偏导数的计算与复合函数求导

偏导数的基本计算

计算偏导数时,将其他自变量视为常数,按一元函数求导法则进行。基本求导公式与一元函数完全相同。

常用求导公式

x(c)=0,x(xn)=nxn1,x(ex)=ex\frac{\partial}{\partial x}(c) = 0, \quad \frac{\partial}{\partial x}(x^n) = nx^{n-1}, \quad \frac{\partial}{\partial x}(e^x) = e^x x(lnx)=1x,x(sinx)=cosx,x(cosx)=sinx\frac{\partial}{\partial x}(\ln x) = \frac{1}{x}, \quad \frac{\partial}{\partial x}(\sin x) = \cos x, \quad \frac{\partial}{\partial x}(\cos x) = -\sin x

复合函数求导法则(链式法则)

情形一:中间变量为一元函数

定理1

z=f(u,v)z = f(u, v)u=u(t)u = u(t)v=v(t)v = v(t),则复合函数 z=f(u(t),v(t))z = f(u(t), v(t))tt 的导数(全导数)为:

dzdt=zududt+zvdvdt\frac{dz}{dt} = \frac{\partial z}{\partial u} \cdot \frac{du}{dt} + \frac{\partial z}{\partial v} \cdot \frac{dv}{dt}

几何解释
推论
证明
符号说明

情形二:中间变量为二元函数

定理2

z=f(u,v)z = f(u, v)u=u(x,y)u = u(x, y)v=v(x,y)v = v(x, y),则复合函数 z=f(u(x,y),v(x,y))z = f(u(x,y), v(x,y)) 的偏导数为:

zx=zuux+zvvx\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}

zy=zuuy+zvvy\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}

几何解释
推论
证明
符号说明

情形三:特殊情形

  1. 一个中间变量z=f(u,x,y)z = f(u, x, y)u=u(x,y)u = u(x, y),则 zx=fuux+fx,zy=fuuy+fy\frac{\partial z}{\partial x} = \frac{\partial f}{\partial u}\frac{\partial u}{\partial x} + \frac{\partial f}{\partial x}, \quad \frac{\partial z}{\partial y} = \frac{\partial f}{\partial u}\frac{\partial u}{\partial y} + \frac{\partial f}{\partial y} 注意 fx\frac{\partial f}{\partial x} 是将 uu 视为常数对 xx 求偏导,与 zx\frac{\partial z}{\partial x} 不同。

  2. 全导数推广z=f(u,v,w)z = f(u, v, w)u=u(t),v=v(t),w=w(t)u=u(t), v=v(t), w=w(t),则 dzdt=zududt+zvdvdt+zwdwdt\frac{dz}{dt} = \frac{\partial z}{\partial u}\frac{du}{dt} + \frac{\partial z}{\partial v}\frac{dv}{dt} + \frac{\partial z}{\partial w}\frac{dw}{dt}

隐函数求导

一个方程的情形

定理3

设函数 F(x,y,z)=0F(x, y, z) = 0 确定隐函数 z=z(x,y)z = z(x, y),且 Fz0F_z \neq 0,则

zx=FxFz,zy=FyFz\frac{\partial z}{\partial x} = -\frac{F_x}{F_z}, \quad \frac{\partial z}{\partial y} = -\frac{F_y}{F_z}

几何解释
推论
证明
符号说明

类似地,F(x,y)=0F(x,y)=0 确定 y=y(x)y=y(x) 时,dydx=FxFy\frac{dy}{dx} = -\frac{F_x}{F_y}Fy0F_y \neq 0)。

推导方法

对方程 F(x,y,z(x,y))=0F(x, y, z(x,y)) = 0 两边对 xx 求偏导:

Fx+Fzzx=0zx=FxFzF_x + F_z \cdot \frac{\partial z}{\partial x} = 0 \Rightarrow \frac{\partial z}{\partial x} = -\frac{F_x}{F_z}

典型例题

例题1:复合函数求导

z=eusinvz = e^u \sin vu=xyu = xyv=x+yv = x + y,求 zx\frac{\partial z}{\partial x}zy\frac{\partial z}{\partial y}

参考答案(3 个标签)
复合函数链式法则偏导数

详细步骤

  1. zu=eusinv\frac{\partial z}{\partial u} = e^u \sin vzv=eucosv\frac{\partial z}{\partial v} = e^u \cos v
  2. ux=y\frac{\partial u}{\partial x} = yuy=x\frac{\partial u}{\partial y} = xvx=1\frac{\partial v}{\partial x} = 1vy=1\frac{\partial v}{\partial y} = 1
  3. zx=eusinvy+eucosv1=exy[ysin(x+y)+cos(x+y)]\frac{\partial z}{\partial x} = e^u \sin v \cdot y + e^u \cos v \cdot 1 = e^{xy}[y\sin(x+y) + \cos(x+y)]
  4. zy=eusinvx+eucosv1=exy[xsin(x+y)+cos(x+y)]\frac{\partial z}{\partial y} = e^u \sin v \cdot x + e^u \cos v \cdot 1 = e^{xy}[x\sin(x+y) + \cos(x+y)]

答案zx=exy[ysin(x+y)+cos(x+y)]\frac{\partial z}{\partial x} = e^{xy}[y\sin(x+y)+\cos(x+y)]zy=exy[xsin(x+y)+cos(x+y)]\frac{\partial z}{\partial y} = e^{xy}[x\sin(x+y)+\cos(x+y)]

例题2:全导数

z=arctan(xy)z = \arctan(xy)y=exy = e^x,求 dzdx\frac{dz}{dx}

参考答案(3 个标签)
全导数复合函数链式法则

详细步骤

方法一:全导数公式

  1. zx=y1+x2y2\frac{\partial z}{\partial x} = \frac{y}{1+x^2y^2}zy=x1+x2y2\frac{\partial z}{\partial y} = \frac{x}{1+x^2y^2}dydx=ex\frac{dy}{dx} = e^x
  2. dzdx=zx+zydydx=y1+x2y2+x1+x2y2ex=ex+xex1+x2e2x=ex(1+x)1+x2e2x\frac{dz}{dx} = \frac{\partial z}{\partial x} + \frac{\partial z}{\partial y}\frac{dy}{dx} = \frac{y}{1+x^2y^2} + \frac{x}{1+x^2y^2} \cdot e^x = \frac{e^x + xe^x}{1+x^2e^{2x}} = \frac{e^x(1+x)}{1+x^2e^{2x}}

方法二:直接代入 z=arctan(xex)z = \arctan(xe^x)dzdx=(xex)1+(xex)2=ex+xex1+x2e2x\frac{dz}{dx} = \frac{(xe^x)'}{1+(xe^x)^2} = \frac{e^x + xe^x}{1+x^2e^{2x}}

答案dzdx=ex(1+x)1+x2e2x\frac{dz}{dx} = \frac{e^x(1+x)}{1+x^2e^{2x}}

例题3:隐函数求导

x2+y2+z2=4zx^2 + y^2 + z^2 = 4z 确定 z=z(x,y)z = z(x,y),求 zx\frac{\partial z}{\partial x}zy\frac{\partial z}{\partial y}

参考答案(3 个标签)
隐函数隐函数求导偏导数

详细步骤

  1. F(x,y,z)=x2+y2+z24zF(x,y,z) = x^2 + y^2 + z^2 - 4z
  2. Fx=2xF_x = 2xFy=2yF_y = 2yFz=2z4F_z = 2z - 4
  3. zx=FxFz=2x2z4=x2z\frac{\partial z}{\partial x} = -\frac{F_x}{F_z} = -\frac{2x}{2z-4} = \frac{x}{2-z}
  4. zy=FyFz=2y2z4=y2z\frac{\partial z}{\partial y} = -\frac{F_y}{F_z} = -\frac{2y}{2z-4} = \frac{y}{2-z}

答案zx=x2z\frac{\partial z}{\partial x} = \frac{x}{2-z}zy=y2z\frac{\partial z}{\partial y} = \frac{y}{2-z}


练习题

练习1

z=u2lnvz = u^2 \ln vu=xyu = \frac{x}{y}v=3x2yv = 3x - 2y,求 zx\frac{\partial z}{\partial x}

参考答案(3 个标签)
复合函数链式法则偏导数

详细步骤

  1. zu=2ulnv\frac{\partial z}{\partial u} = 2u\ln vzv=u2v\frac{\partial z}{\partial v} = \frac{u^2}{v}
  2. ux=1y\frac{\partial u}{\partial x} = \frac{1}{y}vx=3\frac{\partial v}{\partial x} = 3
  3. zx=2ulnv1y+u2v3=2xy2ln(3x2y)+3x2y2(3x2y)\frac{\partial z}{\partial x} = 2u\ln v \cdot \frac{1}{y} + \frac{u^2}{v} \cdot 3 = \frac{2x}{y^2}\ln(3x-2y) + \frac{3x^2}{y^2(3x-2y)}

答案zx=2xy2ln(3x2y)+3x2y2(3x2y)\frac{\partial z}{\partial x} = \frac{2x}{y^2}\ln(3x-2y) + \frac{3x^2}{y^2(3x-2y)}

练习2

ezxyz=0e^z - xyz = 0 确定 z=z(x,y)z = z(x,y),求 zx\frac{\partial z}{\partial x}

参考答案(2 个标签)
隐函数隐函数求导

详细步骤

  1. F(x,y,z)=ezxyzF(x,y,z) = e^z - xyz
  2. Fx=yzF_x = -yzFz=ezxyF_z = e^z - xy
  3. zx=FxFz=yzezxy=yzxyzxy=zx(z1)\frac{\partial z}{\partial x} = -\frac{F_x}{F_z} = \frac{yz}{e^z - xy} = \frac{yz}{xyz - xy} = \frac{z}{x(z-1)}(利用 ez=xyze^z = xyz

答案zx=yzezxy=zx(z1)\frac{\partial z}{\partial x} = \frac{yz}{e^z-xy} = \frac{z}{x(z-1)}


总结

本文出现的符号

符号类型读音/说明在本文中的含义
dzdt\frac{dz}{dt}全导数total derivative复合函数对单一自变量的导数
Fx,Fy,FzF_x, F_y, F_z偏导数partial derivatives隐函数方程对各变量的偏导数
u,v,wu, v, w中间变量intermediate variables复合函数的中间变量

中英对照

中文术语英文术语音标
复合函数composite function/kəmˈpɒzɪt ˈfʌŋkʃən/
链式法则chain rule/tʃeɪn ruːl/
全导数total derivative/ˈtoʊtl dɪˈrɪvətɪv/
隐函数implicit function/ɪmˈplɪsɪt ˈfʌŋkʃən/
隐函数求导implicit differentiation/ɪmˈplɪsɪt ˌdɪfərɛnʃiˈeɪʃən/
中间变量intermediate variable/ˌɪntərˈmiːdiət ˈvɛriəbəl/
自变量independent variable/ˌɪndɪˈpɛndənt ˈvɛriəbəl/