The Newton filtration and d-determination of bifurcation problems related to C~0 contact equivalence

In this paper, from the Newton filtration's point of view, we construct the singular Riemannian metric and use the method in singular theory to study the bifurcation problems, and give the sufficient condition of d-determination of bifurcation problems with respect to C0 contact equivalence. The special cases of the main result in this paper are the results of Sun Weizhi and Zou Jiancheng.     

基于BP网络分类器的交通标志识别

针对中国全部3大类116个交通标志,即禁令标志、指示标志、警告标志,用BP网络实现分类功能.实验中使用了3种测试集,即加高斯噪声、水平扭曲和日本交通标志实景图,对BP网络的分类性能进行了测试.实验结果表明:用BP网络实现交通标志粗分类功能的效果是比较理想的,对交通标志的颜色失真和形状失真均具有较好的容错性和鲁棒性.     

High spin chemistry underlying organic molecular magnetism: Topological symmetry rule as the first principle of spin alignment in organic open-shell systems of π-conjugation and their ions

This review paper deals with an overview of molecule-based magnetism as a rapidly developing interdisciplinary field, topological symmetry rule as the first principle of spin alignment in organic open-shell systems in the ground state, the proposal of organic through-bond 1D and 2D ferro- and superparamagnets and the detection of the first organic high-spin molecule, m-phenylenebis(phenylmethylene) in the quintet ground state (S = 2), followed by extended organic high-spin systems with π-conjugation such as aromatic hydrocarbons having S = 3, 4, 5. The paper also describes a theoretical approach to the understanding of electronic spin structures of organic high-spin molecules by invoking both Heisenberg and Hubbard model Hamiltonians, weakly interacting intramolecular high-spin systems from both experimental and theoretical sides, the spin density distribution of the first organic high-spin molecule in terms of electron- nuclear multiple resonance spectroscopy and the detection and characterization of ionic high-spin hydrocarbons, emphasizing the establishment of high spin chemistry underlying organic molecular magnetism.     

A multi-target approach for isolated word recognition using artificial neural network

I.TntroductionStatisticalandneuralnetworkmcthodsforpatternclassiflcationusesignificantlydifTerentapproachesintrainingaclassificr.Inthestatistica1approach,thcformationofaclassifier1arge1ydependsonthestatisticsofthetrainingpatternsand,insomecases,theassumptionsaboutthedistributionofthepopulation.Theneuralnetworkmethodisnon-parametricandcanbeadaptivcinthetrainingprocessl'l.Becauseofitssimplicityandflexibility,theneuralnetworkhasbecnincreasing1yusedforpatternclassiflcation.Itisnowwe11knownthatan…     

High sensitivity CW-cavity ring down spectroscopy of water in the region of the 1.5 μm atmospheric window

The absorption spectrum of natural water vapour around 1.5 μm has been recorded with a typical sensitivity of 5 × 10⁻¹⁰ cm⁻¹ by using a CW-cavity ring down spectroscopy set up based on fibred DFB lasers. A series of 31 DFB lasers has allowed a full coverage of the 6130.8-6748.5 cm⁻¹ (1.63-1.48 μm) region corresponding to the H transparency band of the atmosphere. The line parameters (wavenumber and intensity) of a total of 5190 lines, including 4247 lines of water vapor, were derived by a one by one fit of the lines to a Voigt profile. Different isotopologues of water (H₂¹⁶O, H₂¹⁸O, H₂¹⁷O, and HD¹⁶O) present in natural abundance in the sample contribute to the spectrum. For the main isotopologue, H₂¹⁶O, 2130 lines were measured with line intensities as weak as 10⁻²⁹ cm/molecule while only 926 lines (including a proportion of 30% inaccurate calculated lines) with a minimum intensity of 3 × 10⁻²⁷ cm/molecule are provided by the HITRAN and GEISA databases. Our comparison in the whole 5750-7965 cm⁻¹ region, has also evidenced that an error in the process of conversion of the intensity units from cm⁻²/atm to cm⁻¹/(molecule × cm⁻²) at 296 K, has led to H₂¹⁶O line intensities values listed in the HITRAN-2000 database, systematically 8 % below the original FTS values. The rovibrational assignment was performed on the basis of the ab initio calculations by Schwenke and Partridge with a subsequent refinement and validation using the Ritz combination principle together with all previously measured water transitions relevant to this study. This procedure allowed determining 172, 139, 71, and 115 new energy levels for the H₂¹⁶O, H₂¹⁸O, H₂¹⁷O, and HD¹⁶O isotopologues, respectively. The results are compared with the available databases and discussed in regard of previous investigations by Fourier transform spectroscopy. The spectrum analysis has showed that most of the transitions which cannot be assigned to water are very weak and are due to impurities such as carbon dioxide and ammonia, leaving only about 3% of the observed transitions unassigned. The interest of a detailed knowledge of water absorption for trace detectors developed in the 1.5 μm range is underlined: for instance HDO contributes significantly to the considered spectrum while no HDO line parameters are provided by the HITRAN database.     

带有边集限制的最均匀支撑树问题

本在无向网络中，建立了带有边集限制的最均匀支撑树问题的网络模型．中首先解决最均匀支撑树问题，并给出求无向网络中最均匀支撑树的多项式时间算法；然后，给出了求无向网络中带有边集限制的最小树多项式时间算法；最后，在已解决的两个问题的基础上解决了带有边集限制的最均匀支撑树问题．     

On tolerance lattices of algebras in congruence modular varieties

We prove that the tolerance lattice TolA of an algebra A from a congruence modular variety V is 0-1 modular and satisfies the general disjointness property. If V is congruence distributive, then the lattice Tol A is pseudocomplemented. If V admits a majority term, then Tol A is 0-modular. This revised version was published online in June 2006 with corrections to the Cover Date.     

BP人工神经网络用于除草剂活性预测

采用AMI方法研究了10种新型的磺酰脲类除草剂的电子结构，并以原子的Mulliken净电荷和除草剂在不同浓度(100，10mg／L)下对油菜、稗草两种作物的根、茎部位的抑制率为训练样本集。构造并训练得到具有活性预测能力的BP神经网络．结果表明，该BP网络不仅能对训练样本很好拟合。亦能对未知化合物的活性作出很好的预测．     

全光网中基于FBG的OADM设计研究

提出一种全新的OADM(Optical Add—Drop Multiplexers)结构设计，此结构仅采用一个多端口环行器，大大减少了环行器的数量，节约了成本。若要上／下路一个波长．则只需要一个FBG(Fiber Bragg Grating)；若要同时上／下路多个波长，则需要2m个FBG。计算机仿真实验证明此结构比H．Yuan提出的结构具有更低的损耗．对全光网络的实现具有重大的意义。     

网络图在产品配送中的应用

面对多样化的市场需求，企业的销售渠道也变得异常复杂，利用网络图可以简化实际问题，并能形象地描述它。本根据网络知识建立了产品分销模型，探求企业在市场需求固定和随机时的最优收益。最后，章还进行实例验算，并做出相应的结果分析。