不确定条件下基于三元区间规划的生产计划模型

研究了不确定条件下的生产计划问题,当企业面临多种产品消耗的单位产能、单位生产利润以及企业的总产能和交货期等均为不确定变量时,如何确定出的各种产品的合理的投产量,以使得企业在规定的交货期内获得最大的收益.引入了三元区间数来描述生产计划模型中的各种变量,建立了相应的三元区间线性规划模型;然后利用三元区间线性规划的有关知识简化了该模型,求出了其弱最优解.模型与传统的生产计划模型相比更贴近生产的实际过程,能为生产决策者提供更好的技术支持.     

用于真空弧等离子体放电阴极温度场测量的多光谱高温计

阴极表面温度是真空弧等离子体放电过程中一个重要参数,对真空弧等离子体的形成、电极腐蚀预测、热传导以及离子源的寿命都有重要影响。真空弧离子源的阴极具有目标小,放电过程快等特点,其温度的测量,对于时间分辨率和空间分辨率要求都很高,阴极表面温度的测量技术的欠缺,使得仅靠理论解析获得的结果难以得到验证。并且等离子体放电过程中测量仪器极易受到弧光的影响,如何避免放电过程中等离子体的辐射也是采用辐射法测量阴极表面温度要考虑的问题。这无疑给其温度场的测试研究带来困难。针对脉冲真空弧等离子体开展阴极表面温度测试实验有着重要意义,在分析了真空弧等离子体放电特性以及背景辐射特性和等离子体放电阴极测温的实际需求,本文基于高速CCD相机研制了一种新型的多光谱高温计。该高温计采用单色高速CCD相机,主要避免RGB彩色相机不能完全滤除背景辐射的弧光。为使用单色CCD相机实现多光谱辐射测温,设计了高温计的光学系统,该系统采用4孔径分光系统。将4种不同波长的滤光片嵌入到1个滤光片中。该研究设计的高温计可用于2 000~6 000 K的等离子体温度测量。并在中国工程物理研究院电子工程研究所进行现场测试,测试过程中将研制的高温计,通过外部触发形式对等离子体放电过程进行跟踪拍摄,高温计完全拍摄到等离子体放电过程。利用真空弧等离子体金属电极阴极放电的实测数据对高温计进行了验证。实验结果表明,设计的新型多光谱高温计能够用于测量真空弧等离子体放电时阴极温度场信息,测量的温度值低于放电电极的沸点温度,与等离子体放电过程中出现气化现象相符,说明高温计测的是等离子体放电阴极的温度。     

基于双向长短期记忆网络的太赫兹光谱识别

特征提取是太赫兹光谱识别的关键处理步骤,通常利用降维方法作为特征提取手段。然而,当一些化合物的太赫兹光谱曲线整体差异度较小时,降维方法往往会缺失样本差异的重要特征信息,从而导致分类错误。如果不采用降维方法提取特征,传统机器学习分类算法对维数较高的原始太赫兹光谱数据又不能很好的分类。针对此问题,提出了一种基于双向长短期记忆网络（BLSTM-RNN）自动提取太赫兹光谱特征的识别方法。BLSTM-RNN作为一种特殊的循环神经网络,利用其LSTM单元可以有效解决原始太赫兹光谱数据维数较高使得模型难以训练问题。再结合模型的双向频谱信息利用架构模式,可以增强模型对复杂光谱数据自动提取有效特征信息的能力。采用三类、15种化合物太赫兹透射光谱作为测试对象,首先利用S-G滤波和三次样条插值对Anthraquinone,Benomyl和Carbazole等十五种化合物在0.9～6 THz内的太赫兹透射光谱数据进行归一化处理,然后通过构建一个具有双向长短期记忆的循环神经网络对太赫兹光谱的全频谱信息进行自动特征提取并利用Softmax分类器进行分类。通过试验优化网络结构和各项参数,最终获得了针对复杂太赫兹透射光谱数据的预测模型,并与传统机器学习算法SVM,KNN及神经网络算法MLP,CNN进行对比实验。结果表明,dataset-1和dataset-2分别作为差异度较大和无明显峰值特征的五种化合物太赫兹透射光谱数据集,其平均识别率分别为100%和98.51%,与其他方法相比识别率有所提高;最重要的是,dataset-3作为5种化合物谱线极为相似的太赫兹透射光谱数据集,其平均识别率为96.56%,与其他方法相比识别率提高显著;dataset-4作为dataset-1,dataset-2和dataset-3的透射光谱数据集集合,其平均识别率为98.87%。从而验证了BLSTM-RNN模型能自动提取有效的太赫兹光谱特征,同时又能保证复杂太赫兹光谱的预测精度。在选择模型训练优化算法方面,使用Adam优化算法要好于RMSProp,SGD和AdaGrad,其模型的目标函数损失值收敛速度最快。同时随着模型训练迭代次数增加,相似太赫兹透射光谱数据集的预测准确率也不断提升。可为复杂太赫兹光谱数据库的光谱识别检索提供一种新的识别方法。     

Characterization of compounds in the Chinese herbal drug Mu-Dan-Pi by liquid chromatography coupled to electrospray ionization mass spectrometry

Cortex Moutan is a well-known traditional Chinese medicine derived from Paeonia suffruticosa ANDREWS. However, root cortices of P. delavayi and P. decomposita also are used under the name of this drug in some regions such as Yunnan and Sichuan Provinces, respectively. In order to make a comparison of their chemical constituents, the compounds of the three Paeonia species were analyzed by high-performance liquid chromatography-diode array detection/electrospray ionization and quadrupole-time-of-flight tandem mass spectrometry (HPLC-DAD/ESI-MS2). A total of 50 compounds were observed in the 50% (v/v) methanolic extracts, including 17 monoterpenes, 14 galloyl glucoses, 10 acetophenones, 5 phenolic acids, 3 flavonoids and 1 triterpene. These chemical constituents were separated on a C18 column and identified or tentatively characterized based on UV spectra and MS fragmentation behavior. The chemical compositions of the three Paeonia species were found to have many differences. Paeonol was the predominant constituent of P. suffruticosa and P. decomposita, while P. delavayi contained albiflorin and more galloyl glucoses than the other two Paeonia species. Most of these identified compounds have been reported from P. delavayi and P. decomposita for the first time. The ESI-MS fragmentation behavior of monoterpene glycosides, acetophenones and galloyl glucoses was also investigated successively, and appropriate characteristic pathways were proposed. The large differences in chemical compounds among the three Paeonia species strongly encouraged further comparison of the bioactivities of these three species.     

Identification of Acetylcholinesterase Inhibitors from Seeds of Plants of Genus Peganum by Thin-Layer Chromatography-Bioautography

JPC – Journal of Planar Chromatography – Modern TLC - In traditional Chinese medicine, plants of the genus Peganum have been used to treat cough, hypertension, diabetes, asthma,...     

Investigation of the volatile constituents of different Gynura species from two Chinese origins by SPME/GC-MS

GC-MS analyses of the volatile constituents obtained by solid phase microextraction (SPME) of two Gynura species, i.e., G. divaricata and G. bicolor, collected from Nanjing and Nanping areas in the east of China, enabled the identification of more than 50 different components. Generally, a higher contents of sesquiterpene compounds such as beta-caryophyllene, alpha-caryophyllene and alpha-copaene were found in G. bicolor than in G. divaricata, regardless of their origin. Qualitatively, the detected major volatiles of G. bicolor and G. divaricata originating from Nanjing were the same, i.e., beta-caryophyllene, alpha-caryophyllene, alpha-copaene, alpha-pinene and beta-pinene. The volatile profile of a third species, G. medica, also originating from Nanjing was different with sesquiterpenes gamma-cadinene, beta-caryophyllene, elixene and monoterpene limonene as the major components. The volatile profiles of G. bicolor and G. divaricata originating from Nanping were more diverse. For both species beta-caryophyllene, alpha-copaene and limonene were the major volatiles, but G. bicolor also produced gamma-caryophyllene and carvone as major constituents, whereas for G. divariata o-cymene was detected. Overall, these results indicate that the biosynthesis of volatiles by Gynura species is not only species related, but is also possibly influenced by the local environmental conditions of growth.     

基于凸组合核函数的化合物太赫兹透射光谱分类

物质的太赫兹光谱包含着非常丰富的物理和化学信息。它对化合物晶体具有高的灵敏度、 单光子能量低等特点。但受到检测人员知识背景、 背景噪声、 识别算法精度等因素的影响,光谱样本识别准确率和效率较低。为了提高对太赫兹光谱的检测能力,提出应用基于凸组合核函数的support vector machines(SVM)对化合物的THz脉冲透射谱进行分类。在使用小波变换对数据进行滤波预处理之后,提取了传统波峰、 波谷位置特征和term frequency-inverse document frequency (TF-IDF) 最大间隔特征。TF-IDF方法使用信息论的原理确定每个采样点的权重,选择权重较大的点作为特征。针对太赫兹透射谱特征相似、 维数较低带来的分类困难问题,构建基于凸组合核函数的SVM分类模型。并利用核评价的方法,通过高维非线性规划方程求解最优凸组合参数。当最优凸组合参数被确定时,构建分类模型进行分类和预测。相比较于单一核函数,凸组合核函数将透射谱特征与分类模型融合起来。对于不同的检测样本,数据经过凸组合核函数映射到高维空间后,特征具有更显著的区分度。使用不同的太赫兹透射谱样本进行分类实验,结果表明,分类准确率得到极大提高。     

Charm physics — A field full of challenges and opportunities

In this review, we discuss some interesting issues in charm physics, which is full of puzzles and challenges. So far in this field, there exist many problems which have not obtained satisfactory answers yet as more unexpected phenomena Continue to be observed at the current facilities of high energy physics. Charm physics may become an ideal place for searching new resonances and studying non-perturbative QCD effects, moreover it is probably an area for exploring new physics beyond the Standard Model. More data will be available at the BE-SIII, B-factories, LHC and even the future ILC, which may open a wide window to a better understanding of its nature.      

Identification and Kinetics Study on the Transformation of Mevinolin in Acidic Alcohol Solution by High-performance Liquid Chromatography with Photodiode Array Detector or Mass Spectrometry

IntroductionIncreased levels of cholesterol and triglycerides areknown to be the risk factors for developing coronary ar-tery disease. Lipid lowering agents that inhibit theHMG coenzyme A( HMG-CoA) reductase are nowprominent among the drugs for treating h…     

多层陶瓷电容器在脉冲条件下寿命特性的研究

 主要研究了高储能密度多层陶瓷电容器（MLC）在不同充放电频率、充电电压、充放电占空比和放电反峰系数等条件下的寿命特性，并对其失效机理进行分析。实验发现，正常情况下MLC试品的击穿需要经过局部放电通道不断发展，绝缘逐渐被破坏的过程，寿命较长，发热和试品内部的交变应力是推动局部放电通道不断发展的根本因素。充放电频率的升高使试品发热加剧，寿命缩短；充电电压和放电反峰系数的增大使发热和介质内应力加剧，可以使寿命缩短90％以上。改善散热可以有效提高试品寿命，如当充放电频率为136 Hz时，采用油浸散热比空气自然换流散热寿命延长300％。