检索结果-南通市图书馆

Application of Diagonal Recurrent Neural Network toDC Motor Speed Control Systems

在线全文

学校读者我要写书评

暂无评论

International Journal of Minerals,Metallurgy and Materials 2000年第1期14卷 68-71页

作者： Jing Wang Hui Chen information Engineering school, {2. of Science and Technology beijing, beijing 100083, china information Engineering school University of Science and Technology Beijing Beijine 100083 China

A new kind of dynamic neural network--diagonal recurrent neural network (DRNN) and its learning method and architecture are presented. A direct adaptive control scheme is also developed that is applied to a DC (Direct Current) speed control system with the ability to auto-tune PI (Proportion Integral) parameters based on combining DRNN with PI controller. The simulation results of DRNN show better control performances and potential practical use in comparison with PI controller.

关键词： diagonal recurrent neural network PI controller DC Motor speed control system

Sediment transport capacity of hyperconcentrated flow

在线全文

学校读者我要写书评

暂无评论

Science china(Physics,Mechanics & Astronomy) 2008年第8期51卷 961-975页

作者： SHU AnPing1,2.& FEI XiangJun2.1 school of Environment & Key Laboratory of Water and Sediment sciences,Ministry of Education,{2.,beijing 100875,china 2.State Key Laboratory of Hydroscience and Engineering,Tsinghua university,beijing 100084,china school of Environment & Key Laboratory of Water and Sediment sciences Ministry of Education Beijing Normal University Beijing 100875 China State Key Laboratory of Hydroscience and Engineering Tsinghua University Beijing 100084 China

As one of the most important components of river mechanics,sediment transport capacity of sediment-laden flows has attracted much attention from many researchers working on river mechanics and hydraulic engineering. Based on the time-averaged equation for a turbulent energy equilibrium in solid and liquid two-phase flow,an expression for the efficiency coefficient of suspended load movement was derived for the two-dimensional,steady,uniform,fully-developed turbulent flow. A new structural expression of sediment transport capacity was achieved. Using 115 runs of flume experimental data,which were obtained through two kinds of sediment transport experiments in the state of equilibrium,in combination with the basic rheological and sediment transporting characteristics of hyperconcentrated flow,the main parameters in the structural expression of sediment transport capacity were calibrated,and a new formula of sediment transport capacity for hyperconcentrated flow was developed. A large amount of field data from the Yellow River,Wuding River,and Yangtze River,etc. were adopted to verify the new formula and good agreement was obtained. These results above contribute to an improved theoretical system of river mechanics and a reliable tool for management of rivers carrying high concentration of sediments.

关键词： hyperconcentrated flow turbulent energy equilibrium efficiency coefficient sediment transport capacity

Multimedia Data Coding and its Development

在线全文

学校读者我要写书评

暂无评论

International Journal of Minerals,Metallurgy and Materials 2000年第3期14卷 229-232页

作者： Xingye Yu, Yang Yang (information Engineering school, {2. of Science and Technology beijing, beijing 100083, china) information Engineering school University of Science and Technology Beijing Beijing 100083 China

The requirements of data coding in multimedia applications are presented, the current technique of coding and relative standards is introduced, then the work that have been doing is presented, i.e. the wavelet-based coding method and the VE (Visual Entropy)-based coding method. The experiment results prove that these methods have gained a better perceptual quality of a reconstructed image and a lower bit rate. Their performance evaluations are better than JPEG (Joint Photographic Experts Group) coding. Finally, the future topics of study are put forward.

关键词： multimedia data coding MPEG (Moving Picture Experts Group) VE (Visual Entropy)

Node ordinal encoded genetic algorithm for the optimal allocation of water resources

在线全文

学校读者我要写书评

暂无评论

Progress in Natural Science:Materials International 2005年第5期15卷 448-452页

作者： YANG Xiaohua1*, YANG Zhifeng 1,2. SHEN Zhenyao1 and LI Jianqiang3(1. State Key Laboratory of Water Environment Simulation, school of Environment, beijing normal university, {2. 100875, china 2. Key Laboratory for Water and Sediment sciences Ministry of Education, school of Environment, beijing normal university, {2. 100875, china 3. Water Resources and Hydropower Planning and Design General Institute, MWR, {2. 100011, china) State Key Laboratory of Water Environment Simulation School of Environment Beijing Normal University Beijing 100875 China Key Laboratory for Water and Sediment sciences Ministry of Education School of Environment Beijing Normal University Beijing 100875 China Water Resources and Hydropower Planning and Design General Institute MWR Beijing 100011 China

A new method, node ordinal encoded genetic algorithm (NOEGA), is proposed for solving water resources optimal allocation problems, in which the capacity of water resources is split into a number of smaller parts so that successive operations can be overlapped. Our objective is to maximize the whole benefit function. To overcome the “dimensionality and algorithm complexity curse” while searching for solutions and looking for an optimal solution, the operations of one-point crossover operator, gene exchange operator, gene random operator, gene shift operator and node ordinal strings are established. It is proved to be an effective optimal method in searching for global solutions. The NOEGA does not need a diversity of initial population, and it does not have the problem of immature convergence. The results of two cases show that using NOEGA to solve the optimal allocation model is very efficient and robust. In addition, the algorithm complexity of NOEGA is discussed.

关键词： water resources optimal allocation node ordinal code genetic algorithm.

Improvement of Boxcar Integrator and Its Application to Low Resistance Measurement of Inductive Load Coil

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

International Journal of Minerals,Metallurgy and Materials 1998年第3期12卷 180-184页

作者： Xisheng Li Shaochun Wang(information Engineering school, {2. of Science and Technology beijing, beijing 100083, china) information Engineering school University of Science and Technology Beijing Beijing 100083 China

A boxcar integrator is described which is suitable for the low-repetition-rate signal processing. This boxcar integrator, named fixed-interval mode boxcar integrator, is able to reject harmonics other than the first harmonic component. It can also decrease the effective time constant In many situations, the antialiasing filter with narrow bandwidth will cause distortion of the input signal. The fixed-interval mode boxcar integrator with suitable gate width can achieve relative high performance without signal distortion because the bandwidth of its antialiasing filter can be wider than that in the fixed-Point boxcar integrator. ms boxcar integrator is used as majn part of signalprocessing circult in the low resisance measurement of inductive load coil. The results of experiments show that the fixed-interval boxcar integrator is suitable for low-repetition-rate use.

关键词： signal processing resistance measurement boxcar integrator

Identification of Nonlinear Dynamic Systems Using Diagonal Recurrent Neural Networks

在线全文

学校读者我要写书评

暂无评论

International Journal of Minerals,Metallurgy and Materials 1999年第2期13卷 149-151页

作者： Jing Wang Hui Chen(information Engmeering school, {2. of Science and Techaology beijing, beijing 100083, china) information Engineering school University of Science and Technology Beijing Beijing 100083 China

In order to apply a new dynamic neural network- Diagonal Recurrent Neural NetWork (DRNN) to the system identificationof nonlinear dynamic Systems and construct more accurate system models, the structure and learning method (DBP algorithm) of theDRNN are Present6d. Nonlinear system characteriStics can be identified by presenting a set of input / output patterns tO the DRNN andadjusting its weights with the DBP algorithm. Experimental results show that the DRNN has good performances in the identification ofnonlinear dynamic systems in comparison with BP networks.

关键词： neural network system identification intelligent control control system models learning method

Optimization of Fractal Iterated Function System(IFS) with Probability and Fractal Image Generation

在线全文

学校读者我要写书评

暂无评论

Journal of university of Science and Technology beijing 2001年第2期8卷 152-156页

作者： Berthe Kya, Yang Yang (information Engineering school, {2. of Science and Technology beijing, beijing 100083, china) Univ Sci & Technol beijing Informat Engn Sch Beijing 100083 Peoples R China

There are several methods for rendering fractal images based on IFS (Iterated Function System) in computer graphic; but one concern of the computer graphic community has been the efficiency rendering algorithms. The invariant measures arising from IFSP (Iterate Function System with Probability) using the probabilistic algorithm are also known as chaos games. The role of these probabilities to generate the image of the attractor has been investigated using a multi-fractal analysis. The conventional choice of probability associated with each set /S/ to generate the attractor in least time possible is introduced, then a new method based on the self similarity and multi-fractal analysis is presented. The efficiency of the new method has been proved over the conventional method.

关键词： optimization of fractal IFS compression IFSP multi-fractal analysis

Geographical Agglomeration of Chinese Manufacturing Industries

在线全文

学校读者我要写书评

暂无评论

中国人口·资源与环境：英文版 2007年第2期5卷 3-11页

作者： Zhang Hua 1, 2. Zheng Wei3 1. Institute of Geographical sciences and Natural Resources Research, Chinese Academy of sciences, {2. 100101, china 2. Graduate school of the Chinese Academy of sciences, {2. 100039, china 3. school of Geography, beijing normal university, {2. 100875, china Institute of Geographical sciences and Natural Resources Research Chinese Academy of Sciences Beijing 100101 China Graduate school of the Chinese Academy of sciences Beijing 100039 China school of Geography Beijing Normal University Beijing 100875 China

Based on the plant level data from the First National Census of Basic Units, the Second National Census of Basic Units, and the First National Economic Census of china, this paper ex- amines the geographical agglomeration of Chinese manufacturing industries at a more disaggregated level both by industrial classi-fication and spatial unit analysis. Applying the agglomeration index proposed by Maurel and Sedillot, studies show that the most agglomerated industries are resource-intensive, and capital and technology intensive while industries demanding localized inputs or serving localized markets or favored by local governments are fairly dispersed. The more disaggregated industries are more spa- tially agglomerated. At the finer spatial scale, industries are more dispersed. Results indicate that within-county spillovers are stronger than nearby-county spillovers, and within-prefecture spillovers are stronger than nearby prefecture spillovers. Local- ized spillovers are still quite substantial at a range beyond that of counties. Comparing the agglomeration index of Chinese manu- facturing industries in 1996, 2.01 and 2.04, agglomeration seems to be a general tendency. Results also indicate that some industries have experienced remarkable changes in their levels of agglom- eration in the period 1996–2.04.

关键词： industrial agglomeration manufacturing industries the MS index

DISCRIMINANT ANALYSIS BASED ON STATISTICAL DEPTH

在线全文

学校读者我要写书评

暂无评论

Journal of Systems Science & Complexity 2010年第2期23卷 362-371页

作者： Jiao JIN·Hengjian CUI school of Mathematical sciences,Key Laboratory of Mathematics and Complex Systems,Ministry of Education, {2.,beijing 100875,china. school of Mathematical sciences Key Laboratory of Mathematics and Complex Systems Ministry of Education

In the past two decades,many statistical depth functions seemed as powerful exploratoryand inferential tools for multivariate data analysis have been *** this paper,a new depthfunction family that meets four properties mentioned in Zuo and Serfling(2.00)is *** aclassification rule based on the depth function family is *** classification parameter b couldbe modified according to the type-Ⅰ error α,and the estimator of b has the consistency and achievesthe convergence rate n^(-1/2..With the help of the proper selection for depth family parameter c,theapproach for discriminant analysis could minimize the type-Ⅱ error β.A simulation study and a realdata example compare the performance of the different discriminant methods.

关键词： Depth discriminant analysis location parameter MVE scatter matrix.