自组织竞争神经网络在江苏油田有杆抽油系统故障诊断中的应用

有杆抽油系统的故障诊断技术是国内外采油工程技术中的一个重要研究课题。通过示功图的不同形状特征可以反映抽油机的不同工作状态。将自组织竞争神经网络应用于示功图的识别与分类,建立了一个自组织竞争神经网络模型对示功图进行自动聚类,从而实现故障诊断的自动化。应用江苏油田的实测示功图数据进行实验,可以看出自组织竞争神经网络具有良好的分类能力和泛化性能,是实现油田抽油系统故障诊断的有效技术,具有很强的实用价值和广泛的应用前景。     

浅谈提高空中三角测量数据成果质量的3个关键环节

空中三角测量数据主要是为摄影测量提供定向框架数据,其质量决定最终产品的质量。笔者结合生产实践中遇到的实际问题,对提高空中三角测量数据成果质量的3个关键环节——航外控制测量,航内控制点量测,外业控制点的残差分析、空中三角测量成果检查及评定的重要性进行了阐述。     

Network analysis to inform invasive species spread among lakes

Often facilitated by human-mediated pathways,aquatic invasive species are a threat to the health and biodiversity of global ecosystems.We present a novel approach incorporating survey data of watercraft movement in a social network analysis to reconstruct potential pathways of aquatic invasive species spread between lakes.As an example,we use the green alga Nitellopsis obtusa,also known as starry stonewort,an aquatic invasive species affecting the Great Lakes region in the United States and Canada.The movement of algal fragments via human-mediated pathways(i.e.,watercraft)has been hypothesized as the primary driver of starry stonewort invasion.We used survey data collected at boat ramps during the 2013 and 2014 openwater seasons to describe the flow of watercraft from Lake Koronis,where N.obtusa was first detected in Minnesota,to other lakes in the state.Our results suggest that the risk of N.obtusa expansion is not highly constrained by geographic proximity and management efforts should consider highly connected lakes.Estimating human movement via network analysis may help to explain past and future routes of aquatic invasive species infestation between lakes and can improve evidence-based prevention and control efforts.     

Network analysis to inform invasive species spread among lakes

Often facilitated by human-mediated pathways,aquatic invasive species are a threat to the health and biodiversity of global ecosystems.We present a novel approach incorporating survey data of watercraft movement in a social network analysis to reconstruct potential pathways of aquatic invasive species spread between lakes.As an example,we use the green alga Nitellopsis obtusa,also known as starry stonewort,an aquatic invasive species affecting the Great Lakes region in the United States and Canada.The movement of algal fragments via human-mediated pathways(i.e.,watercraft)has been hypothesized as the primary driver of starry stonewort invasion.We used survey data collected at boat ramps during the 2013 and 2014 openwater seasons to describe the flow of watercraft from Lake Koronis,where N.obtusa was first detected in Minnesota,to other lakes in the state.Our results suggest that the risk of N.obtusa expansion is not highly constrained by geographic proximity and management efforts should consider highly connected lakes.Estimating human movement via network analysis may help to explain past and future routes of aquatic invasive species infestation between lakes and can improve evidence-based prevention and control efforts.     

Location choices of Chinese enterprises in Southeast Asia: The role of overseas Chinese networks

Journal of Geographical Sciences - With the implementation of the “Going out” strategy and the Belt and Road Initiative, China’s investments have become increasingly influential...     

中国旅游城市生态安全系统评估与类型划分研究 ——基于节点权重约束网络DEA模型

采用 DPSIR概念模型作为旅游生态安全系统的基本形态,引入网络 DEA模型测度系统运行效率,从“投入-产出”的研究视角探究旅游生态安全系统内部作用机理。采用结构方程模型对网络 DEA模型的结构及权重加以限制,构建具有节点权重约束的网络 DEA评价模型,对主要旅游城市生态安全系统实例进行实证研究。研究结果认为：将生态安全系统分为自然运行阶段、管理反馈阶段2阶段进行效率评价,加权综合得出各旅游城市生态安全系统的综合评价值,以此为依据将样本城市划分为按系统效率的评价结果将其分划为绿色发展型、稳步发展型、高效发展型、双向提升型和管理能力缺失型,针对各类型旅游城生态安全系统特点提出改善其生态安全系统状况的建议。     

基于贝叶斯正则化BP神经网络的DEM趋势面逼近

趋势面从宏观上揭示了研究对象的特性,在各领域发挥着重要作用。BP神经网络可以对复杂系统进行无限逼近,进而进行预测。建立了基于贝叶斯正则化BP神经网络的数字高程模型趋势面,与二次多项式建立的数字高程模型趋势面进行比较分析,证明了该方法的可行性和有效性。     

GA-BP网络模型在GPS似大地水准面精化中的应用

针对误差反向传播（BP）算法训练速度慢和易于陷入局部最小值的缺点,提出了利用遗传算法（GA）的全局寻优性,结合GA和BP的各自优点,分析和建立了进化神经网络（GA-BP）模型,并将该模型应用于似大地水准面模型精化。最后以南方某市E级GPS控制网高程数据为例,进行BP和GA-BP模型的对比实验,通过对内、外符合精度及MAPE（平均绝对误差百分比）指标分析,验证了该方法的可行性。     

基于二分图的温度网络拓扑性质研究

利用中国地区435个台站1961—2002年逐日平均温度序列,将温度变化发生在9天时间尺度上的特征编码在网络中,通过研究二分图温度网络(BGT网络)中节点与项目的关系,揭示出9天时间尺度上温度变化的特征及其在空间上的拓扑统计性质.网络中各节点RRRD, RrDD, eeed, DRRD, DDRR等所代表的温度波动模态在网络中异常频发,对9天尺度温度变化的预报有一定的指导意义.统计网络的节点度分布,集群系数等拓扑结构特征量,发现BGT网络服从正态分布特征.BGT网络项目内节点度的多样性大体上     

无标度网络上队列资源分配研究

研究了节点队列资源有限的条件下,无标度网络上的信息流动力学过程,发现了网络由自由流通到拥塞的相变现象,提出了一种基于节点度的队列资源分配模型.模型的核心是使节点i的队列长度与k^β_i成正比(k_i为节点i的度,β为分配参数).仿真结果表明,在网络使用最短路径算法进行信息包传送的条件下,β近似等于1.25时队列资源分配最合理,网络容量最大,且该最佳值与队列总资源多少以及网络的规模无关.