An ANN‐based model for spatiotemporal groundwater level forecasting

This paper evaluates the feasibility of using an artificial neural network (ANN) methodology for estimating the groundwater levels in some piezometers placed in an aquifer in north‐western Iran. This aquifer is multilayer and has a high groundwater level in urban areas. Spatiotemporal groundwater level simulation in a multilayer aquifer is regarded as difficult in hydrogeology due to the complexity of the different aquifer materials. In the present research the performance of different neural networks for groundwater level forecasting is examined in order to identify an optimal ANN architecture that can simulate the piezometers water levels. Six different types of network architectures and training algorithms are investigated and compared in terms of model prediction efficiency and accuracy. The results of different experiments show that accurate predictions can be achieved with a standard feedforward neural network trained usung the Levenberg–Marquardt algorithm. The structure and spatial regressions of the ANN parameters (weights and biases) are then used for spatiotemporal model presentation. The efficiency of the spatio‐temporal ANN (STANN) model is compared with two hybrid neural‐geostatistics (NG) and multivariate time series‐geostatistics (TSG) models. It is found in this study that the ANNs provide the most accurate predictions in comparison with the other models. Based on the nonlinear intrinsic ANN approach, the developed STANN model gives acceptable results for the Tabriz multilayer aquifer. Copyright © 2008 John Wiley & Sons, Ltd.     

现代综合业务数字网的现状与展望

现代通信网正在从公共电话交换网向综合数字网、综合业务数字网、宽带综合业务数字网发展，以适应电信现代化建设的需要。     

Comparison of radial basis function networks and empirical equations for converting from pan evaporation to reference evapotranspiration

Evapotranspiration (ET) is one of the major processes in the hydrological cycle, and its reliable estimation is essential to water resources management. Numerous equations have been developed for estimating ET, most of which are complex and require numerous items of weather data. In many areas, the necessary data are lacking, and simpler techniques are required. Evaporation pans are used throughout the world because of the simplicity of technique, low cost, and ease of application. In this study, the radial basis function (RBF) network is applied for pan evaporation to evapotranspiration conversions. The adaptive pan‐based RBF network was trained using daily Policoro data from 15 May 1981 to 23 December 1983. The RBF network obtained, Christiansen, FAO‐24 pan, and FAO‐56 Penman–Monteith equations were verified in comparison with lysimeter measurements of grass evapotranspiration using daily Policoro data from 25 February to 18 December 1984. Based on summary statistics, the RBF network ranked first with the lowest RMSE value (0·433 mm day^?1). The RBF network obtained on the basis of the daily data from Policoro, Italy and pan‐based equations were further tested using mean monthly data collected in Novi Sad, Serbia, and Kimberly, Idaho, USA. The overall results favoured use of the RBF network for pan evaporation to evapotranspiration conversions. The use of the RBF network is very simple and does not require any knowledge of ANNs. Users require only code (RBF network), E_pan data and corresponding R_a data. Copyright © 2009 John Wiley & Sons, Ltd.     

基于GIS与SDM技术的可视化空间数据分类研究

提出将GIS与可视化空间数据挖掘技术之集成的基本框架。在此基础上,基于VisualC++6.0和Ma-pObject2.0组件技术设计和开发了一个可视化交互空间数据挖掘分类系统,系统采用决策树方法和贝叶斯网络作为数据挖掘方法的基本算法,采用训练与学习相结合实现空间数据的分类。文中用实例数据对系统性能、算法和规则有效性进行了验证。结果表明,该系统是一个适用的、可扩展的可视化交互空间数据挖掘工具,系统能够实现数据挖掘实时动态的交互控制,实现了数据挖掘过程的可视化、挖掘模型的可视化和结果的可视化显示、可视化思考、可视化分析与评价。     

基于A- CGAN的深反射地震数据随机噪声压制方法研究

基于深度学习的地震数据噪声压制方法是当前地震数据去噪处理的重要方向。深度学习方法突破了传统滤波处理的局限,在对常规地震数据的噪声压制中表现出效率高、信噪分离效果好的特点。但针对深部弱有效反射数据,当前的深度学习方法特征提取能力有限,难以取得较好的去噪效果。笔者等结合深反射地震数据特点,针对当前深度学习噪声压制方法在特征提取及对数据集依赖上的局限,提出了基于注意力循环生成对抗网络(Attention Cycle- Consistent Generative Adversarial Networks,A- CGAN)的深反射地震数据随机噪声压制方法。借助循环一致生成对抗网络(Cycle- Consistent Generative Adversarial Networks,Cycle- GAN)的域映射思想,降低对数据集的要求。为了构建适用于深反射地震数据的去噪网络,从3个方面对Cycle- GAN进行改进：在Cycle- GAN的生成器（去噪器）中加入残差结构和注意力机制,用于加深网络深度和提高其特征提取能力;在Cycle- GAN的鉴别器中使用块判决,提高鉴别精度和准确度;在损失函数部分加入感知一致性损失函数,提升网络模型恢复纹理细节信息的能力。通过合成地震数据和实际深反射地震数据测试,验证了优化算法的有效性,体现了良好的应用价值。     

基于复杂网络的岩土颗粒材料分散性失稳先兆研究

岩土颗粒材料在达到Mohr-Coulomb塑性极限之前可能已经发生了没有明显局部损伤的全域准均匀破坏,即分散性失稳。近年来,网络科学工具在颗粒材料环境中的应用为研究其失稳破坏现象提供了有趣新颖的洞见。采用离散单元法对不同初始密实度的颗粒体系进行等体积应变加载路径的数值试验,发现初始状态越松散的颗粒体系越有可能发生分散性失稳;采用网络科学的理论和方法分析颗粒接触网络的拓扑结构特性和演化规律,发现颗粒材料分散性失稳的结构根源是颗粒接触网络完全崩塌。将颗粒体系划分为强、弱接触系统以构建强接触子网络、弱接触子网络和强−弱接触子网络,研究颗粒接触网络及其3个子网络的特征量演化,结果表明：当颗粒材料开始处于不稳定状态时,易受扰动的弱接触系统的部分接触先失效,导致强接触系统失去其稳定支撑作用而降低承载能力。随着加载的继续,系统通过非局部的自组织过程导致接触结构的全面崩塌,最终触发整体分散性失稳。因此,先于整体失稳的部分弱接触的失效,可视为颗粒材料分散性失稳的关键性临界征兆。     

面向室内WLAN定位的动态自适应模型

提出了一种面向无线局域网(wireless local area network,WLAN)位置指纹匹配定位的动态自适应模型,借助多个基站的实时数据为稳健室内定位提供更新的匹配数据库——radio map。基于基站与radio map采样点之间的空间关联性,将基站数据和radio map分别作为多层神经网络的输入和输出,动态更新radio map;利用多元数据异常探测技术检验基站数据捕获环境的时空变化;根据探测结果采用顾及室内布局的数值内插和再训练的方式更新模型,从而使其适应环境的时空变化。在室内动态环境中进行了多次验证试验。试验结果显示,在时变因素作用下,相比较传统方法,采用所提模型的定位方法的平均误差至少下降10%;在空间变化因素(以信标移动为例)作用下,其他方法平均定位误差大幅增加了至少165%,而采用所提模型方法的平均定位误差只增加了10%~20%,定位精度维持在3m左右(即原始精度)。结果证明采用了所提模型的定位系统能够自适应环境的时空变化而保持原有定位表现。不过,与传统位置指纹匹配定位方法相比,该模型带来了更多的计算负荷。     

基于对偶拓扑结构的路网路段重要性评估方法

现有的路网路段重要性评估方法考虑的是路网中的路段的统计特性或路网的局部结构对重要性的影响。在路段的重要性与路网的全体路段相关联的基础上,提出m阶邻居节点的复杂路网路段重要度评估方法。为验证算法的有效性,实验仿真采用成都市路网的对偶拓扑结构,在1 484个路段中提取10条关键路径对评估方法进行验证。评估结果显示:与度值法、介数法相比,该方法能显著地区分复杂路网中路段之间的重要性差异,准确地确定网络中的关键路径,具有更高的评估准确性。     

From Soda Bottles To Super Labs: An Analysis Of North America's Dual Methamphetamine Production Networks

This article uses the global production network perspective to analyze the various actors involved in the production and distribution of methamphetamine in North America. The paper analyzes the organization and operation of the two dominant network typologies at work in the United States: small toxic labs and Mexican drug‐trafficking organizations. Particular emphasis is placed on how the networks have adapted to changes in the regulatory context in which they operate. I conclude with potential policy recommendations for curtailing methamphetamine production within both networks.     

Grain‐Size Characteristics of Dune Networks in China's Tengger Desert

Grain‐size characteristics of dune networks in the Tengger Desert of northwestern China were investigated. Grain‐size parameters (mean, standard deviation, skew, and kurtosis) were determined on dune surfaces at windward toe, stoss, crest and leeward toe locations. Multiple discriminant analyses were applied to distinguish deposition environments. Results indicated that the aeolian sediment is mainly composed of very fine and fine sand in the dune networks of the Tengger Desert. Sorting improves as grain size becomes finer. However, the relationships between mean grain size, skew, and kurtosis vary in space. There is a negative relationship between skew and sorting; similarly, the relationships between sorting and skew, and skew and kurtosis change in space. The sediment deposition environment includes aeolian sediment and lacustrine sediment. The lacustrine sediment provides the source material for dune windward toe and stoss formation and development, but the aeolian sediment provides the source material for dune crest and leeward toe formation and development. According to the log‐probability of grain size distributions, aeolian sediments in dune networks are composed of two distinct saltation populations.