way which can appraises all types of carbonate sands behavior and describes their various aspects. In this paper, a new approach is presented based on the integration of Genetic Algorithm (GA) into an Artificial Neural Network (ANN) to predict the shear behavior of carbonate sands. In the proposed approach, the GA was utilized to optimize the connection weights of the ANN. The network was trained and tested using a comprehensive set of triaxial tests that were carried out on three different carbonate sands in both grouted and ungrouted (cemented and uncemented) condition. The network prediction was then compared to the experimental results and it was concluded that the GA-based ANN has a good potential in predicting the behavior and generalizing the training data to simulate new unseen data.     

Inversion of magnetic anomalies due to isolated thin dike-like sources using artificial neural networks

A new approach is proposed to interpret magnetic anomalies caused by isolated thin dike-like causative targets. The approach is essentially based on utilizing artificial neural network (ANN) inversion for estimating the problem parameters. Particularly, the modular neural network (MNN) is used for the inversion process in order to quantitatively interpret the magnetic anomalies. The MNN inversion has been first tested on a synthetic data with and without random white Gaussian noise. The effect of random noise has been clearly investigated where it showed that the approach provided satisfactory results. Furthermore, three field examples have been inverted in order to investigate the applicability of the proposed approach. The results showed good agreement with the techniques that have been stated in the literatures.     

Uncertainty assessment of the multilayer perceptron (MLP) neural network model with implementation of the novel hybrid MLP-FFA method for prediction of biochemical oxygen demand and dissolved oxygen: a case study of Langat River

Accurate prediction of the chemical constituents in major river systems is a necessary task for water quality management, aquatic life well-being and the overall healthcare planning of river systems. In this study, the capability of a newly proposed hybrid forecasting model based on the firefly algorithm (FFA) as a metaheuristic optimizer, integrated with the multilayer perceptron (MLP-FFA), is investigated for the prediction of monthly water quality in Langat River basin, Malaysia. The predictive ability of the MLP-FFA model is assessed against the MLP-based model. To validate the proposed MLP-FFA model, monthly water quality data over a 10-year duration (2001–2010) for two different hydrological stations (1L04 and 1L05) provided by the Irrigation and Drainage Ministry of Malaysia are used to predict the biochemical oxygen demand (BOD) and dissolved oxygen (DO). The input variables are the chemical oxygen demand (COD), total phosphate (PO₄), total solids, potassium (K), sodium (Na), chloride (Cl), electrical conductivity (EC), pH and ammonia nitrogen (NH₄-N). The proposed hybrid model is then evaluated in accordance with statistical metrics such as the correlation coefficient (r), root-mean-square error, % root-mean-square error and Willmott’s index of agreement. Analysis of the results shows that MLP-FFA outperforms the equivalent MLP model. Also, in this research, the uncertainty of a MLP neural network model is analyzed in relation to the predictive ability of the MLP model. To assess the uncertainties within the MLP model, the percentage of observed data bracketed by 95 percent predicted uncertainties (95PPU) and the band width of 95 percent confidence intervals (d-factors) are selected. The effect of input variables on BOD and DO prediction is also investigated through sensitivity analysis. The obtained values bracketed by 95PPU show about 77.7%, 72.2% of data for BOD and 72.2%, 91.6% of data for DO related to the 1L04 and 1L05 stations, respectively. The d-factors have a value of 1.648, 2.269 for BOD and 1.892, 3.480 for DO related to the 1L04 and 1L05 stations, respectively. Based on the values in both stations for the 95PPU and d-factor, it is concluded that the neural network model has an acceptably low degree of uncertainty applied for BOD and DO simulations. The findings of this study can have important implications for error assessment in artificial intelligence-based predictive models applied for water resources management and the assessment of the overall health in major river systems.     

基于随机补片和DeepLabV3+的建筑物遥感图像变化检测

为有效解决传统遥感图像变化检测预处理复杂的问题,提出一种基于随机补片和DeepLabV3+的建筑物遥感图像变化检测方法。以ResNet50特征提取网络为基础,创建DeepLabV3+语义分割网络,并在图像和标签中创建大小为224像素×224像素的随机补片作为网络输入,训练建筑物提取网络;修改建筑物提取网络输入层为6通道,通过矩阵运算将两期遥感图像转换为一幅6通道非RGB图像,利用转换后的非RGB图像进行网络训练并验证变化检测精度。实验1利用ENVI5.3软件,采用马氏距离法进行变化检测;实验2采用改进的U-Net网络和随机补片,完成网络训练和精度验证;实验3使用实验2的训练数据和验证数据,采用随机补片和DeepLabV3+网络进行变化检测网络训练及精度验证。实验结果表明,该方法实验1、实验2、实验3建筑物变化检测平均交并比分别为24.43%、83.14%、89.90%,边界轮廓匹配分数分别为61.47%,80.24%、96.51%。     

基于集成学习的阶跃型滑坡阶跃点判别分析

针对阶跃型滑坡阶跃点识别和预测难的问题,提出了一种基于聚类分析和集成学习的阶跃型滑坡阶跃点识别和判别模型。以三峡库区八字门滑坡ZG110钻孔2010年4月至2016年12月80个滑坡位移、库水位和降雨数据为例,通过聚类分析方法识别滑坡累积位移-时间曲线中的阶跃点和平稳点,并利用K均值聚类分析检验分类结果的准确性。基于灰色关联确定了滑坡位移的最佳诱发因素,结合随机森林模型建立阶跃型滑坡阶跃点判别模型并利用八字门滑坡ZG111钻孔验证该模型的准确性。模型阶跃点和平稳点的识别准确率均达90%以上,表明该方法在阶跃型滑坡识别中具有较好的适用性,可为阶跃型滑坡的预测提供参考。     

软基沉降的BP神经网络和灰色系统联合预测

使用BP神经网络插值方法对灰色数据进行了预处理,进而建立了预测软基沉降量的BP神经网络和灰色系统联合模型.实例分析表明,该模型短期沉降预测结果的最大相对误差小于2%,最终沉降预测结果的相对偏差小于5%,且灰色预测时取后期沉降瘦导颇算结果准确度高于取前期沉降数据的计算结果准确度.     

基于粗糙集理论的遗传-RBF神经网络在岩爆预测中的应用

岩爆发生机制复杂,影响因素较多,通过粗糙集理论中的属性约简和条件属性重要性评价,确定特定地质条件下岩爆的主要影响因素,删除冗余数据。使用遗传算法(GA)优化径向基函数（radial basis function,简称RBF）神经网络参数,通过RBF神经网络隐层单元将低维模式输入变换到高维空间内,拟合影响因子和岩爆等级之间的非线性映射关系,建立基于粗糙集理论的遗传-RBF神经网络岩爆预测模型,目前未见其在地下洞室岩爆预测中应用。在根据工程实际情况选取多个理论判据的基础上,将建立的预测模型应用于实际工程的岩爆预测问题,并与实际岩爆发生情况进行对比分析。结果证明,该方法的评价结果与实际情况较为吻合,对后期施工有较好的指导作用。     

Modeling the Population Effects of Hypoxia on Atlantic Croaker (<Emphasis Type="Italic">Micropogonias undulatus</Emphasis>) in the Northwestern Gulf of Mexico: Part 1—Model Description and Idealized Hypoxia

We developed a spatially explicit, individual-based model to analyze how hypoxia effects on reproduction, growth, and mortality of Atlantic croaker in the northwestern Gulf of Mexico lead to population-level responses. The model follows the hourly growth, mortality, reproduction, and movement of individuals on a 300 × 800 spatial grid of 1-km² cells for 140 years. Chlorophyll-a concentration, water temperature, and dissolved oxygen (DO) were specified daily for each grid cell and repeated for each year of the simulation. A bioenergetics model was used to represent growth, mortality was assumed stage- and age-dependent, and the movement behavior of juveniles and adults was modeled based on temperature and avoidance of low DO. Hypoxia effects were imposed using exposure effect submodels that converted time-varying exposures to low DO to reduced hourly growth, increased hourly mortality, and reduced annual fecundity. Results showed that 100 years of either mild or intermediate hypoxia produced small reductions in population abundance, while repeated severe hypoxia caused a 19% reduction in long-term population abundance. Relatively few individuals were exposed to low DO each hour, but many individuals experienced some exposure. The response was dominated by a 5% average reduction in annual fecundity of individuals. Under conditions of random sequences of mild, intermediate, and severe hypoxia years occurring in proportion to their historical frequency, the model predicted a 10% decrease in the long-term population abundance of croaker. A companion paper substitutes hourly DO values from a three-dimensional water quality model for the idealized hypoxia and results in a more realistic population reduction of about 25%.