高等数学（七）- 多元函数微分学(1)【多元微分法：多元求偏导】

本节为高等数学复习笔记的第七部分，多元函数微分学（1），主要包括：多元微分法求多元偏导数 。

1. 多元微分法：多元求偏导

  $（1）f(x,y)，固定一个变量，对另一个变量求导，即$：$\frac{\partial f(x,y)}{\partial x}=f_x^{\prime }(x,y)$，$\frac{\partial f(x,y)}{\partial y}=f_y^{\prime }(x,y)$

$（2）链式求导$：

  $z=f(u,v,w)$，$u=u(x,y)$，$v=v(x)$，$w=w(y)$，$则\frac{\partial z}{\partial x}=\frac{\partial z}{\partial u}\cdot \frac{\partial u}{\partial x}+\frac{\partial z}{\partial v}\cdot \frac{\partial v}{\partial x}$

$（3）高阶偏导数$：

  $二阶偏导$：$\frac{\partial (\frac{\partial z}{\partial x})}{\partial x}=\frac{{\partial }^{2}z}{\partial x}$

  $二阶混合偏导$：$\frac{\partial (\frac{\partial z}{\partial x})}{\partial y}=\frac{{\partial }^{2}z}{\partial x\partial y}$

$eg1(显函数).$$设z=f(e^{x}siny,x^2+y^2)，其中f具有$$二阶连续偏导数$，$求\frac{{\partial }^{2}x}{\partial x\partial y}$。
  $解：$$\frac{\partial z}{\partial x}=e^{x}siny{f}_1^{\prime }+2x{f}_2^{\prime }$，

  $\frac{{\partial }^{2}x}{\partial x\partial y}=\frac{\partial (\frac{\partial z}{\partial x})}{\partial y}=\frac{\partial (f_1^{\prime }{e}^{x}siny)}{\partial y}+\frac{\partial (f_2^{\prime }\cdot 2x)}{\partial y}$$=\frac{\partial f_1^{\prime }}{\partial y}\cdot e^{x}siny+f_1^{\prime }{e}^{x}cosy+2x\frac{\partial f_2^{\prime }}{\partial y}$，

  $又\frac{\partial f_1^{\prime }}{\partial y}=f_{11}^{\prime \prime }\cdot e^{x}cosy+f_{12}^{\prime \prime }\cdot 2y$，
  $\frac{\partial f_2^{\prime }}{\partial y}=f_{21}^{\prime \prime }\cdot e^{x}cosy+f_{22}^{\prime \prime }\cdot 2y$，

  $\therefore \frac{{\partial }^{2}x}{\partial x\partial y}=f_{11}^{\prime \prime }{e}^{2x}sinycosy+2e^x(ysiny+xcosy)f_{12}^{\prime \prime }$$+4xy{f}_{22}^{\prime \prime }+f_1^{\prime }{e}^{x}cosy$。

---

$eg2(隐函数).$$设z=z(x,y)由方程F(x+\frac{z}{y},y+\frac{z}{x})确定$，$其中F有连续偏导数$，$求：x\cdot \frac{\partial z}{\partial x}+y\cdot \frac{\partial z}{\partial y}$.

  $解：$
  $方法一：将方程两边分别对x，y求偏导数得$：
$1）F_1^{\prime }(1+\frac{1}{y}\frac{\partial z}{\partial x})+F_2^{\prime }(-\frac{z}{x^2}+\frac{1}{x}\frac{\partial z}{\partial x})=0$$⟹$

  $\frac{x{F}_1^{\prime }+y{F}_2^{\prime }}{xy}\frac{\partial z}{\partial x}=\frac{z}{x^2}{F}_2^{\prime }-F_1^{\prime }$

$2）F_1^{\prime }(-\frac{z}{y^2}+\frac{1}{y}\frac{\partial z}{\partial y})+F_2^{\prime }(1+\frac{1}{x}\frac{\partial z}{\partial y})=0$$⟹$

  $\frac{x{F}_1^{\prime }+y{F}_2^{\prime }}{xy}\frac{\partial z}{\partial y}=\frac{z}{y^2}{F}_1^{\prime }-F_2^{\prime }$
$⟹$
  $\frac{\partial z}{\partial x}=\frac{\frac{yz}{x}{F}_2^{\prime }-xy{F}_1^{\prime }}{x{F}_1^{\prime }+y{F}_2^{\prime }}$，$\frac{\partial z}{\partial y}=\frac{\frac{xz}{y}{F}_1^{\prime }-xy{F}_2^{\prime }}{x{F}_1^{\prime }+y{F}_2^{\prime }}$
$⟹$
  $x\frac{\partial z}{\partial x}+y\frac{\partial z}{\partial x}=z-xy$

  $方法二：$
  $对方程全微分，F_1^{\prime }\cdot d(x+\frac{z}{y})+F_2^{\prime }\cdot d(y+\frac{z}{x})=0$

  $有F_1^{\prime }\cdot (dx+\frac{ydz-zdy}{y^2})+F_2^{\prime }\cdot (dy+\frac{xdz-zdx}{x^2})=0$

  $有:$
$(F_1^{\prime }\frac{1}{y}+F_2^{\prime }\frac{1}{x})dz=(-F_1^{\prime }+\frac{z}{x^2}{F}_2^{\prime })dx+(-F_2^{\prime }+\frac{z}{y^2}{F}_1^{\prime })dy$

  $两边同乘xy$：
$(x{F}_1^{\prime }+y{F}_2^{\prime })dz=(-xy{F}_1^{\prime }+\frac{yz}{x}{F}_2^{\prime })dx+(-xy{F}_2^{\prime }+\frac{xz}{y}{F}_1^{\prime })dy$
$⟹$
  $\frac{\partial z}{\partial x}=\frac{\frac{yz}{x}{F}_2^{\prime }-xy{F}_1^{\prime }}{x{F}_1^{\prime }+y{F}_2^{\prime }}$，$\frac{\partial z}{\partial y}=\frac{\frac{xz}{y}{F}_1^{\prime }-xy{F}_2^{\prime }}{x{F}_1^{\prime }+y{F}_2^{\prime }}$
$⟹$
  $x\frac{\partial z}{\partial x}+y\frac{\partial z}{\partial x}=z-xy$

---

欢迎扫描二维码关注微信公众号 深度学习与数学   [每天获取免费的大数据、AI等相关的学习资源、经典和最新的深度学习相关的论文研读，算法和其他互联网技能的学习，概率论、线性代数等高等数学知识的回顾]