收敛的阶

Though comparing Canny operator and center B spline dyadic wavelet, the following conclusion is proven in this dissertation: a Center B spline function has tight support and Canny operator hasn't. b Center B spline function asymptotic convergence to Gaussian function and the derivative of Center B spline function asymptotic convergence to Canny operator. c The derivative of fourth order center spline B function is more suitable as a optimal edge detector than Canny operator. d Center B spline function can balance the smoothing and approximation of original data, and the fourth center B spline function is the only optimal solution of two order smoothing problem. e The error between the valve of time-frequency uncertainty of the fourth center B spline function and the lower bound of time-frequency uncertainty does not exceed 0.143% of the lower bound. f The derivative of center spline B function can construct a stability dyadic wavelet and can give a fast algorithm for multiscale edge detection, but Canny operator can do neither.

作者给出了Canny算子与中心B样条二进小波严格的比较证明，得出如下结论：a中心B样条函数具有紧支集，Canny算子不具有紧支集。b中心B样条函数的极限收敛于高斯函数，中心B样条函数的导数收敛于Canny算子。c四阶中心B样条函数的导数比Canny算子更接近最佳边缘检测滤波器。d中心B样条函数比高斯函数更能兼顾对原函数平滑和逼近的折中要求，并且四阶中心B样条函数是二阶逼近问题的唯一最优解。e四阶中心B样条函数的时频测不准关系值与时频测不准关系下界的逼近误差不超过0.143％。f中心B样条函数的导数可以构成稳定的二进小波，存在快速的多尺度算法；而Canny算子不构成稳定的二进小波，无法给出快速的多尺度算法。

The main theory results includes:(1) Using the properties of Hilbert transform, perfectly reconstruction and new type of lifting scheme, a new type of dual-tree binary coefficients complex wavelet with linear phase is achieved.(2) For linear systems that can be diagonalized by GFT and DST-II matrices, an efficient MGM method is proposed, convergence is proved.(3) We discuss the algebraic structure when Toeplitz matrix is transformed by multi-band wavelet,show that Toeplitz matrix is composed of generating function is transformed to a band and sparse matrix when wavelet applied to this matrix, based on the above results, an efficient solution of Toeplitz equations is obtained, and the computational complex is O,where N is the order of matrix.

理论成果主要包括：(1)对于对偶树二进制系数复数小波，利用Hilbert变换对性质、完全重构条件并结合新的提升格式构造研究了含参系数多进制小波构造方法，作为特例得到具有线性相位的对偶树二进制系数复数小波构造方法；(2)对于广义离散傅立叶变换与正弦变换对角化系统，提出了高效、快速的多重网格算法，理论上证明了算法的收敛性；(3)研究了Toeplitz矩阵在多进制小波变换下的代数结构，验证了多项式生成函数构成的Toeplitz系统在小波变换下的稀疏带宽性质，从而建立基于小波变换求解Toeplitz系统的快速求解方法，运算量级控制在O，其中N为系统的阶。

In this paper perturbation of eigenvalue and eigenfunction for higher order ordinary differential equation in the case of boundary perturbation combined with operator perturbation we considered.

本文研究了在边界和算子摄动相结合的情况下高阶常微分方程特征值与特征函数的摄动，把摄动问题A_6，μ的特征值λ_ε，μ和特征函数y_6，μ按两参数ε、μ的次幂展开，并且证明这些展开式当ε、μ充分小时是收敛的。

In this paper, we study the rate of convergence of Durrmeyer-Bézier Operaters for functions of bounded variation f and obtain a accurate estimation of coefficient. Our result improves the result of Zeng by making use of probable property of basic functions.

在Zeng等人对有界变差函数f的Durrmeyer-Bézier算子在区间(0，1)上收敛于 1/(a+1f+a/(a+1f 的收敛阶进行研究的基础上，利用基函数的概率性质等方法，对其所给的Durrmeyer-Bézier算子收敛阶估计结果作进一步的改进，得到其收敛阶的精确估计。

The optimal convergence rates in H-1 and L-2 norms are proved, and the convergence order in L-2 norm is one order higher than that in H-1 norm.

并分别给出了解的H-1模和L-2模的最优收敛阶估计， L-2模收敛阶比H-1模收敛阶高一阶。

The resulting linear system, the coefficient matrix of which is, typically, a 2 x 2 block symmetric or non-symmetric and singular or non-singular indefinite matrix, is of high order and ill conditioned. In order to solve such system we have to use iterative methods with fast convergence speed.

所得的离散方程组（其系数矩阵典型地是 2 x 2 块对称或非对称，奇异或非奇异的不定矩阵）是高阶和病态的，为了求解这样的方程组我们必需使用快速收敛的迭代方法。

We have known that the finite element solution can get the convergence accuracy order of m+1 in global interval and of 2m in the nodes.

我们已知有限元解在整体区间上能达到m+1阶的收敛精度，节点处能达到2m阶的收敛精度。

In this project,we primarily investigated the growth of solutions of some types of differential equations,and the hyper order,the exponents of convergence of zeros and fixed points of solutions with infinite order, especially, expononts of convergence of fixed points of first, second order derivatives and differential polynomials of solutions.

中文摘要：在本项目中，重点研究了几种类型的微分方程解的增长性，及无穷级解的超级，零点收敛指数，不动点收敛指数，特别是解的一阶，二阶导数及微分多项式的不动点收敛指数。

In this paper,based on an improved orthogonal expansion in an clement , using the new idea of Ref.[3] ,a new error expression of n-degree Hermite finite element approximation to one-dimensional 4-degrec 2-point bounded problem and 2-degree ordinary differential problem, and then optimal order superconvergence for their first derivatives is obtained.

本文针对在改进的单元正交性估计的基础上，利用文[3]提出的新想法，得到一维四阶两点边值问题和二阶常微初值问题的n次赫米特有限元u_h∈C~1的新误差估计式，以及导数误差的最佳阶超收敛，并且两者有相同的超收敛结果。

With the guide of non-linear program theory, by using interpolative steps and inexact line search, an improved conjugate gradient method was found, by which the training rate of BP networks increases by tens or hundreds of times. Moreover, the improved method is effective to solve non-linear equations for which Newton's method does not converge owing to the problems of the quadratic derivative and inverse matrix.

通过大量的数值模拟试验发现，在非线性规划理论的指导下采用间插步骤和不精确的一维搜索技术改进的共轭梯度法，是基于梯度和共轭方向的连续搜索算法中最有效的算法，这种算法使BP网络的训练速度提高几十到几百倍，使BP网络的实际应用效果大为改善；而且这种算法对于用牛顿法由于求二阶导数和求逆矩阵等问题难于收敛的非线性方程组的求解也是很有效的。

They weren't aggressive, but I yelled and threw a rock in their direction to get them off the trail and away from me, just in case.

他们没有侵略性，但我大喊，并在他们的方向扔石头让他们过的线索，远离我，以防万一。

In slot 2 in your bag put wrapping paper, quantity does not matter in this case.

在你的书包里槽2把包装纸、数量无关紧要。