导数

We investigate mainly solving the hyperbolic and parabolic equations by using the MQ quasi-interpolation. The underlying idea of our means is that: employing the derivative of the MQ quasi-interpolation to approximate the spatial derivative of the PDE, while the approach of the temporal derivative of the PDE is used a finite difference.

我们主要就MQ拟插值在双曲型方程和抛物型方程中的应用进行了研究，其基本思路是：利用MQ拟插值的导数逼近微分方程的空间导数，而微分方程的时间导数则采用有限差分逼近。

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。

First, the partial derivative is defined, based on the convergence in probability measure in the probability measure space.

本文首先在概率测度空间上，利用依概率测度收敛，定义了多随机变量函数的偏导数，并且偏导数的计算式可以通过替换实值函数的对应的偏导数中所有的实变量替换为随机变量得到。

Application of differential coefficient in inequation verification;2. It also elaborates what should be pay attention to while making use of this method to obtain differential coefficient .

本章给出了由分段函数直接求导数再取极限求函数在分段点处导数的方法，并就利用该方法求导数时需要注意的问题进行了说明。

By making use of the definition of derivative ,one can calculate the derivative of demarcation point ,the derioative of the special elementary function,the derivative of parametric equation.

利用导数的定义可求分界点的导数，特殊初等函数的导数和某些参数方程的导

With the help of the notation of the skew-symmetric tensor, the expression for the 1st order absolute and relative derivatives of a tensor was presented, and the expression for the 2nd order absolute and relative derivatives of a tensor was offered further.

由此，并借助于反对称张量，给出了张量的一阶绝对导数和相对导数之间的关系式，并又进一步给出了张量的二阶绝对导数和相对导数之间的关系。

Optimization; parametric quadratic convex programming; set-valued map; directional derivative; linear stability; solution-set map; parametric linear programming; error bound; subdifferential map; lower locally directionally Lipschitzian; upper locally di-rectionally Lipschitzian; locally directionally Lipschitzian; convex function; quasidiferential; kernelled quasidiferential; quasi-kernel; star-kernel; star-diferential; Penot diferential; subderivative; superderivative; epiderivative; set-valued optimization; set-valued analysis; subdifferential; optimization condition;ε-dual; scalization; generalized subconvexlike-cone;ε-Lagrange multiplier

基础科学，数学，运筹学最优化；集值映射；方向导数；线性稳定；最优解集映射；参数线性规划；参数凸二次规划；误差界；次微分映射；下局部方向Lipschitzian；上局部方向Lipschitzian；局部方向Lipschitzian；凸函数；拟微分；核拟微分；拟核；星核；星微分； Penot-微分；上导数；下导数； Epi-导数；集值优化；集值分析；集值映射的次微分；最优性条件；广义锥次类凸；ε-对偶；数乘；ε-Lagrange乘子

The horizontal and vertical first derivatives of magnetic anomalies of an infinite cylinder are calculated by the cosine transform method, in which the maximum errors are -028 nT/m and 047nT/m, respectively and the percent errors are generally within -357%～327% and -194%～188%, respectively except several data of the boundary and part are bigger because of remains of Gibbus effect. The calculating curve and theoretical curve are approximately coincident, and there is no influence by effective magnetic dip angle in computing.

利用余弦变换法计算的无限长水平圆柱体磁异常水平和垂向一阶导数的最大误差分别为-028 nT/m、047 nT/m；水平一阶导数的误差一般在-357%～327%之间，垂向一阶导数的误差一般在-194%～188%之间；计算的磁异常一阶导数值与理论值大致重合，而且不受有效磁化倾角的影响。

In model experiments, we find that the deviation of the first derivative calculated by Fourier transform is very large to comparing with the theoretical derivative, but the fitting effect of the derivative of anomalies calculated by cosine transform is very good. The calculation accuracy of data are all very high except that errors of several data on the boundary are large because of the residual Gibbus effect induced by finite truncation of gravity anomalies. Errors are -009%～5%.

模型实验中发现，用Fourier变换计算的一阶导数与理论导数偏差很大，而余弦变换计算的导数与理论异常导数拟合效果非常好，除边界几个数据因重力异常的有限截断产生的吉布斯效应残留使误差较大外，数据的计算精度均很高，误差为-009%～5%。

When the output branch number is more than 2, we can construct a cascade connection of Y-branch waveguides. However, it occupies a large optical circuit area. The power dividers with 2N output branches can be setup by arranging (2N-1) branch waveguide. However, the design of dividing structures with output branches of odd number such as three or five are difficult.

当输出波导数大於2时，可以用串接的方式来增加输出波导数，但是所需空间会变大，而光功率分配器的2N分岔输出是可以由(2N-1)的分岔输出结构改变而来的，所以在设计奇数波导输出像是3或是5的波导数输出结构就成为一个必须的课题。

It has been put forward that there exists single Ball point and double Ball points on the symmetrical connecting-rod curves of equilateral mechanisms.

从鲍尔点的形成原理出发，分析对称连杆曲线上鲍尔点的产生条件，提出等边机构的对称连杆曲线上有单鲍尔点和双鲍尔点。

The factory affiliated to the Group primarily manufactures multiple-purpose pincers, baking kits, knives, scissors, kitchenware, gardening tools and beauty care kits as well as other hardware tools, the annual production value of which reaches US$ 30 million dollars.

集团所属工厂主要生产多用钳、烤具、刀具、剪刀、厨具、花园工具、美容套等五金产品，年生产总值3000万美元，产品价廉物美、选料上乘、质量保证，深受国内外客户的青睐