自相关过程协方差阵的残差MEWMA控制图

本文研究多变量自相关过程协方差矩阵的质量监控问题.针对服从VAR(1)模型的多变量自相关过程,基于残差分析思想,将多变量指数加权移动平均(MEWMA)控制图用于对过程的监控.首先对残差MEWMA控制图统计量进行了推导,其次采用MonteCarlo仿真的方法,以平均运行链长为评价准则对残差MEWMA控制图的监控效果进行了分析.仿真结果表明,残差MEWMA控制图具有较好的监控效果.     

Fenton试剂处理水中有机物的特性及其应用

阐述了Fenton试剂在水处理应用中的作用、机理和特点,分析了过氧化氢投加量、亚铁离子投加量、水的pH值、水中共存离子等影响因素及Fenton试剂在水处理应用中的某些缺陷。另外简要地说明了Fenton试剂在废水和饮用水中的应用概况,最后对Fenton试剂的发展方向进行了简单预测。     

基于核函数PCA的非线性过程实时监控方法

为了克服基于主元分析的过程监控方法非线性处理能力弱的缺点和降低基于非线性主元分析的过程监控方法的计算复杂度,提出了将核函数PCA监控方法用于复杂工业过程实时监控系统的开发研究,并讨论了核函数参数选择对系统性能的影响。核函数PCA能有效地提取过程变量的非线性关系,而且计算复杂度低,便于在线实施。仿真结果表明该方法是一种有前途的复杂过程非线性实时监控技术。     

实值可料过程关于集值Wiener过程的伊藤积分

为研究实值可料过程关于可积有界紧凸集值Wiener过程的伊藤积分,先从简单实值可料过程入手,以支撑函数为工具,给出相关定义及性质。然后借助于简单过程构造出一般过程的轨道,把结论推广到一般实值可料过程,给出其关于可积有界紧凸集值Wiener过程的伊藤积分的定义及性质。     

循环平稳过程与其离散化后过程统计量的关系

探讨了连续循环平稳过程的离散化问题,详细推导了其离散化后过程的时变统计量及循环统计量的表达式,比原有表达式更为简单和明确。根据这些表达式所得到的样值序列统计量和原连续过程统计量的定量关系,在满足不发生混迭的条件下,可由离散过程的统计量得到连续过程的统计量,在实际中有着较好的应用价值。     

Response of soil heat-water processes to vegetation cover on the typical permafrost and seasonally frozen soil in the headwaters of the Yangtze and Yellow Rivers

The response of soil temperature and moisture to vegetative cover in the active layer of permafrost and seasonally frozen soil were assessed and compared. Soil temperature and moisture, under a range of vegetation covers (92%, 65% and 30%) in the permafrost and vegetation covers (95%, 70%―80%, 40%―50% and 10%) in the seasonally frozen soil, were measured on a daily basis. A decline in vege-tation cover led to a decrease in the integral of freezing depth of active permafrost layer, but an in-crease in seasonally frozen soil. The maximum invasion depth and duration of the negative isotherm during the frozen period and of the positive isotherm during the non-frozen period clearly increased when vegetation cover declined. With a reduction of vegetation cover, the soil moisture in the active layer of the permafrost decreased for depths of 0.20―0.60 m, but increased for depths of 0.60―0.80 m, while for seasonally frozen soil, soil moisture of the entire profile (0.10―1.20 m) increased. Variation in vegetation cover alters soil heat-water processes, but the response to it is different between permafrost and seasonally frozen soil.     

Global monsoon in a geological perspective

Monsoon is now considered as a global system rather than regional phenomena only. For over 300 years, monsoon has been viewed as a gigantic land-sea breeze, but now satellite and conventional observations support an alternative hypothesis which considers monsoon as a manifestation of sea-sonal migration of the intertropical convergence zone (ITCZ) and, hence, a climate system of the global scale. As a low-latitude climate system, monsoon exists over all continents but Antarctica, and through all the geological history at least since the Phenorozoic. The time is ripe for systematical studies of monsoon variations in space and time. As evidenced by the geological records, the global monsoon is controlled by the Wilson cycle on the tectonic time scale (10⁶―10⁸ a). A “Mega-continent” produces “Mega-monsoon”, and its breakdown leads to weakening of the monsoon intensity. On the time scales of 10⁴―10⁵ a, the global monsoon displays the precessional cycles of ~20 ka and eccentricity cycles of 100- and 400-ka, i.e. the orbital cycles. On the time scales of 10³ a and below, the global monsoon intensity is modulated by solar cy-cles and other factors. The cyclicity of global monsoon represents one of the fundamental factors re-sponsible for variations in the Earth surface system as well as for the environmental changes of the human society. The 400-ka long eccentricity cycles of the global monsoon is likened to “heartbeat” of the Earth system, and the precession cycle of the global monsoon was responsible for the collapse of several Asian and African ancient cultures at ~4000 years ago, whereas the Solar cycles led to the de-mise of the Maya civilization about a thousand years ago. Therefore, paleoclimatology should be fo-cused not only on the high-latitude processes centered at ice cap variations, but also on the low-latitude processes such as monsoons, as the latter are much more common in the geological his-tory compared to the glaciations.     

一种在线自适应控制马氏链的强化学习算法

讨论平均准则控制马氏链的强化学习算法。目的是寻找使得长期每阶段期望平均报酬最大的最优控制策略,由于事先未知状态转移矩阵及报酬向量,故必需使用自适应控制方法,通过引入称之为行动器和评判器的神经网络构造,使得学习单元在不断学习中,最终能发现最优策略。行动器的参数在学习中不断被修正,每一时刻的参数的值均对应着一个随机控制策略。评判器用来估计这些参数以找出最优控制策略。     

轧制过程自动化工程支撑体系——工程科学的前沿技术

本文论述现代工程必须以工程科学的方法来设计和实施,实现工程科学方法的基础和前沿技术是一类工程的支撑体系.以轧制过程自动化为例,概述了工程支撑体系的组成和各部分内容.采用这套工程支撑体系,以工程科学方法承担大量的轧制过程自动化工程项目,取得了重大成功和显著的经济效益.     

基于时间反演的多臂赌博机跳空图案生成算法

在多输入多输出(multiple input multiple output, MIMO)系统中,为了实现系统的防窃听能力,并提高系统容量,提出了一种基于时间反演的多臂赌博机跳空图案生成算法。该算法考虑信道状态信息的误差,以时间反演和跳空通信为基础,结合使用多臂赌博机算法选取最优跳空图案以提高系统容量的同时降低因防窃听对合法接收端性能的影响。理论分析和仿真结果表明,该算法在防窃听性能、系统容量方面具有优势。