在求F'' x的时候可以在F'x基础上求。
也可以在前面的中间步骤求二阶导。
F(x,y,z)=0,求 $$\dfrac{\partial z}{\partial x},\dfrac{\partial z}{\partial x}$$
关键:由题意分清自变量和函数! 然后对自变量偏导得到方程组。
自变量x,y 函数z 。自变量偏导得到方程组:
$$\begin{cases} F_x'+F_z'\cdot \dfrac{\partial z}{\partial x} =0 \\ \\ F_y'+F_z'\cdot \dfrac{\partial z}{\partial y} =0 \end{cases}$$
轻易可解得 $$\dfrac{\partial z}{\partial x}=-\dfrac{F_x'}{F_z'},\dfrac{\partial z}{\partial x}=-\dfrac{F_y'}{F_z'}$$
公式可以避免形如 $$z^x=y^z$$变为超越函数无法直接对x求导的尴尬。见 七、隐函数全微分
1、一元隐函数F(x,y)=0 , $$\frac{dy}{dx}=-\dfrac{F_x'}{F_y'}$$
2、多元隐函数F(x,y,z)=0, $$\dfrac{\partial z}{\partial x}=-\dfrac{F_x'}{F_z'},\dfrac{\partial z}{\partial x}=-\dfrac{F_y'}{F_z'}$$
F(x,y,u,v)=0;G(x,y,u,v)=0则 隐函 u=u(x,y),v=v(x,y)
自变量x,y 函数u,v 。以x为例求偏导:
$$\begin{cases} F_x'+F_u' \frac{\partial u}{\partial x}+F_v' \frac{\partial v}{\partial x}=0 \\ \\ G_x'+G_u' \frac{\partial u}{\partial x}+G_v' \frac{\partial v}{\partial x} =0 \end{cases}$$
notes:
1、行列式对调两行或两列,值相反。
2、克莱姆法则x=D1/D,y=D2/D.
3、行列式得一行或者一列可以提取公共因子。
$$D = \begin{vmatrix}F_u'&F_v'\\ \\ G_u'&G_v' \end{vmatrix}$$
$$D_1 = -\begin{vmatrix}F_x'&F_v'\\ \\ G_x'&G_v' \end{vmatrix}$$
$$D_2 = -\begin{vmatrix}F_u'&F_x'\\ \\ G_u'&G_x' \end{vmatrix}$$
$$\dfrac{\partial u}{\partial x}=\dfrac{D_1}{D},\ \dfrac{\partial v}{\partial x}=\dfrac{D_2}{D}$$
补充:雅可比矩阵
$$J = \dfrac{\partial (F,G)}{\partial (u,v)} = \begin{bmatrix} \frac{\partial F}{\partial u}&\frac{\partial F}{\partial v}\\ \\\frac{\partial G}{\partial u}&\frac{\partial G}{\partial v} \end{bmatrix} $$
$$X = \dfrac{\partial (u,v)}{\partial (x,y)} = \begin{pmatrix} \frac{\partial u}{\partial x}&\frac{\partial u}{\partial y}\\ \\\frac{\partial v}{\partial x}&\frac{\partial v}{\partial y} \end{pmatrix} $$
$$B = \dfrac{\partial (F,G)}{\partial (x,y)} = \begin{pmatrix} -\frac{\partial F}{\partial x}&-\frac{\partial F}{\partial y}\\ \\-\frac{\partial G}{\partial x}&-\frac{\partial G}{\partial y} \end{pmatrix} $$
由以上分析易得JX=B .
$$\begin{cases} xu+yv =1 \\ x^2 + y^2+u^2+v^2=6 \end{cases},find\ \dfrac{\partial u}{\partial x}$$
2个方程2个函数u、v。两个自变量x、y。
对两式的x求偏导:
$$\begin{cases} u+x\frac{\partial u}{\partial x} + y \frac{\partial v}{\partial x} \\ 2x+2u\frac{\partial u}{\partial x}+2v\frac{\partial v}{\partial x}\end{cases}$$
$$D = \begin{vmatrix} x&y \\ u &v \end{vmatrix}=xv-yu$$
$$D_1 = \begin{vmatrix} -u&y \\ -x &v \end{vmatrix}=xy-uv$$
$$D_2 = \begin{vmatrix} x&-u \\ u & -x \end{vmatrix}=u^2-x^2$$
所以$$\begin{cases} \dfrac{\partial u}{\partial x} = \dfrac{xy-uv}{xv-yu} \\ \\ \dfrac{\partial v}{\partial x} = \dfrac{u^2-x^2}{xv-yu} \end{cases}$$
$$\begin{cases} x-2y+3z = 2 \\ x^2 + 2y^2+3z^2 =21 \end{cases},find\ \dfrac{dz}{dx}$$
两个方程2个函数y,z,一个自变量x.
对x求偏导可得:
$$\begin{cases} 1-2\dfrac{dy}{dx}+3\dfrac{dz}{dx}=0\\ \\ 2x+4y\dfrac{dy}{dx}+3z\dfrac{dz}{dx} =0\end{cases}$$
解二元一次方程组,可得 $$\dfrac{dz}{dx}$$
解法和“零、一元隐函数求导”没有什么大的区别
如果直接对x偏导就比较麻烦:
先将原式化为 : $$e^{xlnz} =y^z$$
再对x偏导
$$e^{xlnz} \cdot (lnz +x\cdot \dfrac{1}{z}\cdot z') =y^z\cdot lny \cdot z'$$ , 其中$$e^{xlnz} = z^x$$
解得 $$z'=\cdots\cdots$$同上。。。
第七节多元隐函数求偏导.ppt
见上面ppt附件