多元函数微分学（公式，不含推导）

多元函数微分学

多元函数

n维欧氏空间

R n = {   ( x 1 , x 2 , ⋯   , x n ) ∣ x i ∈ R , i = 1 , 2 , ⋯   , n   } 且记 d = ∣ A B ∣ = ( x 1 − y 1 ) 2 + ( x 2 − y 2 ) 2 + ⋯ + ( x n − y n ) 2 则称 R 为 n 维欧氏空间 \mathbb R^n = \set{(x_1,x_2,\cdots,x_n)|x_i\in\mathbb R,i=1,2,\cdots,n}\\ 且记d=|AB|=\sqrt{(x_1-y_1)^2+(x_2-y_2)^2+\cdots+(x_n-y_n)^2 }\\ 则称\mathbb R为n维欧氏空间 Rn={(x1​,x2​,⋯,xn​)∣xi​∈R,i=1,2,⋯,n}且记d=∣AB∣=(x1​−y1​)2+(x2​−y2​)2+⋯+(xn​−yn​)2 ​则称R为n维欧氏空间

$\delta$邻域

圆形

U ( P 0 , δ ) = {   ( x , y ) ∈ R 2 ∣ ( x − x 0 ) 2 + ( y − y 0 ) 2 < δ   } 去心领域： U ˚ ( P 0 , δ ) = {   ( x , y ) ∈ R 2 ∣ 0 < ( x − x 0 ) 2 + ( y − y 0 ) 2 < δ   } U(P_0,\delta)=\set{(x,y)\in\mathbb R^2|\sqrt{(x-x_0)^2+(y-y_0)^2}<\delta}\\ 去心领域：\mathring{U}(P_0,\delta)=\set{(x,y)\in\mathbb R^2|0<\sqrt{(x-x_0)^2+(y-y_0)^2}<\delta} U(P0​,δ)={(x,y)∈R2∣(x−x0​)2+(y−y0​)2 ​<δ}去心领域：U˚(P0​,δ)={(x,y)∈R2∣0<(x−x0​)2+(y−y0​)2 ​<δ}

方形

U ( P 0 , δ ) = {   ( x , y ) ∈ R 2 ∣ ∣ x − x 0 ∣ < δ , ∣ y − y 0 ∣ < δ   } U(P_0,\delta)=\set{(x,y)\in\mathbb R^2||x-x_0|<\delta,|y-y_0|<\delta}\\ U(P0​,δ)={(x,y)∈R2∣∣x−x0​∣<δ,∣y−y0​∣<δ}

平面点集

$$\{\begin{array}{l}内点:\exists \delta >0,U(P_0,\delta )\subset D\\ 外点:\forall \delta >0,U(P_0,\delta )\cap D=\varnothing \\ 界点:\forall \delta >0,U(P_0,\delta )中既有属于D的点，也有不属于D的点\end{array}$$

点集D的全部内点组成的集合称为D的内部，记作$intD$
$$设D\in {\mathbb{R}}^2\{\begin{array}{l}ifD=intD,D为{\mathbb{R}}^2的开集\\ ifD的补集D^c为开集，D为{\mathbb{R}}^2的闭集\\ if\exists M>0,D\subset U(O,M),O为坐标原点，D为有界集，否则为无界集\end{array}$$

多元函数

定义域 值域

二元函数的极限与连续

$$\begin{gathered} 设z=f(x,y)P_0(x_0,y_0)\in D, \\ \exists A\in \mathbb{C},\forall \epsilon >0,\exists \delta >0,when0<\sqrt{(x-x_0{)}^2+(y-y_0{)}^2}<\delta ,and(x,y)\in D \\ |f(x,y)-A|<\epsilon \\ 则记\underset{(x,y)\to (x_0,y_0)}{\lim }f(x,y)=A \end{gathered}$$

累次积分

$$\underset{(x,y)\to (x_0,y_0)}{\lim }f(x,y)$$

重极限

$$\underset{x\to x_0}{\lim }\underset{y\to y_0}{\lim }f(x,y)$$

若函数在$(x_0,y_0)$处存在重极限，则重极限与累次积分相等

二元函数连续性

$$\underset{(x，y)\to (x_0,y_0)}{\lim }f(x,y)=f(x_0,y_0),则z=f(x,y)在(x_0,y_0)处连续$$

全增量

$$\Delta z=f(x_0+\Delta x,y_0+\Delta y)-f(x_0,y_0)$$

偏增量

$$\begin{gathered} \Delta z=f(x_0+\Delta x,y_0)-f(x_0,y_0) \\ \Delta z=f(x_0,y_0+\Delta y)-f(x_0,y_0) \end{gathered}$$

有界闭区域上连续函数的性质

1. 极值可取到
2. 零点存在

偏导数和全微分

偏导数

$$\underset{\Delta x\to 0}{\lim }\frac{f(x_0+\Delta x,y_0)-f(x_0,y_0)}{\Delta x}=f_x^{\prime }(x_0,y_0)=\frac{\partial f}{\partial x}{|}_{(x_0,y_0)}$$

全微分

$$\Delta z=f(x_0+\Delta x,y_0+\Delta y)-f(x_0,y_0)=A\Delta x+B\Delta y+o(\rho )$$

性质

1. 全微分存在即可微
2. 若$z=f(x,y)$在P处可微，则可偏导
3. 可微则$dz=\frac{\partial z}{\partial x}\Delta x+\frac{\partial z}{\partial y}\Delta y$
4. 若可微，则连续
5. 若偏导在P点连续，则可微

高阶偏导数

$$\begin{gathered} \frac{{\partial }^{2}z}{\partial x^2}=\frac{\partial }{\partial x}\left(\frac{\partial z}{\partial x}\right)=f_{xx}^{\prime \prime }(x,y) \\ \frac{{\partial }^{2}z}{\partial x\partial y}=\frac{\partial }{\partial y}\left(\frac{\partial z}{\partial x}\right)=f_{xy}^{\prime \prime }(x,y) \\ \frac{{\partial }^{2}z}{\partial y\partial x}=\frac{\partial }{\partial x}\left(\frac{\partial z}{\partial y}\right)=f_{yx}^{\prime \prime }(x,y) \\ \frac{{\partial }^{2}z}{\partial y^2}=\frac{\partial }{\partial y}\left(\frac{\partial z}{\partial y}\right)=f_{yy}^{\prime \prime }(x,y) \end{gathered}$$

若函数在P处有二阶连续偏导数，则$f_{xy}^{\prime \prime }(x,y)=f_{yx}^{\prime \prime }(x,y)$

多元复合函数的偏导数

链式法则

$$\begin{gathered} 设u=g(x,y),v=h(x,y),z=f(u,v) \\ u,v在P(x,y)处偏导数存在，且z在对应点处有一阶连续偏导数 \\ \frac{\partial z}{\partial x}=\frac{\partial z}{\partial u}\cdot \frac{\partial u}{\partial x}+\frac{\partial z}{\partial v}\cdot \frac{\partial v}{\partial x} \\ \frac{\partial z}{\partial y}=\frac{\partial z}{\partial u}\cdot \frac{\partial u}{\partial y}+\frac{\partial z}{\partial v}\cdot \frac{\partial v}{\partial y} \end{gathered}$$

一阶全微分形式不变性

$$dz=(\frac{\partial z}{\partial u}\cdot \frac{\partial u}{\partial x}+\frac{\partial z}{\partial v}\cdot \frac{\partial v}{\partial x})dx+(\frac{\partial z}{\partial u}\cdot \frac{\partial u}{\partial y}+\frac{\partial z}{\partial v}\cdot \frac{\partial v}{\partial y})dy=\frac{\partial z}{\partial u}du+\frac{\partial z}{\partial v}dv$$

隐函数的偏导数

若函数$F(x,y)$满足：

1. 函数$F$在以$P_0(x_0,y_0)$为内点的某一区域$D\subset {\mathbb{R}}^2$上连续
2. $F(x_0,y_0)=0$
3. 在$D$内存在连续的偏导数$F_x^{\prime }(x,y)$，$F_y^{\prime }(x,y)$
4. $F_y^{\prime }(x,y)\ne 0$

则，
$$\frac{dy}{dx}=-\frac{F_x^{\prime }(x,y)}{F_y^{\prime }(x,y)}$$
若函数$F(x,y,z)$满足：

1. 函数$F$在以$P_0(x_0,y_0,z_0)$为内点的某一区域$V\subset {\mathbb{R}}^3$上连续
2. $F(x_0,y_0,z_0)=0$
3. 在$D$内存在连续的偏导数$F_x^{\prime }(x,y,z)$，$F_y^{\prime }(x,y,z)$,$F_z^{\prime }(x,y,z)$
4. $F_z^{\prime }(x,y,z)\ne 0$

则
$$\begin{gathered} \frac{dz}{dx}=-\frac{F_x^{\prime }(x,y,z)}{F_z^{\prime }(x,y,z)} \\ \frac{dz}{dy}=-\frac{F_y^{\prime }(x,y,z)}{F_z^{\prime }(x,y,z)} \end{gathered}$$

多元函数的极值

二元函数的极值

条件：
$$\{\begin{array}{l}{f}_x^{\prime }(x_0,y_0)=0\\ f_y^{\prime }(x_0,y_0)=0\end{array}$$

$$\begin{gathered} 此时，计A=f_{xx}^{\prime \prime }(x_0,y_0),B=f_{xy}^{\prime \prime }(x_0,y_0),C=f_{yy}^{\prime \prime }(x_0,y_0) \\ \{\begin{array}{l}{B}^2-AC<0\{\begin{array}{l}A>0,极小值\\ A<0,极大值\end{array}\\ B^2-AC=0，不是极值点\\ B^2-AC>0,不一定\end{array} \end{gathered}$$

多元函数的条件极值

指在$g(x,y)=0$下$f(x_0,y_0)$为极值
$$\{\begin{array}{l}{f}_x^{\prime }(x_0,y_0)+\lambda g_x^{\prime }(x_0,y_0)=0\\ f_y^{\prime }(x_0,y_0)+\lambda g_y^{\prime }(x_0,y_0)=0\\ g_x(x_0,y_0)=0\end{array}$$
引进
$$L(x,y,\lambda )=f(x,y)+\lambda g(x,y)$$
则对所有参数偏导均为0

泰勒展开

$${\phi }^{(k)}(0)=(\Delta x\frac{\partial }{\partial x}+\Delta y\frac{\partial }{\partial y}{)}^{k}f(x_0,y_0)$$

所以，佩亚诺余项泰勒展开
$$f(x,y)=f(x_0,y_0)+\sum _{k=1}^n\frac{1}{k!}(\Delta x\frac{\partial }{\partial x}+\Delta y\frac{\partial }{\partial y}{)}^{k}f(x_0,y_0)+o({\rho }^n)$$

方向导数和梯度

方向导数

$$\begin{gathered} \underset{\rho \to 0^{+}}{\lim }\frac{f(M+\rho \vec{l})-f(M)}{\rho }=\frac{\partial f}{\partial \vec{l}}{|}_M \\ \frac{\partial f}{\partial \vec{l}}=f_x^{\prime }(M_0)\cos \alpha +f_y^{\prime }(M_0)\cos \beta +f_z^{\prime }(M_0)\cos \gamma \end{gathered}$$

梯度

$$\begin{gathered} gradf=(f_x^{\prime }(M),f_y^{\prime }(M),f_z^{\prime }(M)) \\ \frac{\partial f}{\partial \vec{l}}=|gradf|\cos \theta =gradf\cdot {\vec{l}}^0 \end{gathered}$$

切平面与法线

$$\begin{gathered} F(x,y,z)=0 \\ C:\{\begin{array}{l}x=x(t)\\ y=y(t)\\ z=z(t)\end{array} \end{gathered}$$

则切向量$\vec{s}=(x^{\prime }(t_0),y^{\prime }(t_0),z^{\prime }(t_0)$,法向量$\vec{n}=(F_x^{\prime }(M_0),F_y^{\prime }(M_0),F_z^{\prime }(M_0))$

法线方程
$$\frac{x-x_0}{F_x^{\prime }(M_0)}=\frac{y-y_0}{F_y^{\prime }(M_0)}=\frac{z-z_0}{F_z^{\prime }(M_0)}$$