分析集

Based on a kind of orthogonal complete functions in L2[0,1], another one (called V-system) which has equivalence relation with U-system is created. It is a kind of new and effective mathematical tool on signal multi-resolution analysis. Besides that V-system keeps the favorable properties of U-system and can accurately reconstruct geometry information which expressed in piecewise polynomials by finite sum of functions in V-system, it has more special properties, such as simple construction, abundant levels, convenience of application, quick calculation, local support, and so on. V-system is the generalization of Haar orthogonal functions, and to be a class of wavelet base. It has a good future for some applications in the fields of wavelet analysis and signal processing.

从L2[0，1]空间的一类正交完备函数系U系统出发，构造了另一类与之等价的正交完备函数系，称之为V系统，它是信号多分辨分析方面的一种新型的、有效的数学工具。V系统不仅保持了U系统的优良特性，对多项式表达的几何信息能够做到有限项精确重构，并且较之U系统，它更有结构简单、层次分明、计算快捷、局部支集等特点，应用起来将更加灵活方便。V系统可以看作是Haar函数系的推广，是一类小波基，在某些数字信号处理及小波分析问题中有良好的应用前景。

During the migrating and optimizing process of targeted platform, this paper starts with the analysis on complexity of H.264/AVC algorithms, carries on plenty of tests, uses a series of optimizing methods aiming at the special hardware resources of PWBSP-16 platform: optimizing modules including SAD Calculate, Integer Transform and Quantization in the instruction level by using FIRtree instruction set; optimizing it by improving coding process and enabling co-processor such as VLx、DMA in system level by the thinking of modules parallelism maximizing; putting forward the future optimizing mind in the level of algorithms.

在目标平台的移植优化过程中，本文从H.264/AVC编码算法的复杂度分析入手，进行了大量的测试分析，并针对PWBSP-16 DSP处理器特殊的硬件资源，采取了如下一系列的优化策略：在指令级方面，采用FIRtree指令集对SAD计算、整型变换、量化等主要耗时模块进行优化；在系统级方面，采用模块并行最大化的思想，通过改进编码流程，启用VLx、DMA等协处理器等方法进行优化；在算法级方面，提出后续的优化思路。

And by using the single square root operator for wavefield extrapolation and the windowed Fourier frame expansion and reconstruction of wavefield, a method for local angle domain common image gather is proposed, i.e.

提出的波动方程反射角度域共成像道集方法可为进一步的叠前偏移数据振幅随角度变化分析和偏移速度分析工作提供基础。

Firstly, the research studies the coding rule of the PDF417 barcode and designs related computer code algorithm, with making it programmed, and completes the coding part of the system software. Next, the coding and decoding principle of Reed-Solomon error-correcting code is briefly introduced. According to thecoding and decoding principle in galois field GF(929), to analyze and study its practical application with emphasis in this topic. Then, the PDF417 barcode decoding process is analyzed and studied detailed according to the barcode coding principle, with the corresponding realization method designed, improving the design in theory aspect by using material example, at the same time programming the design method to complete the decoding part of the system. Finally, validated the system program in the practical application of image information aspect, by using the nonrepeatability and the randomicity characteristic of Julia mark shape artwork.

本课题首先研究PDF417条码的编码规则并设计相关的计算机编码算法使其程序化，完成系统软件的编码部分；其次，简要介绍Reed—Solomon错误控制码的编码与译码原理，在伽罗瓦域GF(929)内根据其编译码的原理重点分析与研究其在本课题中的实际应用；然后根据条码的编码原则对PDF417条码的译码过程进行详细的分析与研究，并设计出相应的实现方法，利用具体实例对设计方法进行理论方面的实践同时将设计方法程序化完成系统软件的译码部分；最后利用Julia集分形图形的不可重复性和随机性的特点对系统软件在图像信息方面的实际应用进行了验证。

The connection degree is used for describing indeterminateness of the system,thus the assessment of the system is realized.

集对分析是一种处理不确定性问题的系统分析方法，在一个系统中确定性与不确定性相互联系、影响、制约，并在一定条件下相互转化，用联系度来描述系统的不确定性，从而实现系统的评价。

Finally, it analyses the time complexity of the algorithm and researches how it is influenced by culture operator.3、The paper presents the evaluation standard of the GA's application capability. Basing on characters such as continuity, multi-peak, vibration, randomicity as well as large-scale, five functions are selected to test search ability and robustcity of co-evolution algorithm. Finally, it analyses the simulation result and researches the influence of algorithm brought by culture operators.4、Basing on the concept of collection overcast, it researches the task distribution issue and constitutes delaminated math model on task distribution issue. It puts out the co-evolution algorithm of subtask's decomposing. The experiment compare IGA、SGA to CN and validates the efficiency of co-evolution algorithm on the NP completeness issue.5、Being aim at the optimization issue of load of antenna near ground, it combines many GA strategies and puts forward strategic meme. And it puts out co-evolution algorithm of load of antenna design. And it emulates the optimization design of load of antenna near ground. Finally, it valuates the co-evolution algorithm's efficiency on the continuum search issues of multi-variable and multi-peak value.6、Being aiming at the knowledge of image model matching, it adopts single meme and real code. It puts out fast co-evolution matching algorithm strategy. Basing on NPROD resemble measurement, I emulate the indiscrimination model matching and discrimination model matching. Finally, it valuates the co-evolution algorithm's efficiency on the real code and real time search issue.

分析了算法的时间复杂度，研究了文化算子对算法时间复杂度的影响。3、提出了GA的应用性能评价标准，从连续性、多峰性、随机性、振荡性、广域性多个角度出发选择五个测试函数，对共同进化算法的搜索性能和鲁棒性作了函数优化的性能测试，分析了仿真结果，研究了文化算子对算法的影响。4、基于集覆盖的概念，研究了MAS中的任务分配问题，建立了任务分配问题的分层数学模型，给出了子任务分解共同进化算法，实验比较了IGA、SGA、CN，验证了共同进化算法对NP完全问题的有效性。5、针对近地天线加载优化设计问题，结合多种改进GA策略，提出了策略型拟子，给出天线加载设计的共同进化算法，对有耗半空间对称偶极子天线加载优化设计作了仿真实验，验证了共同进化算法对多变量多峰连续搜索问题上的有效性。6、针对图像模板匹配问题的领域知识，采用单类拟子和实数编码，给出快速共同进化匹配算法策略，基于NPROD相似度测度，仿真试验了无差别和有差别模板匹配，验证了共同进化算法对实数编码和实时性搜索问题上的有效性。

In the following, this dissertation does some research works about the other system—HDS with uncertainties of polytope using the existent methods, and proves that the system is exponential stable with respect to its invariant set under some conditions making use of Lyapunov function.

然后，本文针对另一类具体的混合动态系统——具有多面体不确定性的混合动态系统，沿用已有的稳定性分析方法对其模型进行分析，最终借助Lyapunov函数证明了在一定的条件下系统关于其不变集是指数稳定的。

First, it begins with the study of the centralized CFAR detection with multiple sensors. The non-CFAR problem is analyzed; the necessity for local CFAR processing is established. The effective local CFAR processing scheme is proposed, i.e. the local test statistic. The upper limit of the optimal performance is proved for the distributed CFAR detection based on LTS. Then, the new schemes based on LTS are analyzed in condition of signal pulse, and then the analysis is expanded to the case of multiple pulses. Next, the feedback mechanism is introduced. At last, the invariant test is used to reveal the essence of CFAR processing. It explains the necessity of local CFAR processing in the distributed CFAR detection at a higher level.

本文深入研究了这方面的问题，首先从多传感器集中式 CFAR 检测入手，分析其存在的非 CFAR 问题，论述局部 CFAR 处理的必要性，提出有效的局部 CFAR 处理方法：局部检测统计量，证明基于局部检测统计量的分布式 CFAR 检测的最优性能上限，分析基于局部检测统计量的新方案在单脉冲条件下的性能，并且进一步拓展到多脉冲非相干积累的情况，然后引入反馈机制，最后利用不变检验的概念揭示 CFAR 处理的机理，从更高的层次上解释分布式 CFAR 检测做局部 CFAR 处理的必要性。

In sample data, we use the database of Taiwan Finance Database of Taiwan Economic Journal and collect 14,125 company's' finance data which include 153 default company's' data to study. We use these financial data to make Variance Inflation Factor test, and use the result of credit scoring model to make Kolmogorov–Smirnov test and Receiver Operating Characteristic check. Finally, the conclusions are listed as following

在样本资料中，主要是用台湾经济新报中，上市上柜与曾经上市上柜公司资料库，其最后，整理出来的资料结果，共有14125笔公司财务资料可以进行完整的实证分析，违约公司资料有153笔；并将这此公司的财务变数进行线性相关变异数影响因子检定，并将评分模型所得出的结果进行K–S检定与ROC验证，最后再将评分结果进行集区分析以及评等稳定度的检定，其结果所得出的结果包括下列几点

Aimed at the characteristic of multiple types of faults possibly happened in nuclear power plant, large scale of training sample, and requirement of quick and accurate diagnosing, after the analysis of large sample number generated by large fault type exceeds limitation of 64K paragraph in DOS and large number of local minmum in error surface, measurements of same format of data file, Win95/NT operating system platform, resetting weight learning rate, dynamic training set in quick learning algorithm and improvement on quick learning algorithm using homotopy method which can avoid local minmum points in error surface have been adopted to ensure quickly and effective process of the course of neural network's training and testing.

针对核动力装置可能发生的故障种类多，训练样本规模大，故障诊断需要快速准确的特点等，分析了样本多和故障种类多产生的训练样本量超出DOS段大小和网络误差曲面上局部极小值多的情况，采取了使用相同格式数据文件、Win95/NT操作系统平台，对快速学习算法采用重置神经网络权值学习率和动态训练集、并采用能够有效克服网络误差曲面上局部极小点的同伦方法对学习算法进行改进等措施保证神经网络训练测试过程的快速、有效地进行；针对装置发生的故障须快速、准确诊断的需要，分析了故障的产生对装置参数变化的影响及操纵员对故障诊断的基础，在核动力装置发生故障时参数曲线的变化量与正常运行时参数曲线的变化量存在明显差异的基础上，提出采用参数曲线的变化量作为神经网络的输入，并围绕参数变化量的方法采用二次曲线拟合滤波求变化量和阈值技术来保证神经网络得到精确的装置参数变化量，从而得到准确的诊断结果。

In the United States, chronic alcoholism and hepatitis C are the most common ones.

在美国，慢性酒精中毒，肝炎是最常见的。

If you have any questions, you can contact me anytime.

如果有任何问题，你可以随时联系我。