CAL-6-多元函数微分学
2021-07-30 11:55:17 1 举报
AI智能生成
高等数学微积分 第六章 多元函数微分学 知识点梳理
作者其他创作
大纲/内容
多元理论实践(3)
求偏导(3)
多元显函数求偏导
把其他变量看做常数<br>
二阶偏导数仍然连续<span class="equation-text" data-index="0" data-equation="\to\color{red}\frac{\partial^2\zeta}{\partial x\partial y}=\frac{\partial^2\zeta}{\partial y\partial x}" contenteditable="false"><span></span><span></span></span>
对于二阶连续可偏导的f:<span class="equation-text" data-index="0" data-equation="f_{12}=f_{21}" contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="\mathrm d\zeta(x,y)=\frac{\partial\zeta}{\partial x}\mathrm dx+\frac{\partial\zeta}{\partial y}\mathrm dy" contenteditable="false"><span></span><span></span></span>
多元复合函数求偏导(3)<br>
先辨析:<span class="equation-text" data-index="0" data-equation="\left\{\begin{aligned}\zeta=f(x^2+y^2)\rightarrow\zeta二元,f一元\\\zeta=f(t,e^t)\rightarrow\zeta一元,f二元\\\zeta=f(xy,x+y)\to\zeta二元,f二元\end{aligned}\right." contenteditable="false"><span></span><span></span></span><br>
<span class="equation-text" data-index="0" data-equation="\frac{\mathrm d\zeta(u(t),v(t))}{\mathrm dt}=\frac{\mathrm d\zeta(t)}{\mathrm dt}" contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="\frac{\mathrm d\zeta(u,v)}{\mathrm dt}=\frac{\partial\zeta}{\partial u}\frac{\mathrm du}{\mathrm dt}+\frac{\partial\zeta}{\partial v}\frac{\mathrm dv}{\mathrm dt}" contenteditable="false"><span></span></span><br>
注意回代u(t),v(t)
<span class="equation-text" data-index="0" data-equation="\frac{\mathrm d^2\zeta(u,v)}{\mathrm dt^2}=\frac{\mathrm d}{\mathrm dt}\frac{\mathrm d\zeta}{\mathrm dt}=\frac{\mathrm d}{\mathrm dt}(f'_1\frac{\mathrm du}{\mathrm dt}+f'_2\frac{\mathrm dv}{\mathrm dt})" contenteditable="false"><span></span><span></span></span><br>
一般简记:<span class="equation-text" data-index="0" data-equation="\frac{\partial\zeta}{\partial u}\to f'_1,\frac{\partial\zeta}{\partial v}\to f'_2,\frac{\partial^2\zeta}{\partial u^2}\to f''_{11},\frac{\partial^2\zeta}{\partial u\partial v}\to f''_{12}" contenteditable="false"><span></span><span></span></span><br>
注意<span class="equation-text" data-index="0" data-equation="f'_1、f'_2" contenteditable="false"><span></span><span></span></span>亦是以u,v为自变量的复合函数
<span class="equation-text" data-index="0" data-equation="\frac{\partial^2\zeta(u(x,y),v(x,y))}{\partial x\partial y}=\frac{\partial }{\partial y}(f'_1\frac{\partial u}{\partial x}+f'_2\frac{\partial v}{\partial x})\\" contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="{\color{blue}例:}设z=f(xy,x+y),f二阶连续可偏导,求\frac{\partial^2z}{\partial x\partial y}" contenteditable="false"><span></span><span></span></span>
技巧
<span class="equation-text" data-index="0" data-equation="f(x)^{g(x)}" contenteditable="false"><span></span><span></span></span>型要先取对数
多元隐函数求偏导(3)
确定已知函数<span class="equation-text" data-index="0" data-equation="\left\{\begin{aligned}F(x,y)=0&\mathop{\Rightarrow}\limits_{显式化} 一元函数y=f(x)\\F(x,y,z)=0&\mathop{\Rightarrow}\limits_{显式化} 二元函数z=f(x,y)\\ \left\{\begin{aligned}F(x,y,z)=0\\G(x,y,z)=0\end{aligned}\right.&\mathop{\Rightarrow}\limits_{显式化} {\color{red}2个一元函数\left\{\begin{aligned}y=y(x)\\z=z(x)\end{aligned}\right.}\end{aligned}\right." contenteditable="false"><span></span><span></span></span>
对已知函数(等号两侧)直接求导,整理出目标对象
若已知函数是显函数,无需隐函数化
其余自变量均视为函数,应用求导乘法公式
<span class="equation-text" data-index="0" data-equation="\left\{\begin{aligned}F(x,y,z)=0\\G(x,y,z)=0\end{aligned}\right." contenteditable="false"><span></span><span></span></span>解法
视目标导数的自变量为全局自变量
将另外的变量视为自变量的函数
直接求导,作为二元方程组求解目标偏导<br>
<span class="equation-text" data-index="0" data-equation="{\color{blue}例:}设\left\{\begin{aligned}&x^2+2y^2+3z^2&=&21\\&x-3y+2z&=&5\end{aligned}\right.求\frac{\mathrm dy}{\mathrm dx}" contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="{\color{blue}例:}设\left\{\begin{aligned}&y=f(x,t)\\&F(x^2,y,\sin t)=0\end{aligned}\right.求\frac{\mathrm dy}{\mathrm dx}" contenteditable="false"><span></span><span></span></span>
Jacobi行列式<br>
<font color="#F44336">多元微分学代数应用<br>极值<br></font>
参照:一元<span class="equation-text" data-index="0" data-equation="y=f(x)" contenteditable="false"><span></span><span></span></span>
基本步骤:<span class="equation-text" data-index="0" data-equation="\left\{\begin{aligned}&确定定义域\\&探究f'(x)=0/不存在的情况\\&应用判别法\end{aligned}\right." contenteditable="false"><span></span><span></span></span><br>
多元函数极值(2)
无条件极值<br>Case1. z=f(x,y)<br>Case2. F(x,y,z)=0<br>
Case1. <span class="equation-text" data-index="0" data-equation="z=f(x,y),(x,y\in D)" contenteditable="false"><span></span><span></span></span>,D是开区域(不含边界)<br>
令<span class="equation-text" data-index="0" data-equation="\left\{\begin{aligned}\frac{\partial z}{\partial x}=0\\\frac{\partial z}{\partial y}=0\end{aligned}\right.\Rightarrow\left\{\begin{aligned}x=?\\y=?\end{aligned}\right. " contenteditable="false"><span></span><span></span></span>(驻点)<br>
应用判别法
条件极值<br>(有限制条件)<br>
设目标函数<span class="equation-text" data-index="0" data-equation="z=f(x,y)" contenteditable="false"><span></span><span></span></span>,且满足<span class="equation-text" data-index="1" data-equation="\varphi(x,y)=0" contenteditable="false"><span></span><span></span></span>
令<span class="equation-text" data-index="0" data-equation="F(x,y,\lambda)=f(x,y)+\lambda\varphi(x,y)" contenteditable="false"><span></span><span></span></span><br>
令<span class="equation-text" data-index="0" data-equation="\left\{\begin{aligned}&F_x=f_x+\lambda\varphi_x=0(a)\\&F_y=f_y+\lambda\varphi_y=0(b)\\&F_{\lambda}=\varphi(x,y)=0(c)\end{aligned}\right.\Rightarrow\left\{\begin{aligned}x=?\\y=?\end{aligned}\right." contenteditable="false"><span></span><span></span></span>
解法一:由<span class="equation-text" data-index="0" data-equation="\frac{(a)}{(b)}" contenteditable="false"><span></span><span></span></span>消去<span class="equation-text" data-index="1" data-equation="\lambda\Rightarrow y=y(x)" contenteditable="false"><span></span><span></span></span>,代入(c)解得(x,y)<br>
解法二:由(a)(b)得到<span class="equation-text" data-index="0" data-equation="\left\{\begin{aligned}x=h(\lambda)\\y=t(\lambda)\end{aligned}\right. " contenteditable="false"><span></span><span></span></span>,代入(c)解得<span class="equation-text" data-index="1" data-equation="\lambda" contenteditable="false"><span></span><span></span></span>,从而解出(x,y)<br>
技巧:<span class="equation-text" data-index="0" data-equation="F_x" contenteditable="false"><span></span><span></span></span>恰好与y无关时,可以单独分析直到解出对应(x,y)。<span class="equation-text" data-index="1" data-equation="F_y" contenteditable="false"><span></span><span></span></span>同理
应用判别法
<font color="#F44336">判别法</font>
令<span class="equation-text" data-index="0" data-equation="\left\{\begin{aligned}&A=\frac{\partial^2 z}{\partial x^2}\big|_{(x_0,y_0)}\\&B=\frac{\partial^2 z}{\partial x\partial y}\big|_{(x_0,y_0)}\\&C=\frac{\partial^2 z}{\partial y^2}\big|_{(x_0,y_0)}\end{aligned}\right." contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="AC-B^2\left\{\begin{aligned}&<0\Rightarrow不是极值点\\&>0\Rightarrow是极值点\left\{\begin{aligned}&A<0\Rightarrow极大点\\&A>0\Rightarrow极小点\end{aligned}\right.\end{aligned}\right." contenteditable="false"><span></span><span></span></span>
混合情况
<span class="equation-text" data-index="0" data-equation="{\color{green}例}:求z=2x^2-y^2+3在D:x^2+4y^2\leq4上的{\color{red}最大、最小值}" contenteditable="false"><span></span><span></span></span>
取整与否对应了两种不同的极值:取整:条件;不取整:无条件;<br>
最大最小直接代值即可
多元微分学物理<br>与几何应用<br>
方向导数与梯度(3)
方向余弦
当自变量正交时,方向余弦的平方和=1<br>
方向导数<br>
<span class="equation-text" data-index="0" data-equation="\frac{\partial f}{\partial l}\big|_{M_0}=\frac{\partial f}{\partial x}\big|_{M_0}\cos\alpha+\frac{\partial f}{\partial y}\big|_{M_0}\cos\beta" contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="\frac{\partial f}{\partial l}\big|_{M_0}=\frac{\partial f}{\partial x}\big|_{M_0}\cos\alpha+\frac{\partial f}{\partial y}\big|_{M_0}\cos\beta+\frac{\partial f}{\partial z}\big|_{M_0}\cos\gamma" contenteditable="false"><span></span><span></span></span>
梯度
<span class="equation-text" data-index="0" data-equation="通常\bold{grad}u=\{\frac{\partial f}{\partial x},\frac{\partial f}{\partial y},\frac{\partial f}{\partial z}\}" contenteditable="false"><span></span><span></span></span>
几何应用(2)
空间曲面
<span class="equation-text" data-index="0" data-equation="{F'_x(M_0)}(x-x_0)+{F'_y(M_0)}(y-y_0)+{F'_z(M_0)}(z-z_0)=0" contenteditable="false"><span></span><span></span></span>:切平面
<span class="equation-text" data-index="0" data-equation="\boldsymbol{n}=\{F'_x(M_0),F'_y(M_0),F'_z(M_0)\}" contenteditable="false"><span></span><span></span></span>:法向量
<span class="equation-text" data-index="0" data-equation="\frac{x-x_0}{F'_x(M_0)}=\frac{y-y_0}{F'_y(M_0)}=\frac{z-z_0}{F'_z(M_0)}" contenteditable="false"><span></span><span></span></span>:法线
空间曲线(2)<br>
<span class="equation-text" data-index="0" data-equation="L:\left\{\begin{aligned}x=\varphi(t)\\y=\psi(t)\\z=\omega(t)\end{aligned}\right." contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="\varphi'(t_0)(x-x_0)+\psi'(t_0)(y-y_0)+\omega'(t_0)(z-z_0)=0" contenteditable="false"><span></span><span></span></span>:法平面
<span class="equation-text" data-index="0" data-equation="\boldsymbol{T}=\{\varphi'(t_0),\psi'(t_0),\omega'(t_0)\}" contenteditable="false"><span></span><span></span></span>:切向量
<span class="equation-text" data-index="0" data-equation="\frac{x-x_0}{\varphi'(t_0)}=\frac{y-y_0}{\psi'(t_0)}=\frac{z-z_0}{\omega'(t_0)}" contenteditable="false"><span></span><span></span></span>:切线<br>
<span class="equation-text" data-index="0" data-equation="L:\left\{\begin{aligned}F(x,y,z)=0\\G(x,y,z)=0\end{aligned}\right." contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="\boldsymbol{T}=\{F'_x,F'_y,F'_z\}\times\{G'_x,G'_y,G'_z\}\big|_{M_0}" contenteditable="false"><span></span><span></span></span>:切向量
基本概念
极限<br>
一元:<span class="equation-text" data-index="0" data-equation="\forall\varepsilon>0,\exist\delta>0,当x\in\mathring{U}(a,\delta),|f(x)-a|<\varepsilon" contenteditable="false"><span></span><span></span></span><br>
二元:<span class="equation-text" data-index="0" data-equation="\forall\varepsilon>0,\exist\delta>0,当0<\sqrt{(x-x_0)^2+(y-y_0)^2}<\delta时,|f(x,y)-a|<\varepsilon" contenteditable="false"><span></span><span></span></span>
连续<br>
一元:<span class="equation-text" data-index="0" data-equation="\lim\limits_{x\rightarrow x_0}f(x)=f(x_0) " contenteditable="false"><span></span><span></span></span><br>
等价于:左右极限均有,=函数值<br>
二元:<span class="equation-text" data-index="0" data-equation="\lim\limits_{x\rightarrow x_0 ,y\rightarrow y_0}f(x,y)=f(x_0,y_0)" contenteditable="false"><span></span><span></span></span>
偏导数<br><span class="equation-text" data-index="0" data-equation="\zeta=f(x,y)\\在(x_0,y_0)邻域\\有定义" contenteditable="false"><span></span><span></span></span><br>
偏增量<br>
<span class="equation-text" data-index="0" data-equation="\Delta\zeta _x" contenteditable="false"><span></span><span></span></span>:<span class="equation-text" data-index="1" data-equation="f(x_0+\Delta x,y_0)-f(x_0,y_0)" contenteditable="false"><span></span><span></span></span><br>
<span class="equation-text" data-index="0" data-equation="\Delta\zeta _y" contenteditable="false"><span></span><span></span></span>:<span class="equation-text" data-index="1" data-equation="f(x_0,y_0+\Delta y)-f(x_0,y_0)" contenteditable="false"><span></span><span></span></span><br>
全增量<br>
<span class="equation-text" data-index="0" data-equation="\Delta\zeta" contenteditable="false"><span></span><span></span></span>:<span class="equation-text" data-index="1" data-equation="f(x_0+\Delta x,y_0+\Delta y)-f(x_0,y_0)" contenteditable="false"><span></span><span></span></span><br>
偏导数
<span class="equation-text" data-index="0" data-equation="if\lim\limits_{\Delta x\rightarrow0}\frac{\Delta\zeta}{\Delta x}=\lim\limits_{x\rightarrow x_0}\frac{f(x,y_0)-f(x_0,y_0)}{x-x_0}\exist" contenteditable="false"><span></span><span></span></span>,则称f(x,y)在<span class="equation-text" data-index="1" data-equation="(x_0,y_0)" contenteditable="false"><span></span><span></span></span>处对x可偏导。<br>极限值A称该点处函数对x的偏导数,即为<span class="equation-text" data-index="2" data-equation="\frac{\partial\zeta}{\partial x}\big|_{x=x_0}" contenteditable="false"><span></span><span></span></span><br>
可微性
一元:<span class="equation-text" data-index="0" data-equation="y=f(x)(x\in D),x_0\in D\\\Delta y=f(x_0+\Delta x)-f(x_0)" contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="if \Delta y = A\Delta x+o(\Delta x)" contenteditable="false"><span></span><span></span></span>称函数可微<br>并记<span class="equation-text" data-index="1" data-equation="A\mathrm dx=\mathrm dy\big|_{x=x_0}" contenteditable="false"><span></span><span></span></span>为函数的微分<br>
二元:<span class="equation-text" data-index="0" data-equation="\Delta\zeta=f(x_0+\Delta x,y_0+\Delta y)-f(x_0,y_0)" contenteditable="false"><span></span><span></span></span><br>
<span class="equation-text" data-index="0" data-equation="if \Delta\zeta = A\Delta x+B\Delta y+o({\color{red}\sqrt{\Delta x^2+\Delta y^2}})" contenteditable="false"><span></span><span></span></span><br>称函数在<span class="equation-text" data-index="1" data-equation="M_0" contenteditable="false"><span></span><span></span></span>可微<br>并记<span class="equation-text" data-index="2" data-equation="A\mathrm dx+B\mathrm dy=\mathrm d\zeta\big|_{\tiny\begin{aligned} x=x_0\\y=y_0\end{aligned}}" contenteditable="false"><span></span><span></span></span>为函数的<font color="#F44336">全微分</font><br>
有界闭区域<span class="equation-text" data-index="0" data-equation="D" contenteditable="false"><span></span><span></span></span>上<span class="equation-text" data-index="1" data-equation="f(x,y)" contenteditable="false"><span></span><span></span></span>连续的性质(3)
最值定理:若f(x,y)在有界闭区域D连续→<span class="equation-text" data-index="0" data-equation="\exist " contenteditable="false"><span></span><span></span></span><span class="equation-text" data-index="1" data-equation="m,M" contenteditable="false"><span></span><span></span></span><span class="equation-text" data-index="2" data-equation="," contenteditable="false"><span></span><span></span></span><span class="equation-text" data-index="3" data-equation="m\leq f(x,y)\leq M" contenteditable="false"><span></span><span></span></span><br>
有界定理:若f(x,y)在有界闭区域D连续→<span class="equation-text" data-index="0" data-equation="\exist k>0" contenteditable="false"><span></span><span></span></span>使<span class="equation-text" data-index="1" data-equation="|f(x,y)|\leq k" contenteditable="false"><span></span><span></span></span>
介值定理:设f(x,y)在有界闭区域D连续,<span class="equation-text" data-index="0" data-equation="\forall\delta\in[m,M],\exist(\xi,\eta),使f(\xi,\eta)=\delta" contenteditable="false"><span></span><span></span></span>
连续、可偏导、可微之关系
可微→连续<br>
可微→可偏导
连续<span class="equation-text" data-index="0" data-equation="\nRightarrow" contenteditable="false"><span></span><span></span></span>可偏导:<br>
<span class="equation-text" data-index="0" data-equation="f(x,y)=\sqrt{x^2+y^2}在(0,0)" contenteditable="false"><span></span><span></span></span>不可偏导。<br>
可偏导<span class="equation-text" data-index="0" data-equation="\nRightarrow" contenteditable="false"><span></span><span></span></span>连续:
<span class="equation-text" data-index="0" data-equation="\color{red} \zeta=f(x,y)=\left\{\begin{aligned}\frac{xy}{x^2+y^2}&,&(x,y)\neq(0,0)\\0&,&(x,y)=(0,0)\end{aligned}\right." contenteditable="false"><span></span><span></span></span><br>
(0,0)不连续,可偏导
方向导数与梯度
方向导数<br>
<span class="equation-text" data-index="0" data-equation="就是函数f在定义域内某个方向上的导数" contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="n维空间的方向导数,指的是定义域的维数" contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="设函数f在点M_0的邻域内有定义" contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="在定义域内,过点M_0做射线l,\\并在l上取一点M\neq M_0" contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="射线l不一定与函数f相切" contenteditable="false"><span></span><span></span></span>
<span class="equation-text" data-index="0" data-equation="若\lim\limits_{|MM_0|\to0}\frac{f(M)-f(M_0)}{|MM_0|}存在\\称此极限为函数f在点M_0处\\沿射线l的方向导数,记为\frac{\partial f}{\partial l}\big|_{M_0}" contenteditable="false"><span></span><span></span></span>
方向余弦
<span class="equation-text" data-index="0" data-equation="将\frac{\partial f}{\partial l}\big|_{M_0}表示为\sum\limits_{i=x,y...}\frac{\partial f}{\partial i}\big|_{M_0}\cos A_i,\\则cos A_x,cos A_y...即为射线l的方向余弦" contenteditable="false"><span></span><span></span></span>
当自变量正交时,方向余弦的平方和=1<br>
方向余弦即为方向角的余弦
通常利用方向余弦求方向导数
梯度
<span class="equation-text" data-index="0" data-equation="记\frac{\partial f}{\partial l}\big|_{M_0}=\bold{grad}u\big|_{M_0}\cdot\hat{l},其中两者分别为\\f在点M_0处的梯度和射线l的单位向量" contenteditable="false"><span></span><span></span></span>
梯度即为点<span class="equation-text" data-index="0" data-equation="M_0" contenteditable="false"><span></span><span></span></span>处,取最大值的方向导数
大小
方向
<span class="equation-text" data-index="0" data-equation="通常\bold{grad}u=\{\frac{\partial f}{\partial x},\frac{\partial f}{\partial y},\frac{\partial f}{\partial z}\}" contenteditable="false"><span></span><span></span></span>
0 条评论
下一页