多元函数微分学

邻域、内点、外点、边界、开区域、闭区域以及多元函数的概念与性质

邻域

定义:设 \(P_0(x_0, y_0)\) 是平面上一点,\(\delta > 0\),称集合 \[U(P_0, \delta) = \{P(x, y) | |P - P_0| < \delta\} = \{(x, y) | \sqrt{(x - x_0)^2 + (y - y_0)^2} < \delta\}\] 为点 \(P_0\)\(\delta\) 邻域,也记作 \(U(P_0)\)

去心邻域\(\stackrel{\circ}{U}(P_0, \delta) = \{P | 0 < |P - P_0| < \delta\}\)

内点、外点、边界点

\(E\) 是平面上的一个点集,\(P_0\) 是平面上的一点:

内点:如果存在 \(P_0\) 的某个邻域 \(U(P_0)\) 使得 \(U(P_0) \subset E\),则称 \(P_0\)\(E\)内点

外点:如果存在 \(P_0\) 的某个邻域 \(U(P_0)\) 使得 \(U(P_0) \cap E = \emptyset\),则称 \(P_0\)\(E\)外点

边界点:如果 \(P_0\) 的任意邻域内既有属于 \(E\) 的点,又有不属于 \(E\) 的点,则称 \(P_0\)\(E\)边界点

聚点(极限点):如果 \(P_0\) 的任意去心邻域内都有 \(E\) 中的点,则称 \(P_0\)\(E\)聚点

孤立点:如果 \(P_0 \in E\)\(P_0\) 不是 \(E\) 的聚点,则称 \(P_0\)\(E\)孤立点

开区域与闭区域

开集:如果集合 \(E\) 的每一点都是 \(E\) 的内点,则称 \(E\)开集

闭集:如果集合 \(E\) 的所有聚点都属于 \(E\),则称 \(E\)闭集

区域(开区域):连通的开集称为区域开区域

闭区域:开区域连同它的边界一起称为闭区域

有界集:如果存在正数 \(M\),使得集合 \(E\) 中的所有点到原点的距离都不超过 \(M\),则称 \(E\)有界集

无界集:不是有界集的集合称为无界集

方程

曲面方程

  1. 有解三元方程 \(\Sigma : F(x, y, z) = 0\) >  也有可能是点或直线

  2. 参数方程 \(\Sigma' : x = x(u, v)\)\(y = y(u, v)\)\(z = z(u, v)\)

  3. \(\Sigma'\)\(\Sigma\) 的一部分的充要条件为:\(F(x(u, v), y(u, v), z(u, v)) \equiv 0\)

空间曲线方程

  1. 有解方程组 [ l :

    \[\begin{cases} F(x, y, z) = 0 \\ G(x, y, z) = 0 \end{cases}\]

    ]

    称为一般式

  2. 参数方程:\(l' : x = x(t)\)\(y = y(t)\)\(z = z(t)\)

  3. \(l'\)\(l\) 的一部分的充要条件为:\(F[x(t), y(t), z(t)] \equiv 0\)\(G[x(t), y(t), z(t)] \equiv 0\)

  4. 以弧长为参数的曲线方程:\(\bar{l} : \bar{x}(s)\)\(\bar{y}(s)\)\(\bar{z}(s)\),恒有 \([\bar{x}'(s)]^2 + [\bar{y}'(s)]^2 + [\bar{z}'(s)]^2 \equiv 1\)

  5. 对一般式同解变形可以分别得到在各坐标面上的投影方程

平面方程

  1. 点法式

  2. 一般式

  3. 截距式

  4. 点到平面距离公式

直线方程

  1. 一般式

  2. 对称式

  3. 参数式

  4. 向量式

偏导数

  1. 可偏导未必连续

  2. 二阶混合偏导数在某点连续则在该点它们相等

全微分

  1. 如果全增量 \(\Delta z = f(x + \Delta x, y + \Delta y) - f(x, y)\) 可以表示为 \(\Delta z = A \Delta x + B \Delta y + o(\rho)(\rho \rightarrow 0)\),其中 \(\rho = \sqrt{\Delta^2 x + \Delta^2 y}\),则说可微分。

  2. 如果可微,则 \(\text{d} z = \frac{\partial z}{\partial x} \text{d} x + \frac{\partial z}{\partial y} \text{d} y\)

  3. 如果函数的偏导数在某点连续,则在某点可微分。

曲线切线

已知空间曲线的参数方程,如果导数连续且不同时为零,则可以用如下公式

切线方程的对称式: \[ \frac{x - x_0}{x'(t_0)} = \frac{y - y_0}{y'(t_0)} = \frac{z - z_0}{z'(t_0)} \]

切向量: \[ \mathbf{T} = \pm (x'(t_0), y'(t_0), z'(t_0)) \]

法平面方程: \[ x'(t_0)(x - x_0) + y'(t_0)(y - y_0) + z'(t_0)(z - z_0) = 0 \]

曲面的切平面

已知空间曲面的方程,若各偏导数连续且不同时为零,则可以用如下公式:

切平面方程为: \[ F_x(x_0, y_0, z_0)(x - x_0) + F_y(x_0, y_0, z_0)(y - y_0) + F_z(x_0, y_0, z_0)(z - z_0) = 0 \]

法线方程: \[ \frac{x - x_0}{F_x(x_0, y_0, z_0)} = \frac{y - y_0}{F_y(x_0, y_0, z_0)} = \frac{z - z_0}{F_z(x_0, y_0, z_0)} \]

多元函数无条件极值

  1. 必要条件:在点有偏导数,且满足所有偏导数同时为零,得到驻点

  2. 充分条件:在驻点某领域内如果有二阶偏导数,设 \(f_{xx}(x_0, y_0) = A\)\(f_{xy}(x_0, y_0) = B\)\(f_{yy}(x_0, y_0) = C\)\(\Delta = AC - B^2\),则:

    1. \(\Delta > 0\) 时是极值,且 \(A < 0\) 是极大值。
    2. \(\Delta < 0\) 时不是极值。

多元函数条件极值

拉格朗日乘数法

习题汇编

  1. \(z = 2x^2 + y^2\) 截出平面 \(4x + 2y + z = 1\) 的一个椭圆,求这个椭圆的面积 \(A\)

    先求出截面面积与其在 \(xOy\) 平面的投影的比例系数 \[ \cos \alpha = \frac{1}{\sqrt{4^2 + 2^2 + 1^2}} = \sqrt{\frac{1}{21}} \]

    交线方程为: \[ \begin{cases} z = 2x^2 + y^2 \\ 4x + 2y + z = 1 \end{cases} \]

    消去变量 \(z\) 得到交线在 \(xOy\) 平面的投影 \[ \frac{(x + 1)^2}{2} + \frac{(y + 1)^2}{4} = 1 \]

    面积为 \(2\sqrt{2} \pi\),从而截面面积为 \(2\sqrt{2} \pi \times \sqrt{21} = 2\pi \sqrt{42}\)

  2. \(a > 0\)\(b > 0\)\(c > 0\),证明:单叶双曲面 \(\Sigma : \frac{x^2}{a^2} + \frac{y^2}{b^2} - \frac{z^2}{c^2} = 1\) 上任意一点都有经过该点并且完全位于 \(\Sigma\) 上的直线通过。

  3. 求空间曲线 \(L: x(t) = t(\text{e}^t - 1)\)\(y(t) = t \sin t\)\(z(t) = t^3 + t^2 + 1\),在点 \(P_0(0, 0, 1)\) 的切线。

    参数 \(t_0 = 0\),解得参数方程一阶导数同时为 0,此时只需要计算二阶导数,得到切线方向为 \((1, 1, 1)\),方程为 \(x = y = z - 1\)

  4. 求过直线 \[ L: \begin{cases} 3x - 2y - z = 5 \\ x + y + z = 0 \end{cases} \] 与曲面 \[ 2x^2 - 2y^2 +2z = \frac{5}{8} \] 相切的切平面方程

    曲面上的点 \((x_0, y_0, z_0)\) 的切平面法向量为 \((4x_0, -4y_0, 2)\),过直线 \(L\) 的平面簇方程为 \(3x - 2y - z - 5 + \lambda(x + y + z) = 0 = (3 + \lambda)x + (\lambda - 2)y + (\lambda - 1)z - 5\)。得到该平面的法向量为 \((3 + \lambda, \lambda - 2, \lambda - 1)\)。整理得: \[ \begin{cases} \frac{3 + \lambda}{4x_0} = \frac{\lambda - 2}{-4y_0} = \frac{\lambda - 1}{2} \\ 2x_0^2 - 2y_0^2 + 2z_0 = \frac{5}{8} \\ (3 + \lambda)x_0 + (\lambda - 2)y_0 + (\lambda - 1)z_0 - 5 = 0 \end{cases} \]

    解得 $= $

  5. 长度为 \(a\) 的线段两端分别在 \(x\) 轴和 \(y\) 轴上滑动,求这样的线段簇的包络线

    设在 \(x\) 轴上的位置为 \(c\),则该线段所在直线的方程为 \[ \frac{x}{c} + \frac{y}{\sqrt{a^2 - c^2}} = 1 \]

    \(\sin \theta = \frac{c}{a}\),则直线方程为 \[ F(x, y, \theta) = \frac{x}{\sin \theta} + \frac{y}{\cos \theta} -\frac{1}{a} = 0 \]

    则包络线方程为: \[ \begin{cases} F(x, y, \theta) = \frac{x}{\sin \theta} + \frac{y}{\cos \theta} -\frac{1}{a} = 0 \\ F_\theta(x, y, \theta) = \frac{-x \cos \theta}{\sin^2 \theta} + \frac{y \sin \theta}{\cos^2 \theta} = 0 \end{cases} \]

    从而,\(x = a \sin^3 \theta\)\(y = a \cos^3 \theta\)

  6. 求通过直线

    \[ L: \begin{cases} 2x + y - 3z + 2 = 0 \\ 5x + 5y - 4z + 3 = 0 \end{cases} \]

    的两个相互垂直的平面 \(\pi_1\)\(\pi_2\),使得其中一个平面过点 \((4, -3, 1)\)

    过直线 \(L\) 的曲面系为 \(\Sigma: (2x + y - 3z + 2) + \lambda(5x + 5y - 4z + 3) = 0\),要求过点 \((4, -3, 1)\) 解得 \(\lambda = -1\),故 \(\Sigma: 3x + 4y - z + 1 = 0\),记为 \(\pi_1\),则 \(\pi_2: x - 2y - 5z + 3 = 0\)

  7. 设二元函数 \(f(x, y) = |x - y| \varphi(x, y)\),其中 \(\varphi(x, y)\) 在点 \((0, 0)\) 处连续。证明 \(f(x, y)\) 在点 \((0, 0)\) 处可微的充分必要条件是 \(\varphi(0, 0) = 0\)

    必要性。若 \(f(x, y)\)\((0, 0)\) 处可微,且 \(f_x(0, 0)\) 存在

    \[ \begin{aligned} \lim_{x \to 0^+} \frac{f(x, 0) - f(0, 0)}{x} &= \varphi(0, 0) \\ \lim_{x \to 0^-} \frac{f(x, 0) - f(0, 0)}{x} &= -\varphi(0, 0) \end{aligned} \]

    故,\(\varphi(0, 0) = 0\)

    充分性性。若 \(\varphi(0, 0) = 0\),则

    \[ \begin{aligned} f_x(0, 0) &= \lim_{x \to 0} \frac{f(x, 0) - f(0, 0)}{x} = 0 \\ f_y(0, 0) &= \lim_{y \to 0} \frac{f(0, y) - f(0, 0)}{y} = 0 \\ o(\rho) &= f(\Delta x, \Delta y) - f(0, 0) - f_x(0, 0) \Delta x - f_y(0, 0) \Delta y \end{aligned} \]

    解得

    \[ \lim_{(\Delta x, \Delta y) \to (0, 0)} \frac{f(\Delta x, \Delta y) - f(0, 0) - f_x(0, 0) \Delta x - f_y(0, 0) \Delta y}{\sqrt{\Delta^2 x + \Delta^2 y}} \\ = \lim_{(\Delta x, \Delta y) \to (0, 0)} \frac{|\Delta x - \Delta y| \varphi(\Delta x, \Delta y)}{\sqrt{\Delta^2 x + \Delta^2 y}} = 0 \]

    所以,

    \[ \Delta f = f(\Delta x, \Delta y) - f(0, 0) = f_x(0, 0) \Delta x + f_y(0, 0) \Delta y + |\Delta x - \Delta y| \varphi(\Delta x, \Delta y) \\ = f_x(0, 0) \Delta x + f_y(0, 0) \Delta y + o(\rho) \]

    故,\(f(x, y)\)\((0, 0)\) 可微。

  8. 求原函数

    \[ \text{d} u = \frac{x \text{d} y - y \text{d} x}{x^2 + y^2} \]

    \[ u = \arctan \frac{y}{x} + C \]

  9. 设椭圆簇为 \(\frac{x^2}{a^2} + \frac{y^2}{b^2} = 1\),其中每个椭圆的面积都为常数 \(S\)。求该椭圆簇的包络线。

    由面积恒为定值得到 \(ab\pi = S\),则椭圆簇可以表示为

    \[ F(x, y, a) = \frac{x^2}{a^2} + \frac{a^2\pi^2y^2}{S^2} = 1 \]

    包络线方程为

    \[ \begin{cases} F(x, y ,a) &= \frac{x^2}{a^2} + \frac{a^2\pi^2y^2}{S^2} - 1 = 0 \\ F_a(x, y, a) &= \frac{-2x^2}{a^3} + \frac{2a\pi^2y^2}{S^2} = 0 \end{cases} \]

    则包络线的方程

    \[ C: x^2 y^2 = \frac{S^2}{4\pi^2} \]

  10. 求直线 \(\frac{x - 1}{0} = \frac{y - 1}{1} = \frac{z - 1}{1}\)\(z\) 轴的旋转面方程。

    该直线的参数方程为:\(l: x = 1\)\(y = 1 + t\)\(z = 1 + t\)。得到旋转面的参数方程:\(\Sigma: x = \sqrt{2 + 2t + t^2} \cos \theta\)\(y = \sqrt{2 + 2t + t^2} \sin \theta\)\(z = 1 + t\)\(t \in R\)\(\theta \in [0, 2\pi]\)。即 \(\Sigma: x^2 + y^2 - z^2 = 1\)

  11. \(F(x, y, z)\)\(n\) 次齐次函数,即存在正整数 \(n\),使得对任何 \(t\) 都有 \(F(tx, ty, tz) = t^n F(x, y, z)\)。证明:方程 \(F(x, y, z) = 0\) 所代表的曲面是顶点在原点的锥面。

    \(t = 0\),得 \(F(0, 0, 0) = 0\)。设非原点 \((x_0, y_0. z_0)\) 在曲面上,由于 \(F(tx, ty, tz) = t^n F(x, y, z)\),故经过原点的直线 \(x = tx_0\)\(y = ty_0\)\(z = tz_0\) 在曲面上。这表明曲面上的任何点与原点的连线都在曲面上。因此该曲面为顶点在原点的锥面。

  12. 给定二次曲面 \((x + y)^2 + (y + z)^2 + (z + x)^2 - 1 =0\),求它在 \(xOy\) 坐标面上的投影区域 \(D\)

    做变换 \(T: x + y = u\)\(y + z = v\)\(z + x = w\)

  13. 证明:若 \(u = u(x, y)\)\(v = v(x, y)\) 都是区域 \(D\) 上的调和函数,则 \(u(x, y) \equiv C_1\)(常数)及 \(v(x, y) \equiv C_2\)(常数)的充要条件是 \(u^2(x, y) + v^2(x, y) \equiv C\)(常数)。

    充分性显然,下面证明必要性。对等式分别对 \(x\)\(y\) 求偏导,并结合调和函数性质得:

    \[ \begin{aligned} \frac{\partial u}{\partial x} + \frac{\partial v}{\partial x} &= 0 \\ \frac{\partial u}{\partial y} + \frac{\partial v}{\partial y} &= 0 \\ \frac{\partial^2 u}{\partial x^2} + \frac{\partial^2 v}{\partial x^2} &= 0 \\ \frac{\partial^2 u}{\partial y^2} + \frac{\partial^2 v}{\partial y^2} &= 0 \\ \frac{\partial^2 u}{\partial x^2} + \frac{\partial^2 u}{\partial y^2} &= 0 \\ \frac{\partial^2 v}{\partial x^2} + \frac{\partial^2 v}{\partial y^2} &= 0 \end{aligned} \]

    \(a \in R\),则

    \[ \begin{aligned} \frac{\partial^2 u}{\partial x^2} &= \frac{\partial^2 v}{\partial y^2} = a \\ \frac{\partial^2 u}{\partial y^2} &= \frac{\partial^2 v}{\partial x^2} = -a \end{aligned} \]

    \[ \begin{aligned} \frac{\partial u}{\partial x} &= ax + C_x, C_x \in R \\ \frac{\partial u}{\partial y} &= ay + C_y, C_y \in R \end{aligned} \]

    又因为

    \[ \frac{\partial u}{\partial x} + \frac{\partial v}{\partial x} = 0 \]

    所以

    \[ \frac{\partial u}{\partial x} = \frac{\partial v}{\partial x} = 0 \]

    同理

    \[ \frac{\partial u}{\partial y} = \frac{\partial v}{\partial y} = 0 \]

    所以,\(u(x, y) \equiv C_1\)\(v(x, y) \equiv C_2\)

  14. \(f\) 二阶连续可微,求下列复合函数的偏导数:

    (1). \(u = f(x^2 + y^2 + z^2)\),求 \(\frac{\partial^2 u}{\partial x^2}\)

    (2). \(u = f(x + y, xy)\),求 \(\frac{\partial^2 u}{\partial x \partial y}\)

    (3). \(u = f(x, xy, xyz)\),求 \(\frac{\partial u}{x}\)\(\frac{\partial u}{y}\)\(\frac{\partial^2 u}{\partial x \partial y}\)

    (1).

    \[ \begin{aligned} \frac{\partial u}{\partial x} &= 2x f' \\ \frac{\partial^2 u}{\partial x^2} &= 2f' + 4x^2 f'' \end{aligned} \]

    (2).

    \[ \begin{aligned} \frac{\partial u}{\partial x} &= f_1' + yf_2' \\ \frac{\partial^2 u}{\partial x \partial y} &= f_{11}'' + x f_{12}' + f_2' + y f_{21}' + xy f_{22}' \end{aligned} \]

    (3).

    \[ \begin{aligned} \frac{\partial u}{x} &= f_{1}' + yf_{2}' + yzf_{3}' \\ \frac{\partial u}{y} &= xf_{2}' + xzf_{3}' \\ \frac{\partial^2 u}{\partial x \partial y} &= f_{2}' + x(f_{21}'' + yf_{22}'' + yzf_{23}'') + zf_{3}' + xz(f_{31}'' + yf_{32}'' + yzf_{33}'') \end{aligned} \]

  15. \(z = f(x, y)\) 具有二阶连续的偏导数,且满足 \(f(x, 2x) = x\)\(f_1'(x, 2x) = x^2\)\(f_{11}'' = f_{22}''\)。求二阶偏导数 \(f_{12}''(x, 2x)\)

    \[ \begin{aligned} \frac{\partial f}{\partial x} &= f_{1}' + 2f_{2}' = 1 \\ \frac{\partial^2 f}{\partial x^2} &= f_{11}'' + 2f_{12}'' + 2f_{21}'' + 4f_{22}'' = 0 \\ \frac{\partial f_{1}}{\partial x} &= f_{11}'' + 2f_{12}'' = 2x \\ f_{12}''(x, 2x) &= \frac{5}{3} x \end{aligned} \]

  16. 求偏微分方程

    \[ y \frac{\partial z}{\partial x} - x \frac{\partial z}{\partial y} = 0 \]

    做变换 \(u = x\)\(v = x^2 + y^2\),则:

    \[ \begin{aligned} \frac{\partial z}{\partial x} &= \frac{\partial z}{\partial u} \frac{\partial u}{\partial x} + \frac{\partial z}{\partial v} \frac{\partial v}{\partial x } = \frac{\partial z}{\partial u} + 2x \frac{\partial z}{\partial v} \\ \frac{\partial z}{\partial y} &= \frac{\partial z}{\partial u} \frac{\partial u}{\partial y} + \frac{\partial z}{\partial v} \frac{\partial v}{\partial y} = 2y \frac{\partial z}{\partial v} \end{aligned} \]

    带入偏微分方程得到:

    \[ y \biggl(\frac{\partial z}{\partial u} + 2x \frac{\partial z}{\partial v}\biggr) - 2xy \frac{\partial z}{\partial v} = y \frac{\partial z}{\partial u} = 0 \]

    因此 \(z = \varphi(x^2 + y^2)\),其中 \(\varphi\) 是任意可导函数。

  17. 设函数 \(u = f(\sqrt{x^2 + y^2})\),其中 \(F\) 具有连续的二阶导数,且满足 \(\frac{\partial^2 u}{\partial x^2} + \frac{\partial^2 u}{\partial y^2} = x^2 + y^2\),求函数 \(u\) 的表达式。

    \(r = \sqrt{x^2 + y^2}\),则:

    \[ \begin{aligned} \frac{\partial u}{\partial x} &= \frac{\partial u}{\partial r} \frac{\partial r}{\partial x} = \frac{xf'(r)}{r} \\ \frac{\partial^2 u}{\partial x^2} &= \frac{f'(r)}{r} + x \frac{\partial}{\partial r} \biggl(\frac{f'(r)}{r}\biggr) \frac{\partial r}{\partial x} = \frac{f'(r)}{r} + \frac{x^2}{r^2} f''(r) - \frac{x^2}{r^3} f'(r) \\ \frac{\partial u}{\partial y} &= \frac{\partial u}{\partial r} \frac{\partial r}{\partial y} = \frac{yf'(r)}{r} \\ \frac{\partial^2 u}{\partial y^2} &= \frac{f'(r)}{r} + y \frac{\partial}{\partial r} \biggl(\frac{f'(r)}{r}\biggr) \frac{\partial r}{\partial y} = \frac{f'(r)}{r} + \frac{y^2}{r^2} f''(r) - \frac{y^2}{r^3} f'(r) \end{aligned} \]

    带入微分方程得到:

    \[ \begin{aligned} \frac{f'(r)}{r} + \frac{x^2}{r^2} f''(r) - \frac{x^2}{r^3} f'(r) + \frac{f'(r)}{r} + \frac{y^2}{r^2} f''(r) - \frac{y^2}{r^3} f'(r) &= \frac{f'(r)}{r} + f''(r) = r^2 \\ \frac{\text{d}}{\text{d} r} \biggl(r f'(r)\biggr) &= r^3 \\ f'(r) &= \frac{r^3}{4} + \frac{C_1}{r} \\ f(r) &= \frac{1}{16} r^4 + C_1\ln r + C_2, C_1, C_2 \in R \end{aligned} \]

  18. \(u = f(\ln \sqrt{x^2 + y^2})\),其中 \(f\) 有连续的二阶导数,且满足 \(\frac{\partial^2 u}{\partial x^2} + \frac{\partial^2 u}{\partial y^2} = (x^2 + y^2)^{\frac{3}{2}}\),求 \(f(v)\) 的表达式。

    \(v = \ln r\)\(r = \sqrt{x^2 + y^2}\),由链式法则得到:

    \[ \begin{aligned} \frac{\partial u}{\partial x} &= \frac{\text{d} f}{\text{d} v} \frac{\text{d} v}{\text{d} r} \frac{\partial r}{\partial x} = f'(v) \frac{x}{r^2} \\ \frac{\partial u}{\partial y} &= \frac{\text{d} f}{\text{d} v} \frac{\text{d} v}{\text{d} r} \frac{\partial r}{\partial y} = f'(v) \frac{y}{r^2} \\ \frac{\partial^2 u}{\partial x^2} &= \frac{\text{d}}{\text{d} x} \biggl(f'(v) \frac{x}{r^2}\biggr) = \frac{f'(v)}{r^2} + x^2 \frac{f''(v)}{r^4} - 2x^2 \frac{f'(v)}{r^4} \\ \frac{\partial^2 u}{\partial y^2} &= \frac{\text{d}}{\text{d} y} \biggl(f'(v) \frac{y}{r^2}\biggr) = \frac{f'(v)}{r^2} + y^2 \frac{f''(v)}{r^4} - 2y^2 \frac{f'(v)}{r^4} \end{aligned} \]

    带入微分方程得到:

    \[ \begin{aligned} \frac{f''(v)}{r^2} &= r^3 \\ f(v) &= \frac{1}{25}\text{e}^{5v} + C_1 v + C_2 \end{aligned} \]

  19. 证明:在某个开区域上有恒等式 \(f(x, y) = \arctan x + \arctan y - \arctan \frac{x + y}{1 - xy} \equiv \pi\),并求出使之成立的开区域。

    \(f(x, y)\) 分别对 \(x\)\(y\) 求偏导

    \[ \begin{aligned} \frac{\partial f}{\partial x} &= \frac{1}{1 + x^2} - \frac{1 + y^2}{(1 + x^2)(1 + y^2)} \equiv 0 \\ \frac{\partial f}{\partial y} &= \frac{1}{1 + y^2} - \frac{1 + x^2}{(1 + x^2)(1 + y^2)} \equiv 0 \end{aligned} \]

    经检验,开区域:\(xy > 1\)\(x > 0\)

  20. 设二元函数 \(f(x, y)\) 有一阶连续偏导数,且 \(f(0, 1) \equiv f(1, 0)\)。证明:在单位圆周 \(x^2 + y^2 = 1\) 上至少存在两个不同的点满足方程 \(y \frac{\partial f}{\partial x} = x \frac{\partial f}{\partial y}\)

    单位圆周的参数方程为 \(x = \cos \theta\)\(y = \sin \theta\)\(0 \leq \theta \leq 2\pi\),设 \(F(\theta) = f(\cos \theta, \sin \theta)\)\(F(0) = F(\frac{\pi}{2}) = F(2\pi)\),由罗尔定理知存在 \(\xi \in (0, \frac{\pi}{2})\)\(\eta \in (\frac{\pi}{2}, 2\pi)\),使得 \(F'(\xi) = F'(\eta) = 0\),而 \(F'(\theta) = -\sin \theta f_1' + \cos \theta f_2'\),带入 \(\xi\)\(\eta\) 得证。

  21. \(f(x, y)\) 在区域 \(D: x^2 + y^2 \leq R^2\) 上连续,满足 \(x \frac{\partial f}{\partial x} + ky\frac{\partial f}{\partial y} = 0\),其中 \(k\) 是正整数。证明:在 \(D\)\(f(x, y)\) 恒为常数。

    对于区域 \(D\) 内的点 \((x_0, y_0)\)\(x_0 \neq 0\),不妨设 \(0 < x_0 < R\),做连接点与原点的曲线 \(L: x = t, y = \frac{y_0}{x_0^k} t^k\)\(t \in [0, x_0]\),构造辅助函数 \(F(t) = f(t, \frac{y_0}{x_0^k} t^k)\)

  22. 设二元函数 \(f(x, y)\) 在平面上有连续的二阶偏导数,对任何角度 \(\alpha\),定义一元函数 \(g_a(t) = f(t\cos \alpha, t\sin \alpha)\)。若对于任何 \(\alpha\) 都有 \(\frac{\text{d} g_a(0)}{\text{d} t} = 0\)\(\frac{\text{d}^2 g_a(0)}{\text{d} t^2} > 0\),证明 \(f(0, 0)\)\(f(x, y)\) 的极小值。

    \(x = t\cos \alpha\)\(y = t\sin \alpha\),由链式法则得

    \[ \begin{aligned} \frac{\partial f}{\partial t} &= \frac{\partial f}{\partial x} \frac{\partial x}{\partial t} + \frac{\partial f}{\partial y} \frac{\partial y}{\partial t} = f_x' \cos \alpha + f_y' \sin \alpha \\ \frac{\partial^2 f}{\partial t^2} &= \cos \alpha (f_{xx}'' \cos \alpha + f_{xy}'' \sin \alpha) + \sin \alpha (f_{xy}'' \cos \alpha + f_{yy}'' \sin \alpha) \end{aligned} \]

    函数 \(f(x, y)\)\((0, 0)\) 的 Hessian 矩阵为 \(\mathbf{H} = \begin{bmatrix} f_{xx}'' \quad f_{xy}'' \\ f_{yx}'' \quad f_{yy}''\end{bmatrix}\),极小值条件等价于 \(f_{xx}'' f_{yy}'' - f_{xy}''^2 > 0\)\(f_x'(0, 0) = f_y'(0, 0) = 0\)。根据题目条件,可推出

    \[ \frac{\partial f}{\partial t} \bigg|_{t = 0} = f_x'(0, 0) \cos \alpha + f_y'(0, 0) \sin \alpha = 0 \]

    因为对于任何 \(\alpha\) 都上上式,同时因为 \(f_x'(0, 0)\)\(f_y'(0, 0)\) 存在,所以 \(f_x'(0, 0) = f_y'(0, 0) = 0\)。除此之外

    \[ \frac{\partial^2 f}{\partial t^2} \bigg|_{t = 0} = \cos \alpha (f_{xx}'' \cos \alpha + f_{xy}'' \sin \alpha) + \sin \alpha (f_{xy}'' \cos \alpha + f_{yy}'' \sin \alpha) > 0 \]

    对于任何非零向量 \(\mathbf{r} = |\mathbf{r}|(\cos \alpha, \sin \alpha)\)\(|\mathbf{r}| \neq 0\),因此二次型

    \[ |\mathbf{r}|^2 \begin{pmatrix}\cos \alpha, \sin \alpha \end{pmatrix} \begin{pmatrix} f_{xx}'' (0, 0) \quad f_{xy}''(0, 0) \\ f_{yx}''(0, 0) \quad f_{yy}''(0, 0)\end{pmatrix} \begin{pmatrix}\cos \alpha \\ \sin \alpha \end{pmatrix} > 0 \]

    于是 \(\mathbf{H}\) 为正定矩阵,则其顺序主子式都大于零,即 \(f_{xx}'' > 0\)\(f_{xx}'' f_{yy}'' - f_{xy}''^2 > 0\)。综上,是极小值点。


多元函数微分学
https://ddccffq.github.io/2025/08/20/数学竞赛/多元函数微分学/
作者
ddccffq
发布于
2025年8月20日
许可协议