高数 07.02 偏导数

$\color{blue}{\S 第七章 第二节 偏导数}$

一、偏导数概念及其计算
二、高阶偏导数

$\color{blue}{一、偏导数概念及其计算法}$

$定义1.设函数z = f(x, y)在点(x_0, y_0)的某邻域内极限\\ \lim_{\Delta{x} \rightarrow 0}{\dfrac{f(x_0 + \Delta{x}, y_0) - f(x_0, y_0)}{\Delta{x}}} 存在,则称此极限为函数\\ z = f(x, y)在点(x_0, y_0)对x的偏导数,记为\\ \left. \dfrac{\partial z}{\partial x} \right|_ {(x_0, y_0)}; \left. \dfrac{\partial f}{\partial x} \right|_ {(x_0, y_0)}; \left. z_x \right|_ {(x_0, y_0)}; f_x(x_0, y_0); f_1^{\prime}(x_0, y_0).$

$同样可以定义对y的偏导数\\ f_y(x_0, y_0) = \lim_{\Delta{y} \rightarrow 0}{\dfrac{f(x_0, y_0 + \Delta{y}) - f(x_0, y_0)}{\Delta{y}}} \\ = \left. \dfrac{d}{dy}f(x_0, y) \right|_ {y = y_0}$
$若函数z = f(x, y)在域D内每一点(x, y)处对于x或y偏导数存在,\\ 则该偏导数称为偏导函数,也简称偏导数,记作\\ \dfrac{\partial z}{\partial x}， \dfrac{\partial f}{\partial x}, z_x, f_x(x, y), f_1^{\prime}(x, y) \\ \dfrac{\partial z}{\partial y}， \dfrac{\partial f}{\partial y}, z_y, f_y(x, y), f_2^{\prime}(x, y)$

$偏导数的概念可以推广到二元以上的函数.$
$例如,三元函数u = f(x, y, z)在点(x, y, z)处对x的偏导数定义为\\ f_x(x, y, z) = \lim_{\Delta{x} \rightarrow 0}{\dfrac{f(x + \Delta{x}, y, z) - f(x, y, z)}{\Delta{x}}} \\ f_y(x, y, z) = \lim_{\Delta{y} \rightarrow 0}{\dfrac{f(x , y+ \Delta{y}, z) - f(x, y, z)}{\Delta{y}}} \\ f_z(x, y, z) = \lim_{\Delta{z} \rightarrow 0}{\dfrac{f(x , y, z+ \Delta{z}) - f(x, y, z)}{\Delta{z}}}$

$例1.求z = x^2 + 3xy + y^2在点(1, 2)处的偏导数.$
$解:\\ 解法1\\ \dfrac{\partial z}{\partial x} = 2x + 3y \\ \dfrac{\partial z}{\partial y} = 3x + 2y \\ \left. \dfrac{\partial z}{\partial x} \right|_ {(1, 2)} = 2 \cdot 1 + 3 \cdot 2 = 8 \\ \left. \dfrac{\partial z}{\partial y} \right|_ {(1, 2)} = 3 \cdot 1 + 2 \cdot 2 = 7 \\ 解法2：\\ \left. \dfrac{\partial z}{\partial x} \right|_ {(1, 2)} = \lim_{\Delta{x} \rightarrow 0}{\dfrac{f(1 + \Delta{x}, 2) - f(1, 2)}{\Delta{x}}} \\ = \lim_{\Delta{x} \rightarrow 0}{\dfrac{[(1 + \Delta{x})^2 + 3 \cdot (1 + \Delta{x}) \cdot 2 + 2^2] - [1^2 + 3 \cdot 1 \cdot 2 + 2^2]}{\Delta{x}}} \\ = \lim_{\Delta{x} \rightarrow 0}{(8 + \Delta{x})} \\ = 8 \\ \left. \dfrac{\partial z}{\partial y} \right|_ {(1, 2)} = \lim_{\Delta{y} \rightarrow 0}{\dfrac{f(1, 2 + \Delta{y}) - f(1, 2)}{\Delta{y}}} \\ = \lim_{\Delta{y} \rightarrow 0}{\dfrac{[1^2 + 3\cdot1\cdot(2+ \Delta{y}) + (2 + \Delta{y})^2] - [1^2 + 3 \cdot 1 \cdot 2 + 2^2]}{\Delta{y}}} \\ = \lim_{\Delta{x} \rightarrow 0}{(7 + \Delta{y})} \\ = 7$

$例2.设z = x^y (x > 0, 且 x \neq 1),求证\dfrac{x}{y} \dfrac{\partial z}{\partial x} + \dfrac{1}{\ln{x}} \dfrac{\partial z}{\partial y} = 2z$
$证:\\ \because \dfrac{\partial z}{\partial x} = yx^{y - 1}; \dfrac{\partial z}{\partial y} = x^y \ln{x}\\ \dfrac{x}{y} \dfrac{\partial z}{\partial x} + \dfrac{1}{\ln{x}} \dfrac{\partial z}{\partial y} \\ = \dfrac{x}{y} \cdot yx^{y -1} + \dfrac{1}{\ln{x}} \cdot x^y\cdot \ln{x} \\ = x^y + x^y \\ = 2z$

$例3.设z = xy + x^3,求\dfrac{\partial z}{\partial x} + \dfrac{\partial z}{\partial y}$
$解:\\ \dfrac{\partial z}{\partial x} = y + 3x^2 \\ \dfrac{\partial z}{\partial y} = x \\ \dfrac{\partial z}{\partial x} + \dfrac{\partial z}{\partial y} = 3x^2 + x + y$

$例4.设f(x, y) = \sin{x^2y},求f_x^{\prime}(x, y),f_y^{\prime}(x, y).$
$解:\\ f_x^{\prime}(x, y) = (\sin{x^2y}) _x^{\prime} = \cos{x^2y} \cdot 2xy = 2xy \cos{x^2y} \\ f_y^{\prime}(x, y) = (\sin{x^2y}) _y^{\prime} = \cos{x^2y} \cdot x^2 = x^2 \cos{x^2y}$

$\color{blue}{二、高阶偏导数}$

$设z = f(x, y)在域D内存在连续的偏导数\\ \dfrac{\partial z}{\partial x} = f_x(x, y), \dfrac{\partial z}{\partial y} = f_y(x, y) \\ 若这两个偏导数仍存在偏导数,则称它们是\\ z = f(x, y)的二阶偏导数.按求导顺序不同,有\\ 下列四个二阶偏导数:$
$\dfrac{\partial }{\partial x}(\dfrac{\partial z }{\partial x}) = \dfrac{\partial ^2 z}{\partial x^2} = f_{xx}(x, y); \\ \dfrac{\partial }{\partial y}(\dfrac{\partial z }{\partial x}) = \dfrac{\partial ^2 z}{\partial x \partial y} = f_{xy}(x, y); \\ \dfrac{\partial }{\partial x}(\dfrac{\partial z }{\partial y}) = \dfrac{\partial ^2 z}{\partial y \partial x} = f_{yx}(x, y); \\ \dfrac{\partial }{\partial y}(\dfrac{\partial z }{\partial y}) = \dfrac{\partial ^2 z}{\partial y^2} = f_{yy}(x, y);$

$类似可以定义更高阶的偏导数$
$例如,z = f(x, y)关于x的三阶偏导数为\\ \dfrac{\partial }{\partial x}(\dfrac{\partial ^2 z}{\partial x^2}) = \dfrac{\partial ^3 z}{\partial x^3}$

$z = f(x, y)关于x的n - 1阶偏导数,再关于y的一阶偏导数为\\ \dfrac{\partial }{\partial y}(\dfrac{\partial ^{n - 1} z}{\partial x^{n -1}}) = \dfrac{\partial ^n z}{\partial x^{n - 1} \partial y}$

$例1.求函数z = e^{x + 2y}的二阶偏导数及\dfrac{\partial ^3 z}{\partial y \partial x^2}$
$解：\\ \dfrac{\partial z}{\partial x} = e^{x + 2y} \\ \dfrac{\partial z}{\partial y} = 2e^{x + 2y} \\ \dfrac{\partial^2 z}{\partial x^2} = (e^{x + 2y})_x^{\prime} = e^{x + 2y}\\ \dfrac{\partial ^2 z}{\partial x \partial y} = (e^{x + 2y})_y^{\prime} = 2e^{x + 2y} \\ \dfrac{\partial ^2 z}{\partial y \partial x} = (2e^{x + 2y})_x^{\prime} = 2e^{x + 2y} \\ \dfrac{\partial^2 z}{\partial y^2} = (2e^{x + 2y})_y^{\prime} = 4e^{x + 2y}\\ \dfrac{\partial ^3 z}{\partial y \partial x^2} = \dfrac{\partial }{\partial x}(\dfrac{\partial ^2 z}{\partial y \partial x}) = 2e^{x + 2y}$

$例2.求函数z = x^2y -\sin{xy}的二阶偏导数$
$解:\\ \dfrac{\partial z}{\partial x} = 2xy - y\cos{xy} \\ \dfrac{\partial z}{\partial y} = x^2 - x\cos{xy} \\ \dfrac{\partial ^2 z}{\partial x^2} = 2y + y^2\sin{xy} \\ \dfrac{\partial ^2 z}{\partial x \partial y} = 2x - \cos{xy} + xy\sin{xy} \\ \dfrac{\partial ^2 z}{\partial y \partial x} = 2x -\cos{xy} +xy\sin{xy} \\ \dfrac{\partial ^2 z}{\partial y^2} = x^2 \sin{xy}$

$例3.求函数z = xe^x\sin{y}的二阶导数$
$解:\\ \dfrac{\partial z}{\partial x} = (x + 1)e^x\sin{y} \\ \dfrac{\partial z}{\partial y} = xe^x\cos{y} \\ \dfrac{\partial ^2 z}{\partial x^2} = (x + 2)e^x\sin{y} \\ \dfrac{\partial ^2 z}{\partial x \partial y} = (x + 1)e^x \cos{y} \\ \dfrac{\partial ^2 z}{\partial y \partial x} = (x + 1)e^x\cos{y} \\ \dfrac{\partial ^2 z}{\partial y^2} = -xe^x \sin{y}$

$例4.函数z = f[\ln(x^2 + y^2)],求\dfrac{\partial ^2 z}{\partial x \partial y }$
$解:\\ \dfrac{\partial z}{\partial x} = f^{\prime}[\ln(x^2 + y^2)] \cdot \dfrac{2x}{x^2 + y^2} \\ \dfrac{\partial ^2 z}{\partial x \partial y } = f^{\prime \prime}[\ln(x^2 + y^2)]\cdot \dfrac{2y}{x^2 + y^2} \cdot \dfrac{2x}{x^2 + y^2} + f^{\prime}[\ln(x^2 + y^2)] \cdot (\dfrac{2x}{x^2 + y^2})_y^{\prime} \\ = \dfrac{4xy}{(x^2 + y^2)^2} \lbrace f^{\prime \prime}[\ln(x^2 + y^2)] - f^{\prime}[\ln(x^2 + y^2)] \rbrace$

$定理.若f_{xy}(x, y)和f_{yx}(x, y)在点(x_0, y_0)连续,则\\ f_{xy}(x_0, y_0) = f_{yx}(x_0, y_0)$

$本定理对n元函数的高阶混合导数也成立.$
$例如,对三元函数u = f(x, y, z),当三阶混合偏导数在点(x, y, z)\\ 连续时,有\\ f_{xyz}(x, y, z) = f_{yzx}(x, y, z) = f_{zxy}(x, y, z) \\ = f_{xzy}(x, y, z) = f_{yxz}(x, y, z) = f_{zyx}(x, y, z)$

内容小结
1.偏导数的概念机有关结论
定义；记号；
函数在一点偏导数存在，函数在此点不一定连续
$混合偏导数连续 \Longrightarrow 与求导顺序无关$
2.偏导数的计算方法
$求一点处偏导数的方法 \left \{ \begin{array}{l}先代后求 \\ 先求后代 \\ 利用定义 \end{array} \right.$
求高阶偏导数的方法 –逐次求导法