估计

The dissertation turns out the software of Helmert variance component estimation and applied it into Sutong Bridge GPS control network adjustment. The control points coordinate precision are increased, it shows that Helmert variance component estimation is valuable to GPS control network adjustment in practice.

本文完成了Helmert方差分量估计软件的研制，并把该估计方法成功应用于苏通大桥GPS控制网平差处理中，经过验后估计的网点坐标精度得到提高，说明了赫尔默特方差分量估计在GPS网平差中的实用价值。

The results showed that:(1) we can mapping quantitative trait locus while estimating the variance component of QTL;(2) granddaughter design is better than daughter design when mapping QTL;(3) it is easy to map a QTL for trait with a high heriability and a large QTL variance contribution;(4) we can estimate the variance component of a QTL by TM-BLUP based on ML method whether the QTL has only 2 alleles or QTL has normal distributed alleles effects;(5) the estimation accuracy of variance component contributed by QTL was improved by using of grand daughter design;(6) the higher the heritability and the QTL variance contribution was, the more accurate estimation of QTL variance component.

结果表明：（1）采用随机QTL效应模型和最大似然法，在估计QTL方差组分的同时，能够定位QTL；（2）孙女设计与女儿设计相比，在其它因素相同时，容易检出QTL；（3）遗传力高，QTL方差贡献较大的性状，QTL检出的效果优于遗传力低，QTL方差贡献较小的性状；（4）无论QTL上有2个等位基因，还是QTL上等位基因的效应服从正态分布，都可将其看作随机效应，采用基于TM-BLUP的ML法，估计其方差组分和定位QTL；（5）QTL方差组分估计的准确性，孙女设计高于女儿设计；（6）遗传力高的性状，QTL方差贡献大的QTL，QTL方差组分估计的准确性高。

In mathematical that is to say,after obtaining the sampleχ,we may obtain the conditional distribution Hwhen X=χ.

在传统的Bayes估计中，参数λ的先验分布都是已知的，但有时候我们不一定能得到λ的精确的先验分布，于是就产生了新的Bayes估计方法，比如E Bayes估计和多层Bayes估计。

Additionally, this research used One-Pass Method for first time to estimate jointly the conditional mean equation and conditional variance equation of Bivariate GARCH Model to avoid estimating error in previous relative studies with Two-Pass Method.

此外，本文在进行模型估计时，首度采用一阶段估计法，来联合估计双变量GARCH模型中的条件平均数方程式与条件变异数方程式，以避免过去相关文献将两条方程式个别估计时所造成的估计误差。

Third, the discussion of the corresponding relationship of the real angles to the apparent real angles is performed, the asymptotic bias and MSE of real angles are given. Fourth, based on above conclusions, the analysis of two typical cases is made by numerical analysis and computer simulation: Case A when the impinging signals are divided into desired signals and interfering signals: it is shown that similar to Cyclic estimation of DOA, the smaller the separation between interfering direction and signal direction, in general, the smaller the estimation variance.

首先，研究了无穷快拍时信号数目有误的最大似然估计，引入了视在真实角度的概念；其次，导出了视在真实角度最大似然估计大样本时的分布以及渐近均方误差的解析表达式，而且还证明了：Stoica性能分析结果是现在的分析结果在信号数目准确时的特例；第三，讨论了真实角度与视在真实角度的关系，给出了真实角度的渐近偏和渐近均方误差；第四，依据前面的结果，利用数值分析和模拟实验的方法对两种典型情况做了分析： 1 在入射信号被分为干扰信号和所需信号情况下：结果表明：与CyclicDOA估计一样，干扰方向与所需信号方向靠的越近，所需信号估计的方差越小，越远离，方差越大。

About the problem of hypothesis testing for the problem of estimate interval,first of all,we set statistic and the distributions of the statistic.Fron this ,we derivate the methods of testing.

对于点估计问题，讨论了估计的优良性质；对于区间估计问题，讨论了估计的精度；关于参数的检验问题，首先建立检验用的统计量及其分布，由此导出检验的方法。

In particular, for recurrent event data, we proposed a novel nonparametric estimator of the survival function and obtained an efficient estimator of the expected total number of events for counting processes. Also we established asymptotic properties of these estimators and investigated nonparametric estimation of survival functions, density functions and hazard functions under various types of incomplete data.

在重复事件发生的情况下，创新地提出了一种生存函数的非参数估计，获得了计数过程中预期事件总数的估计以及这些估计的大样本性质，并在其他各种不完全数据下对生存函数，概率密度和失效率函数等非参数估计进行了深刻研究。

Thirdly, on the base of the existent error estimate of elliptical Galerkin projection operator, obtain semi-discrete and complete-discrete error estimates betwent EFG solution and exact solution of parabolic partial differential equation. Semi-discrete error estimate show that the rank of r is accordance with the approximation rank of subspace S_h. Complete-discrete error estimates show that it not only have relation with radius r of domain of influence, but also have relation with the step length of time variable and the way of discrete.Lastly, give typical examples and draw up procedures of MATLAB.

再次，在"椭圆Galerkin投影"算子的误差估计的基础上，对用EFG法解抛物型偏微分方程的EFG解与精确解之间作了半离散和全离散的误差估计，半离散的误差估计表明所给出的误差界限关于r的阶是与子空间S_h的逼近阶相一致的，全离散的误差估计表明所产生的误差不但与影响域半径r有关，而且与离散时间变量的步长τ及其离散方式有关。

The dissertation focuses on performance gradient estimation and SRO algorithm of DES aswell as other related topics,such as modeling of DES and discrete event simulation,performancemeasure Hessian matrix estimation,Stochastic Approximationand so on.General Semi-Markov Processis adapted to describe behaviour of DES and discrete eventsimulation.A general result is obtained and applied to analyze strong consistency of performancegradient Perturbation Analysisestimator.Conditions for effectiveness of InfinitesimalPerturbation Analysisare also discussed.New PA algorithms,such as server PA,PA usingDirac generalized function,job sequence PA etc.,are proposed to deal with discreteparameters.Furthermore,strategies of combining PA algorithms with certain iterative rules areanalyzed widely in order to optimize DES parameters and some SRO algorithms are constructed toslove DES optimization problems.

论文将DES性能分析领域中性能梯度估计和单运行仿真优化作为主要的研究对象，探讨了与性能梯度估计有关的DES和离散事件仿真过程建模、高阶导数估计以及随机逼近等理论和方法；建立了具有一般性的DES和离散事件仿真的广义半Markov过程模型；获得了分析性能梯度估计量强相合性的理论结果和无穷小摄动分析算法应用的条件；提出了诸如服务台数目的摄动分析、奇异摄动分析、串联生产线和加工车间中工件加工序的摄动分析等一系列摄动的思想和算法，对连续参数和离散参数都进行了深入的研究；对摄动分析算法同优化规则结合实现仿真优化的策略也进行了广泛的分析，构造了新型的变界截尾的随机逼近算法，并用于DES的单运行仿真优化研究；最后，给出了DES性能分析方法的实证性研究结果。

From the results of expanded state estimation application, which is a representative parameter estimation method proposed before, it often happened that the parameters could not be estimated or estimated imprecisely.

扩展状态估计法是一种有代表性的参数估计方法，在该算法的实际应用中，发现支路参数不可估计或者估计结果不合理的现象时有发生。

Methods: Five patients with parkinsonism or dystonia were assigned to general anesthesia using an modified endotracheal tube.

本实验依照人体实验之相关规定进行，五位患有帕金森氏症或肌张力异常的病人接受神经立体定位手术。

If you can benefit from this book, it is our honour.

如果您能从本书获益，这将是我们的荣幸。