本文主要内容如下:
- 1. 复变函数的导数与微分
- 1.1. 复变函数可导、可微、解析与奇点的定义
- 1.2. 复变函数可微的充要条件
- 1.3. 关于复变函数可微性判定的其它形式
- 1.4. 相关结论
- 1.5. 解析函数的构造
- 2. 解析函数与调和函数
- 2.1. 调和函数与共轭调和函数
- 2.2. 解析函数与调和函数的关系
1. 复变函数的导数与微分
1.1. 复变函数可导、可微、解析与奇点的定义
定义:设复变函数
w
=
f
(
z
)
w=f(z)
w=f(z) 在包含
z
0
z_0
z0 的邻域内有定义,若
z
z
z 按照任意方式趋近于
z
0
z_0
z0 ,比值
Δ
w
Δ
z
=
f
(
z
)
−
f
(
z
0
)
z
−
z
0
=
f
(
z
0
+
Δ
z
)
−
f
(
z
0
)
Δ
z
\dfrac{\Delta w}{\Delta z}=\dfrac{f(z)-f(z_0)}{z-z_0}=\dfrac{f(z_0+\Delta z)-f(z_0)}{\Delta z}
ΔzΔw=z−z0f(z)−f(z0)=Δzf(z0+Δz)−f(z0)
的极限均存在,且为有限值,则称复变函数
w
=
f
(
z
)
w=f(z)
w=f(z) 在点
z
0
z_0
z0 处可导。该比值的极限称作
w
=
f
(
z
)
w=f(z)
w=f(z) 在点
z
0
z_0
z0 处的导数,记作
f
′
(
z
0
)
f'(z_0)
f′(z0) 或
d
w
d
z
∣
z
=
z
0
\left.\dfrac{dw}{dz}\right|_{z=z_0}
dzdw
z=z0,即
f
′
(
z
0
)
=
lim
Δ
z
→
0
Δ
w
Δ
z
=
lim
z
→
z
0
f
(
z
)
−
f
(
z
0
)
z
−
z
0
f'(z_0)=\lim_{\Delta z\rightarrow0}\dfrac{\Delta w}{\Delta z}=\lim_{z\rightarrow z_0}\dfrac{f(z)-f(z_0)}{z-z_0}
f′(z0)=Δz→0limΔzΔw=z→z0limz−z0f(z)−f(z0)
或
Δ
w
=
f
′
(
z
0
)
Δ
z
+
o
(
∣
Δ
z
∣
)
(
Δ
z
→
0
)
\Delta w=f'(z_0)\Delta z+o(|\Delta z|)\quad(\Delta z\rightarrow 0)
Δw=f′(z0)Δz+o(∣Δz∣)(Δz→0)
称
f
′
(
z
0
)
Δ
z
f'(z_0)\Delta z
f′(z0)Δz 为
w
w
w 在
z
0
z_0
z0 处的微分,记作
d
f
(
z
0
)
df(z_0)
df(z0)。此时也称
w
w
w 在
z
0
z_0
z0 处可微。
定义:若复变函数
w
=
f
(
z
)
w=f(z)
w=f(z) 在
z
0
z_0
z0及其领域内处处可导,则称
f
(
z
)
f(z)
f(z) 在
z
0
z_0
z0 处解析;若
w
w
w在区域D内每点均解析,称
w
w
w在区域D内解析,或称
w
w
w 是区域D内的解析函数 / 全纯函数 / 正则函数;称函数在闭区域
D
ˉ
\bar D
Dˉ内解析是指函数在包含
D
ˉ
\bar D
Dˉ的某个更大的区域内解析。
Remark: 根据定义:
{
点解析
⇒
点可导
区域解析
⇔
区域可导
\left\{\begin{aligned} &点解析\Rightarrow点可导\\ \\ &区域解析\Leftrightarrow区域可导 \end{aligned}\right.
⎩
⎨
⎧点解析⇒点可导区域解析⇔区域可导
定义:若
f
(
z
)
f(z)
f(z) 在点
z
0
z_0
z0 处不解析,但在
z
0
z_0
z0 的任意一邻域内总存在
f
(
z
)
f(z)
f(z) 的解析点,则称
z
0
z_0
z0 为函数
f
(
z
)
f(z)
f(z) 的奇点。
1.2. 复变函数可微的充要条件
定理:(可微的充要条件) 设复变函数
f
(
z
)
=
u
(
x
,
y
)
+
i
v
(
x
,
y
)
f(z)=u(x,y)+iv(x,y)
f(z)=u(x,y)+iv(x,y) 在区域D内有定义,且在D内一点
z
=
x
+
i
y
z=x+iy
z=x+iy 可微的充要条件是:二元函数
u
(
x
,
y
)
、
v
(
x
,
y
)
u(x,y)、v(x,y)
u(x,y)、v(x,y) 在
(
x
,
y
)
(x,y)
(x,y) 处的可微,且满足Cauchy-Riemann方程:
{
∂
u
∂
x
=
∂
v
∂
y
∂
u
∂
y
=
−
∂
v
∂
x
\left\{\begin{aligned} \dfrac{\partial u}{\partial x}= \dfrac{\partial v}{\partial y}\\ \\ \dfrac{\partial u}{\partial y}=-\dfrac{\partial v}{\partial x} \end{aligned}\right.
⎩
⎨
⎧∂x∂u=∂y∂v∂y∂u=−∂x∂v
证明:(必要性)由于
f
(
z
)
f(z)
f(z) 在点
z
0
=
x
0
+
i
y
0
z_0=x_0+iy_0
z0=x0+iy0 处可微,则:
Δ
w
=
f
(
z
0
+
Δ
z
)
−
f
(
z
0
)
=
Δ
u
+
i
Δ
v
=
f
′
(
z
)
(
Δ
x
+
i
Δ
y
)
+
o
(
∣
Δ
z
∣
)
\Delta w=f(z_0+\Delta z)-f(z_0)=\Delta u+i\Delta v=f'(z)(\Delta x+i\Delta y)+o(|\Delta z|)
Δw=f(z0+Δz)−f(z0)=Δu+iΔv=f′(z)(Δx+iΔy)+o(∣Δz∣)
对比实部与虚部可得:
{
Δ
u
=
u
(
x
0
+
Δ
x
,
y
0
+
Δ
y
)
−
u
(
x
0
,
y
0
)
=
R
e
(
f
′
)
Δ
x
−
I
m
(
f
′
)
Δ
y
+
o
(
∣
Δ
z
∣
)
Δ
v
=
v
(
x
0
+
Δ
x
,
y
0
+
Δ
y
)
−
v
(
x
0
,
y
0
)
=
R
e
(
f
′
)
Δ
y
+
I
m
(
f
′
)
Δ
x
+
o
(
∣
Δ
z
∣
)
\left\{\begin{aligned} \Delta u=u(x_0+\Delta x,y_0+\Delta y)-u(x_0,y_0)=Re(f')\Delta x-Im(f')\Delta y+o(|\Delta z|)\\ \\ \Delta v=v(x_0+\Delta x,y_0+\Delta y)-v(x_0,y_0)=Re(f')\Delta y+Im(f')\Delta x+o(|\Delta z|) \end{aligned}\right.
⎩
⎨
⎧Δu=u(x0+Δx,y0+Δy)−u(x0,y0)=Re(f′)Δx−Im(f′)Δy+o(∣Δz∣)Δv=v(x0+Δx,y0+Δy)−v(x0,y0)=Re(f′)Δy+Im(f′)Δx+o(∣Δz∣)
故可知复变函数的实部与虚部可微,且满足:
{
R
e
(
f
′
)
=
∂
u
∂
x
=
∂
v
∂
y
I
m
(
f
′
)
=
−
∂
u
∂
y
=
∂
v
∂
x
\left\{\begin{aligned} Re(f')=\dfrac{\partial u}{\partial x}=\dfrac{\partial v}{\partial y}\\\\ Im(f')=-\dfrac{\partial u}{\partial y}=\dfrac{\partial v}{\partial x} \end{aligned}\right.
⎩
⎨
⎧Re(f′)=∂x∂u=∂y∂vIm(f′)=−∂y∂u=∂x∂v
(充分性)设
u
(
x
,
y
)
、
v
(
x
,
y
)
u(x,y)、v(x,y)
u(x,y)、v(x,y)可微,则
u
(
x
,
y
)
、
v
(
x
,
y
)
u(x,y)、v(x,y)
u(x,y)、v(x,y)的偏导数存在,且有:
{
Δ
u
=
u
x
Δ
x
+
u
y
Δ
y
+
o
(
∣
Δ
z
∣
)
Δ
v
=
v
x
Δ
x
+
v
y
Δ
y
+
o
(
∣
Δ
z
∣
)
\left\{\begin{aligned} \Delta u=u_x\Delta x+u_y\Delta y+o(|\Delta z|)\\ \\ \Delta v=v_x\Delta x+v_y\Delta y+o(|\Delta z|) \end{aligned}\right.
⎩
⎨
⎧Δu=uxΔx+uyΔy+o(∣Δz∣)Δv=vxΔx+vyΔy+o(∣Δz∣)
又
u
(
x
,
y
)
、
v
(
x
,
y
)
u(x,y)、v(x,y)
u(x,y)、v(x,y)的偏导数满足 C-R 方程,那么
Δ
w
=
Δ
u
+
i
Δ
v
=
(
u
x
+
i
v
x
)
Δ
x
+
(
u
y
+
i
v
y
)
Δ
y
+
o
(
∣
Δ
z
∣
)
=
(
u
x
−
i
u
y
)
Δ
x
+
(
u
y
+
i
u
x
)
Δ
y
+
o
(
∣
Δ
z
∣
)
=
(
u
x
−
i
u
y
)
(
Δ
x
+
i
Δ
y
)
+
o
(
∣
Δ
z
∣
)
\begin{aligned} \Delta w=\Delta u+i\Delta v&=(u_x+iv_x)\Delta x+(u_y+iv_y)\Delta y+o(|\Delta z|) \\ \\ &=(u_x-iu_y)\Delta x+(u_y+iu_x)\Delta y+o(|\Delta z|)\\ \\ &=(u_x-iu_y)(\Delta x+i\Delta y)+o(|\Delta z|) \end{aligned}
Δw=Δu+iΔv=(ux+ivx)Δx+(uy+ivy)Δy+o(∣Δz∣)=(ux−iuy)Δx+(uy+iux)Δy+o(∣Δz∣)=(ux−iuy)(Δx+iΔy)+o(∣Δz∣)
故
lim
Δ
z
→
0
Δ
w
Δ
z
=
u
x
−
i
u
y
\lim_{\Delta z\rightarrow 0}\dfrac{\Delta w}{\Delta z}=u_x-iu_y
Δz→0limΔzΔw=ux−iuy
证毕.
Remark:
1)上述定理必要性的证明过程给出了一点处复变函数导数的求法:
f
′
(
z
)
=
∂
u
∂
x
+
i
∂
v
∂
x
=
∂
v
∂
y
+
i
∂
v
∂
x
=
∂
u
∂
x
−
i
∂
u
∂
y
=
∂
v
∂
y
−
i
∂
u
∂
y
f'(z)=\dfrac{\partial u}{\partial x}+i\dfrac{\partial v}{\partial x} =\dfrac{\partial v}{\partial y}+i\dfrac{\partial v}{\partial x} =\dfrac{\partial u}{\partial x}-i\dfrac{\partial u}{\partial y} =\dfrac{\partial v}{\partial y}-i\dfrac{\partial u}{\partial y}
f′(z)=∂x∂u+i∂x∂v=∂y∂v+i∂x∂v=∂x∂u−i∂y∂u=∂y∂v−i∂y∂u
2)关于复变函数可微与实/虚部二元函数可微的关系可见下图:
3)复变函数在一点可微,则它在该点处必定连续,因为此时实/虚部函数可微(连续)。
1.3. 关于复变函数可微性判定的其它形式
定理: 若采用指数形式表示复数,即
w
=
f
(
z
)
=
f
(
r
e
i
θ
)
=
u
(
r
,
θ
)
+
i
v
(
r
,
θ
)
w=f(z)=f(re^{i\theta})=u(r,\theta)+iv(r,\theta)
w=f(z)=f(reiθ)=u(r,θ)+iv(r,θ)
此时,
w
=
f
(
z
)
可微
⇔
u
(
r
,
θ
)
,
v
(
r
,
θ
)
可微,且满足
{
r
u
r
=
v
θ
r
v
r
=
−
v
θ
w=f(z)可微\Leftrightarrow u(r,\theta),v(r,\theta)可微,且满足 \left\{\begin{aligned} ru_r=v_\theta \\\\ rv_r=-v_\theta \end{aligned}\right.
w=f(z)可微⇔u(r,θ),v(r,θ)可微,且满足⎩
⎨
⎧rur=vθrvr=−vθ
证明: 由于
{
x
=
r
c
o
s
θ
y
=
r
s
i
n
θ
⇒
{
u
r
=
u
x
c
o
s
θ
+
u
y
s
i
n
θ
u
θ
=
−
u
x
r
s
i
n
θ
+
u
y
r
c
o
s
θ
v
r
=
v
x
c
o
s
θ
+
v
y
s
i
n
θ
v
θ
=
−
v
x
r
s
i
n
θ
+
v
y
r
c
o
s
θ
⇒
{
u
x
=
u
r
c
o
s
θ
−
1
r
u
θ
s
i
n
θ
u
y
=
u
r
s
i
n
θ
+
1
r
u
θ
c
o
s
θ
v
x
=
v
r
c
o
s
θ
−
1
r
v
θ
s
i
n
θ
v
y
=
v
r
s
i
n
θ
+
1
r
v
θ
c
o
s
θ
\left\{\begin{aligned} x=r~cos\theta \\\\ y=r~sin\theta \end{aligned}\right. \Rightarrow \left\{\begin{aligned} &u_r=u_xcos\theta+u_ysin\theta \\\\ &u_\theta=-u_xrsin\theta+u_yrcos\theta \\\\ &v_r=v_xcos\theta+v_ysin\theta \\\\ &v_\theta=-v_xrsin\theta+v_yrcos\theta \end{aligned}\right. \Rightarrow \left\{\begin{aligned} &u_x=u_rcos\theta-\dfrac{1}{r}u_\theta sin\theta \\\\ &u_y=u_r sin\theta+\dfrac{1}{r}u_\theta cos\theta \\\\ &v_x=v_rcos\theta-\dfrac{1}{r}v_\theta sin\theta \\\\ &v_y=v_r sin\theta+\dfrac{1}{r}v_\theta cos\theta \end{aligned}\right.
⎩
⎨
⎧x=r cosθy=r sinθ⇒⎩
⎨
⎧ur=uxcosθ+uysinθuθ=−uxrsinθ+uyrcosθvr=vxcosθ+vysinθvθ=−vxrsinθ+vyrcosθ⇒⎩
⎨
⎧ux=urcosθ−r1uθsinθuy=ursinθ+r1uθcosθvx=vrcosθ−r1vθsinθvy=vrsinθ+r1vθcosθ
进一步根据直角坐标系下的Cauchy-Riemann方程可求解得到:
r
u
r
=
v
θ
,
r
v
r
=
−
v
θ
ru_r=v_\theta,\quad rv_r=-v_\theta
rur=vθ,rvr=−vθ
证毕.
复函数
w
=
f
(
z
)
w=f(z)
w=f(z)也可视为实变元
x
,
y
x,y
x,y的函数,若它可微,则有:
d
f
=
(
u
x
+
i
v
x
)
(
d
x
+
i
d
y
)
=
(
u
x
+
i
v
x
)
d
x
+
(
−
v
x
+
i
u
x
)
d
y
=
(
u
x
+
i
v
x
)
d
x
+
(
u
y
+
i
v
y
)
d
y
=
f
x
d
x
+
f
y
d
y
\begin{aligned} df&=(u_x+iv_x)(dx+idy)\\\\ &=(u_x+iv_x)dx+(-v_x+iu_x)dy\\\\ &=(u_x+iv_x)dx+(u_y+iv_y)dy\\\\ &=f_xdx+f_ydy \end{aligned}
df=(ux+ivx)(dx+idy)=(ux+ivx)dx+(−vx+iux)dy=(ux+ivx)dx+(uy+ivy)dy=fxdx+fydy
考虑一个特殊的可微的复变函数
w
(
z
)
=
z
=
x
+
i
y
w(z)=z=x+iy
w(z)=z=x+iy,此时有:
d
z
=
d
x
+
i
d
y
⇒
d
z
‾
=
d
x
−
i
d
y
dz=dx+idy\Rightarrow\overline{dz}=dx-idy
dz=dx+idy⇒dz=dx−idy
故有
d
x
=
d
z
+
d
z
‾
2
,
d
y
=
d
z
−
d
z
‾
2
i
=
−
i
d
z
+
i
d
z
‾
2
dx=\dfrac{dz+\overline{dz}}{2},\quad dy=\dfrac{dz-\overline{dz}}{2i}=\dfrac{-idz+i\overline{dz}}{2}
dx=2dz+dz,dy=2idz−dz=2−idz+idz
进一步可得:
d
f
=
f
x
d
x
+
f
y
d
y
=
1
2
(
f
x
−
i
f
y
)
d
z
+
1
2
(
f
x
+
i
f
y
)
d
z
‾
≜
∂
f
∂
z
d
z
+
∂
f
∂
z
ˉ
d
z
‾
\begin{aligned} df&=f_xdx+f_ydy \\\\ &=\dfrac{1}{2}(f_x-if_y)dz+\dfrac{1}{2}(f_x+if_y)\overline{dz}\\\\ &\triangleq\dfrac{\partial f}{\partial z}dz+\dfrac{\partial f}{\partial \bar{z}}\overline{dz} \end{aligned}
df=fxdx+fydy=21(fx−ify)dz+21(fx+ify)dz≜∂z∂fdz+∂zˉ∂fdz
上式形式上定义了微分算子:
{
∂
∂
z
≜
1
2
(
∂
∂
x
−
i
∂
∂
y
)
∂
∂
z
ˉ
≜
1
2
(
∂
∂
x
+
i
∂
∂
y
)
\left\{\begin{aligned} \dfrac{\partial }{\partial z}\triangleq \dfrac{1}{2}(\dfrac{\partial }{\partial x}-i\dfrac{\partial }{\partial y}) \\\\ \dfrac{\partial }{\partial\bar z}\triangleq \dfrac{1}{2}(\dfrac{\partial }{\partial x}+i\dfrac{\partial }{\partial y}) \end{aligned}\right.
⎩
⎨
⎧∂z∂≜21(∂x∂−i∂y∂)∂zˉ∂≜21(∂x∂+i∂y∂)
显然它们满足线性性质与Leibniz法则,即
∂
(
a
1
w
1
+
a
2
w
2
)
∂
z
=
1
2
[
∂
(
a
1
w
1
+
a
2
w
2
)
∂
x
−
i
∂
(
a
1
w
1
+
a
2
w
2
)
∂
y
]
=
a
1
2
(
∂
w
1
∂
x
−
i
∂
w
1
∂
y
)
+
a
2
2
(
∂
w
2
∂
x
−
i
∂
w
2
∂
y
)
=
a
1
∂
w
1
∂
z
+
a
2
∂
w
2
∂
z
∂
(
a
1
w
1
+
a
2
w
2
)
∂
z
ˉ
=
1
2
[
∂
(
a
1
w
1
+
a
2
w
2
)
∂
x
+
i
∂
(
a
1
w
1
+
a
2
w
2
)
∂
y
]
=
a
1
2
(
∂
w
1
∂
x
+
i
∂
w
1
∂
y
)
+
a
2
2
(
∂
w
2
∂
x
+
i
∂
w
2
∂
y
)
=
a
1
∂
w
1
∂
z
ˉ
+
a
2
∂
w
2
∂
z
ˉ
∂
(
w
1
w
2
)
∂
z
=
1
2
[
∂
(
w
1
w
2
)
∂
x
−
i
∂
(
w
1
w
2
)
∂
y
]
=
1
2
[
(
w
1
∂
w
2
∂
x
+
∂
w
1
∂
x
w
2
)
−
i
(
w
1
∂
w
2
∂
x
+
∂
w
1
∂
x
w
2
)
]
=
∂
w
1
∂
z
w
2
+
w
1
∂
w
2
∂
z
∂
(
w
1
w
2
)
∂
z
ˉ
=
1
2
[
∂
(
w
1
w
2
)
∂
x
+
i
∂
(
w
1
w
2
)
∂
y
]
=
1
2
[
(
w
1
∂
w
2
∂
x
+
∂
w
1
∂
x
w
2
)
+
i
(
w
1
∂
w
2
∂
x
+
∂
w
1
∂
x
w
2
)
]
=
∂
w
1
∂
z
ˉ
w
2
+
w
1
∂
w
2
∂
z
ˉ
\begin{aligned} &\dfrac{\partial(a_1w_1+a_2w_2) }{\partial z} = \dfrac{1}{2}\left[\dfrac{\partial (a_1w_1+a_2w_2)}{\partial x}-i\dfrac{\partial (a_1w_1+a_2w_2) }{\partial y}\right] =\dfrac{a_1}{2}(\dfrac{\partial w_1}{\partial x}-i\dfrac{\partial w_1}{\partial y})+\dfrac{a_2}{2}(\dfrac{\partial w_2}{\partial x}-i\dfrac{\partial w_2}{\partial y}) =a_1\dfrac{\partial w_1}{\partial z}+a_2\dfrac{\partial w_2}{\partial z} \\\\ &\dfrac{\partial(a_1w_1+a_2w_2) }{\partial \bar z} = \dfrac{1}{2}\left[\dfrac{\partial (a_1w_1+a_2w_2)}{\partial x}+i\dfrac{\partial (a_1w_1+a_2w_2) }{\partial y}\right] =\dfrac{a_1}{2}(\dfrac{\partial w_1}{\partial x}+i\dfrac{\partial w_1}{\partial y})+\dfrac{a_2}{2}(\dfrac{\partial w_2}{\partial x}+i\dfrac{\partial w_2}{\partial y}) =a_1\dfrac{\partial w_1}{\partial \bar z}+a_2\dfrac{\partial w_2}{\partial \bar z} \\\\ &\dfrac{\partial (w_1w_2)}{\partial z} = \dfrac{1}{2}\left[\dfrac{\partial (w_1w_2)}{\partial x}-i\dfrac{\partial(w_1w_2) }{\partial y}\right] = \dfrac{1}{2}\left[\left(w_1\dfrac{\partial w_2}{\partial x}+\dfrac{\partial w_1}{\partial x}w_2\right)-i\left(w_1\dfrac{\partial w_2}{\partial x}+\dfrac{\partial w_1}{\partial x}w_2\right)\right] =\dfrac{\partial w_1}{\partial z}w_2+w_1\dfrac{\partial w_2}{\partial z}\\\\ &\dfrac{\partial (w_1w_2)}{\partial\bar z} = \dfrac{1}{2}\left[\dfrac{\partial (w_1w_2)}{\partial x}+i\dfrac{\partial(w_1w_2) }{\partial y}\right] = \dfrac{1}{2}\left[\left(w_1\dfrac{\partial w_2}{\partial x}+\dfrac{\partial w_1}{\partial x}w_2\right)+i\left(w_1\dfrac{\partial w_2}{\partial x}+\dfrac{\partial w_1}{\partial x}w_2\right)\right] =\dfrac{\partial w_1}{\partial\bar z}w_2+w_1\dfrac{\partial w_2}{\partial\bar z} \end{aligned}
∂z∂(a1w1+a2w2)=21[∂x∂(a1w1+a2w2)−i∂y∂(a1w1+a2w2)]=2a1(∂x∂w1−i∂y∂w1)+2a2(∂x∂w2−i∂y∂w2)=a1∂z∂w1+a2∂z∂w2∂zˉ∂(a1w1+a2w2)=21[∂x∂(a1w1+a2w2)+i∂y∂(a1w1+a2w2)]=2a1(∂x∂w1+i∂y∂w1)+2a2(∂x∂w2+i∂y∂w2)=a1∂zˉ∂w1+a2∂zˉ∂w2∂z∂(w1w2)=21[∂x∂(w1w2)−i∂y∂(w1w2)]=21[(w1∂x∂w2+∂x∂w1w2)−i(w1∂x∂w2+∂x∂w1w2)]=∂z∂w1w2+w1∂z∂w2∂zˉ∂(w1w2)=21[∂x∂(w1w2)+i∂y∂(w1w2)]=21[(w1∂x∂w2+∂x∂w1w2)+i(w1∂x∂w2+∂x∂w1w2)]=∂zˉ∂w1w2+w1∂zˉ∂w2
且有:
∂
z
∂
z
=
1
2
(
∂
z
∂
x
−
i
∂
z
∂
y
)
=
1
,
∂
z
∂
z
ˉ
=
1
2
(
∂
z
∂
x
+
i
∂
z
∂
y
)
=
0
\dfrac{\partial z}{\partial z}= \dfrac{1}{2}(\dfrac{\partial z}{\partial x}-i\dfrac{\partial z}{\partial y})=1, \quad\dfrac{\partial z}{\partial\bar z}=\dfrac{1}{2}(\dfrac{\partial z}{\partial x}+i\dfrac{\partial z}{\partial y})=0
∂z∂z=21(∂x∂z−i∂y∂z)=1,∂zˉ∂z=21(∂x∂z+i∂y∂z)=0
需要指出的是:上述微分算子是在函数可微的背景下引入的,但形式上我们同样可以将其扩展地作用于不可微函数进行运算,如
w
(
z
)
=
z
ˉ
w(z)=\bar z
w(z)=zˉ
∂
z
ˉ
∂
z
=
1
2
(
∂
z
ˉ
∂
x
−
i
∂
z
ˉ
∂
y
)
=
0
,
∂
z
ˉ
∂
z
ˉ
=
1
2
(
∂
z
ˉ
∂
x
+
i
∂
z
ˉ
∂
y
)
=
1
\dfrac{\partial \bar z}{\partial z}= \dfrac{1}{2}(\dfrac{\partial \bar z}{\partial x}-i\dfrac{\partial \bar z}{\partial y})=0, \quad\dfrac{\partial \bar z}{\partial\bar z}=\dfrac{1}{2}(\dfrac{\partial \bar z}{\partial x}+i\dfrac{\partial \bar z}{\partial y})=1
∂z∂zˉ=21(∂x∂zˉ−i∂y∂zˉ)=0,∂zˉ∂zˉ=21(∂x∂zˉ+i∂y∂zˉ)=1
不过此时,我们便有必要对函数可导的条件进行重新讨论,即在上述微分算子可作用任意函数的前提下重新考虑可微条件。
定理:设
u
(
x
,
y
)
u(x,y)
u(x,y)与
v
(
x
,
y
)
v(x,y)
v(x,y)在
(
x
,
y
)
(x,y)
(x,y)有一阶连续偏导数,则
f
(
z
)
=
u
+
i
v
f(z)=u+iv
f(z)=u+iv在该点可微的充要条件为:
∂
f
∂
z
ˉ
=
0
\dfrac{\partial f}{\partial\bar z}=0
∂zˉ∂f=0
证明:只需要验证Cauchy-Riemann方程是否得到满足即可。由定义
∂
f
∂
z
ˉ
=
1
2
(
∂
f
∂
x
+
i
∂
f
∂
y
)
=
1
2
[
(
∂
u
∂
x
+
i
∂
v
∂
x
)
+
i
(
∂
u
∂
y
+
i
∂
v
∂
y
)
]
=
1
2
[
(
∂
u
∂
x
−
∂
v
∂
y
)
+
i
(
∂
u
∂
y
+
∂
v
∂
x
)
]
=
0
\begin{aligned} \dfrac{\partial f}{\partial\bar z}&=\dfrac{1}{2}(\dfrac{\partial f}{\partial x}+i\dfrac{\partial f}{\partial y})\\\\ &=\dfrac{1}{2}\left[(\dfrac{\partial u}{\partial x}+i\dfrac{\partial v}{\partial x})+i(\dfrac{\partial u}{\partial y}+i\dfrac{\partial v}{\partial y})\right]\\\\ &=\dfrac{1}{2}\left[(\dfrac{\partial u}{\partial x}-\dfrac{\partial v}{\partial y})+i(\dfrac{\partial u}{\partial y}+\dfrac{\partial v}{\partial x})\right]=0 \end{aligned}
∂zˉ∂f=21(∂x∂f+i∂y∂f)=21[(∂x∂u+i∂x∂v)+i(∂y∂u+i∂y∂v)]=21[(∂x∂u−∂y∂v)+i(∂y∂u+∂x∂v)]=0
则
u
x
=
v
y
,
u
y
=
−
v
x
u_x=v_y,\quad u_y=-v_x
ux=vy,uy=−vx
证毕.
Remark: 上述定理表明:若 w = f ( z ) w=f(z) w=f(z)为解析函数,则它必须与 z ˉ \bar z zˉ无关。
1.4. 相关结论
定理 :若复变函数 w = f ( z ) w=f(z) w=f(z)在区域D内解析,且满足如下条件之一,则该函数在区域D内为常函数:
1) f ′ ( z ) = 0 f'(z)=0 f′(z)=0;
2) R e ( f ) Re(f) Re(f) 或 I m ( f ) Im(f) Im(f)为常数;
3) ∣ f ∣ |f| ∣f∣ 为常数。
此外,若函数在区域D内存在零点,则它恒为零。
证明:1)由已知:
f
′
(
z
)
=
u
x
+
i
v
x
=
v
y
−
i
u
y
=
0
f'(z)=u_x+iv_x=v_y-iu_y=0
f′(z)=ux+ivx=vy−iuy=0
则
u
x
=
u
y
=
v
x
=
v
y
=
0
u_x=u_y=v_x=v_y=0
ux=uy=vx=vy=0
故
u
u
u、
v
v
v 都是常二元函数,即
f
(
z
)
=
c
o
n
s
t
f(z)=const
f(z)=const。
2)若
I
m
(
f
)
=
v
=
0
Im(f)=v=0
Im(f)=v=0,则
v
x
=
v
y
=
0
⇒
u
y
=
u
x
=
0
v_x=v_y=0\Rightarrow u_y=u_x=0
vx=vy=0⇒uy=ux=0
故
f
(
z
)
=
c
o
n
s
t
f(z)=const
f(z)=const,其实部为常数时可同理得证。
3)
∣
f
∣
=
u
2
+
v
2
=
c
o
n
s
t
⇒
{
u
u
x
+
v
v
x
=
0
=
u
v
y
+
v
v
x
u
u
y
+
v
v
y
=
0
=
−
u
v
x
+
v
v
y
⇒
{
(
u
2
+
v
2
)
v
y
=
0
(
u
2
+
v
2
)
v
x
=
0
|f|=u^2+v^2=const\Rightarrow \left\{\begin{aligned} &uu_x+vv_x=0=uv_y+vv_x \\\\ &uu_y+vv_y=0=-uv_x+vv_y \end{aligned}\right. \Rightarrow \left\{\begin{aligned} &(u^2+v^2)v_y=0 \\\\ &(u^2+v^2)v_x=0 \end{aligned}\right.
∣f∣=u2+v2=const⇒⎩
⎨
⎧uux+vvx=0=uvy+vvxuuy+vvy=0=−uvx+vvy⇒⎩
⎨
⎧(u2+v2)vy=0(u2+v2)vx=0
若
u
2
+
v
2
=
0
u^2+v^2=0
u2+v2=0 ,则
u
=
0
,
v
=
0
⇒
f
(
z
)
=
0
u=0,v=0\Rightarrow f(z)=0
u=0,v=0⇒f(z)=0;
若 u 2 + v 2 ≠ 0 u^2+v^2\ne0 u2+v2=0 ,则 v x = v y = 0 ⇒ v ( x , y ) = c o n s t v_x=v_y=0\Rightarrow v(x,y)=const vx=vy=0⇒v(x,y)=const,故其虚部为常数,进一步可知函数本身也为常数。
另外,由函数的连续性可知如函数在区域D内存在零点,则它恒为零。(证毕)
定理 :若复变函数
f
(
z
)
=
u
(
x
,
y
)
+
i
v
(
x
,
y
)
f(z)=u(x,y)+iv(x,y)
f(z)=u(x,y)+iv(x,y) 在区域D内解析,且
f
′
(
z
)
≠
0
(
z
∈
D
)
f'(z)\ne0(z\in D)
f′(z)=0(z∈D),则
u
(
x
,
y
)
=
c
1
,
v
(
x
,
y
)
=
c
2
u(x,y)=c_1,~v(x,y)=c_2
u(x,y)=c1, v(x,y)=c2是
D
D
D 内的两组正交曲线族。
证明:由于 f ′ ( z ) ≠ 0 ( z ∈ D ) f'(z)\ne0(z\in D) f′(z)=0(z∈D) ,故 u x = v y u_x=v_y ux=vy, v x = − u y v_x=-u_y vx=−uy 并不全为零。
1)若某处二者均不为零,曲线
u
(
x
,
y
)
=
c
1
u(x,y)=c_1
u(x,y)=c1 的斜率为:
d
u
=
u
x
d
x
+
u
y
d
y
=
0
⇒
k
1
=
d
y
d
x
=
−
u
x
u
y
du=u_xdx+u_ydy=0\Rightarrow k_1=\dfrac{dy}{dx}=-\dfrac{u_x}{u_y}
du=uxdx+uydy=0⇒k1=dxdy=−uyux
曲线
v
(
x
,
y
)
=
c
2
v(x,y)=c_2
v(x,y)=c2 的斜率为:
d
v
=
v
x
d
x
+
v
y
d
y
=
0
⇒
k
2
=
d
y
d
x
=
−
v
x
v
y
=
u
y
u
x
dv=v_xdx+v_ydy=0\Rightarrow k_2=\dfrac{dy}{dx}=-\dfrac{v_x}{v_y}=\dfrac{u_y}{u_x}
dv=vxdx+vydy=0⇒k2=dxdy=−vyvx=uxuy
故,
k
1
k
2
=
−
1
k_1k_2=-1
k1k2=−1
说明,二者在该处正交。
2)若 u x = v y = 0 , v x = − u y ≠ 0 u_x=v_y=0,v_x=-u_y\ne0 ux=vy=0,vx=−uy=0 或 u x = v y ≠ 0 , v x = − u y = 0 u_x=v_y\ne0,v_x=-u_y=0 ux=vy=0,vx=−uy=0,此时过二者交点的切线一条水平一条竖直,仍正交。(证毕)
1.5. 解析函数的构造
若已知实部(或虚部)函数 u ( x , y ) u(x,y) u(x,y) 则可根据 Cauthy-Riemann 方程得到相应的虚部(或实部)函数 v ( x , y ) v(x,y) v(x,y) ,其中将包含一个待定常数,从而构造出区域 D 中的解析函数 u ( x , y ) + i v ( x , y ) u(x,y)+iv(x,y) u(x,y)+iv(x,y)。
方法一:偏积分法
根据 Cauthy-Riemann 方程:
∂
v
∂
x
=
−
∂
u
∂
y
⇒
v
=
−
∫
∂
u
∂
y
d
x
+
C
(
y
)
(
∗
)
\dfrac{\partial v}{\partial x}=-\dfrac{\partial u}{\partial y}\Rightarrow v=-\int\dfrac{\partial u}{\partial y}dx+C(y)\qquad(*)
∂x∂v=−∂y∂u⇒v=−∫∂y∂udx+C(y)(∗)
其中,
C
(
y
)
C(y)
C(y) 为
y
y
y 的待定函数。又
∂
u
∂
x
=
∂
v
∂
y
=
−
∂
∂
y
(
∫
∂
u
∂
y
d
x
)
+
d
C
(
y
)
d
y
\dfrac{\partial u}{\partial x}=\dfrac{\partial v}{\partial y}=-\dfrac{\partial }{\partial y}\left(\int\dfrac{\partial u}{\partial y}dx\right)+\dfrac{dC(y)}{dy}
∂x∂u=∂y∂v=−∂y∂(∫∂y∂udx)+dydC(y)
上式给出了关于
C
(
y
)
C(y)
C(y) 的常微分方程,求解出
C
(
y
)
C(y)
C(y) 后回代至
(
∗
)
(*)
(∗) 便可得到
v
v
v (含有一个待定常数)。
方法二:曲线积分法
由Cauthy-Riemann 方程:
v
(
x
,
y
)
=
∫
(
x
0
,
y
0
)
(
x
,
y
)
d
v
+
v
(
x
0
,
y
0
)
=
[
∫
(
x
0
,
y
0
)
(
x
,
y
)
v
x
d
x
+
v
y
d
y
]
+
v
(
x
0
,
y
0
)
=
[
∫
(
x
0
,
y
0
)
(
x
,
y
)
−
u
y
d
x
+
u
x
d
y
]
+
v
(
x
0
,
y
0
)
v(x,y)=\int_{(x_0,y_0)}^{(x,y)} dv+v(x_0,y_0)=\left[\int_{(x_0,y_0)}^{(x,y)} v_xdx+v_ydy\right]+v(x_0,y_0)=\left[\int_{(x_0,y_0)}^{(x,y)} -u_ydx+u_xdy\right]+v(x_0,y_0)
v(x,y)=∫(x0,y0)(x,y)dv+v(x0,y0)=[∫(x0,y0)(x,y)vxdx+vydy]+v(x0,y0)=[∫(x0,y0)(x,y)−uydx+uxdy]+v(x0,y0)
其中,
v
(
x
0
,
y
0
)
v(x_0,y_0)
v(x0,y0) 为待定常数。由 Cauthy-Riemann 方程可推知:
−
u
y
y
=
u
x
x
-u_{yy}=u_{xx}
−uyy=uxx。那么,当区域
D
D
D 为单连通区域,上述第二类曲线积分与路径无关,可选择特殊路径进行求解。如区域
D
D
D 为非单连通区域,则上述积分可能确定一个多值函数。
2. 解析函数与调和函数
2.1. 调和函数与共轭调和函数
定义:若
n
n
n 元函数
φ
\varphi
φ 在区域
D
⊂
R
D\subset\mathbb R
D⊂R 内有二阶连续偏导数,且满足 Laplace 方程:
Δ
φ
=
∇
2
φ
=
∂
2
φ
∂
x
1
2
+
∂
2
φ
∂
x
2
2
+
⋯
+
∂
2
φ
∂
x
n
2
=
0
\Delta\varphi=\nabla^2\varphi=\dfrac{\partial^2\varphi}{\partial x_1^2}+\dfrac{\partial^2\varphi}{\partial x_2^2}+\cdots+\dfrac{\partial^2\varphi}{\partial x_n^2}=0
Δφ=∇2φ=∂x12∂2φ+∂x22∂2φ+⋯+∂xn2∂2φ=0
则称
φ
\varphi
φ 为区域
D
D
D 内的调和函数(Harmonic function)。
定义:设二元函数
u
(
x
,
y
)
u(x,y)
u(x,y) 及
v
(
x
,
y
)
v(x,y)
v(x,y) 在区域
D
D
D 内调和,且满足 Cauthy-Riemann 方程:
∂
u
∂
x
=
∂
v
∂
y
,
∂
u
∂
y
=
−
∂
v
∂
x
\dfrac{\partial u}{\partial x}=\dfrac{\partial v}{\partial y},\quad\dfrac{\partial u}{\partial y}=-\dfrac{\partial v}{\partial x}
∂x∂u=∂y∂v,∂y∂u=−∂x∂v
则称
v
v
v 为
u
u
u 在区域
D
D
D 内的共轭调和函数。
2.2. 解析函数与调和函数的关系
定理:复变函数 f ( z ) = u ( x , y ) + i v ( x , y ) f(z)=u(x,y)+iv(x,y) f(z)=u(x,y)+iv(x,y) 在区域 D D D 内解析的充要条件为在区域 D D D 内 v ( x , y ) v(x,y) v(x,y) 为 u ( x , y ) u(x,y) u(x,y) 的共轭调和函数。
证明:( ⇐ \Leftarrow ⇐)若在区域 D D D 内 v ( x , y ) v(x,y) v(x,y) 为 u ( x , y ) u(x,y) u(x,y) 的共轭调和函数,则在区域 D D D 内 v ( x , y ) v(x,y) v(x,y) 和 u ( x , y ) u(x,y) u(x,y) 可微且二者满足 Cauthy-Riemann 方程,故复变函数 f ( z ) = u ( x , y ) + i v ( x , y ) f(z)=u(x,y)+iv(x,y) f(z)=u(x,y)+iv(x,y) 解析。
(
⇒
\Rightarrow
⇒) 若
f
(
z
)
f(z)
f(z) 在区域D内解析,则
v
(
x
,
y
)
v(x,y)
v(x,y) 和
u
(
x
,
y
)
u(x,y)
u(x,y) 满足 Cauthy-Riemann 方程:
∂
u
∂
x
=
∂
v
∂
y
,
∂
u
∂
y
=
−
∂
v
∂
x
\dfrac{\partial u}{\partial x}=\dfrac{\partial v}{\partial y},\quad\dfrac{\partial u}{\partial y}=-\dfrac{\partial v}{\partial x}
∂x∂u=∂y∂v,∂y∂u=−∂x∂v
需要进一步说明
u
(
x
,
y
)
u(x,y)
u(x,y) 及
v
(
x
,
y
)
v(x,y)
v(x,y) 在区域
D
D
D 内调和。由于
f
(
z
)
f(z)
f(z)解析时,
v
(
x
,
y
)
v(x,y)
v(x,y) 和
u
(
x
,
y
)
u(x,y)
u(x,y)有任意阶连续偏导数。对 Cauthy-Riemann 方程两侧分别求导得到:
{
u
x
x
=
v
x
y
,
u
y
y
=
−
v
x
y
⇒
u
x
x
+
u
y
y
=
Δ
u
=
0
u
x
y
=
v
y
y
,
u
x
y
=
−
v
x
x
⇒
v
x
x
+
v
y
y
=
Δ
v
=
0
\begin{cases} u_{xx}=v_{xy},\quad u_{yy}=-v_{xy}\Rightarrow u_{xx}+u_{yy}=\Delta u=0 \\\\ u_{xy}=v_{yy},\quad u_{xy}=-v_{xx}\Rightarrow v_{xx}+v_{yy}=\Delta v=0 \end{cases}
⎩
⎨
⎧uxx=vxy,uyy=−vxy⇒uxx+uyy=Δu=0uxy=vyy,uxy=−vxx⇒vxx+vyy=Δv=0
故
v
v
v 是
u
u
u 的共轭调和函数。
推论:任意一个二元调和函数的任意阶偏导数均为调和函数。 这是因为对任意一个调和函数可作为解析函数的实部,而虚部可根据 Cauthy-Riemann 方程构造,又因为解析函数的任意阶导数均为解析函数,根据上述定理可得推论成立。