索引
闭区间上的连续函数
闭区间上的连续函数的性质
定理5.1 有界定理
函数 f ( x ) f\left ( x \right ) f(x)在闭区间 [ a , b ] \left [ a,b \right ] [a,b]上连续,则函数 f ( x ) f\left ( x \right ) f(x)在闭区间 [ a , b ] \left [ a,b \right ] [a,b]上有界。
反证法,设函数
f
(
x
)
f\left ( x \right )
f(x)在闭区间
[
a
,
b
]
\left [ a,b \right ]
[a,b]上无界。
将区间
[
a
,
b
]
(
i
∈
N
)
\left [ a,b \right ](i \in \mathbb{N} )
[a,b](i∈N)分为
[
a
,
a
+
b
2
]
\left [ a,\frac{a+b }{2} \right ]
[a,2a+b]和
[
a
+
b
2
,
b
]
\left [ \frac{a+b }{2},b \right ]
[2a+b,b],则
f
(
x
)
f\left ( x \right )
f(x)至少在其中一个区间上无界,选择无界区间并标记为
[
a
1
,
b
1
]
\left [ a_{1},b_{1} \right ]
[a1,b1],
以此类推,循环以下流程:
- 将区间 [ a n , b n ] ( n ∈ N + ) \left [ a_{n},b_{n} \right ](n \in \mathbb{N}^{+} ) [an,bn](n∈N+)二等分为 [ a n , a n + b n 2 ] \left [ a_{n},\frac{a_{n}+b_{n} }{2} \right ] [an,2an+bn]和 [ a n + b n 2 , b n ] \left [ \frac{a_{n}+b_{n} }{2},b_{n} \right ] [2an+bn,bn];
- 则 f ( x ) f\left ( x \right ) f(x)至少在其中一个区间上无界,选择无界区间并标记为 [ a n + 1 , b n + 1 ] \left [ a_{n+1},b_{n+1} \right ] [an+1,bn+1]。
由闭区间套定理,闭区间套
{
[
a
n
,
b
n
]
}
\left \{ \left [ a_{n},b_{n} \right ] \right \}
{[an,bn]}中有唯一的
ξ
∈
R
\xi \in \mathbb{R}
ξ∈R存在于每个闭区间
[
a
n
,
b
n
]
\left [ a_{n},b_{n} \right ]
[an,bn],
因为函数
f
(
x
)
f \left ( x \right )
f(x)在
ξ
∈
[
a
,
b
]
\xi \in \left [ a,b \right ]
ξ∈[a,b]处连续,所以
lim
x
→
ξ
f
(
x
)
\lim_{x \to \xi } f\left ( x \right )
limx→ξf(x)存在,
由函数极限的局部有界性,
f
(
x
)
f\left ( x \right )
f(x)在区间
[
a
,
b
]
∪
U
(
ξ
,
δ
)
\left [ a,b \right ]\cup U \left ( \xi,\delta \right )
[a,b]∪U(ξ,δ)内有界,
令
n
→
∞
n\to \infty
n→∞,则当
n
n
n充分大时,
[
a
n
,
b
n
]
\left [ a_{n} ,b_{n} \right ]
[an,bn]以及其后构造的闭区间都包含于
[
a
,
b
]
∪
U
(
ξ
,
δ
)
\left [ a,b \right ]\cup U \left ( \xi,\delta \right )
[a,b]∪U(ξ,δ)内,
f
(
x
)
f\left ( x \right )
f(x)在区间
[
a
,
b
]
∪
U
(
ξ
,
δ
)
\left [ a,b \right ]\cup U \left ( \xi,\delta \right )
[a,b]∪U(ξ,δ)的子区间
[
a
n
,
b
n
]
\left [ a_{n} ,b_{n} \right ]
[an,bn]内无界,这与函数
f
(
x
)
f \left ( x \right )
f(x)在
[
a
,
b
]
∪
U
(
ξ
,
δ
)
\left [ a,b \right ]\cup U \left ( \xi,\delta \right )
[a,b]∪U(ξ,δ)内有界矛盾。
定理5.2 最值定理
f ( x ) f\left ( x \right ) f(x)在 [ a , b ] \left [ a,b \right ] [a,b]上连续,则 f ( x ) f\left ( x \right ) f(x)在 [ a , b ] \left [ a,b \right ] [a,b]上必有最大值 M M M和最小值 m m m,也就是 ∃ ξ \exists \xi ∃ξ, η ∈ [ a , b ] \eta \in \left [ a,b \right ] η∈[a,b]: ∀ x ∈ [ a , b ] \forall x\in \left [ a,b \right ] ∀x∈[a,b], f ( ξ ) = m = min f ( x ) f\left ( \xi \right )=m= \min f\left ( x \right ) f(ξ)=m=minf(x), f ( η ) = M = max f ( x ) f\left ( \eta \right )=M=\max f\left ( x \right ) f(η)=M=maxf(x)。
设
f
(
x
)
f\left ( x \right )
f(x)的值域为
R
f
R_{f}
Rf,
R
f
R_{f}
Rf中任取数列
x
n
x_{n}
xn,
因为
f
(
x
)
f\left ( x \right )
f(x)在
[
a
,
b
]
\left [ a,b \right ]
[a,b]上连续,所以根据有界定理,值域
R
f
R_{f}
Rf有界,
又根据确界存在原理,
R
f
R_{f}
Rf有下确界
α
=
inf
R
f
\alpha =\inf R_{f}
α=infRf,上确界
β
=
sup
R
f
\beta =\sup R_{f}
β=supRf,
以上确界
β
=
sup
R
f
\beta =\sup R_{f}
β=supRf为典型,
取无穷小量数量数列
ε
n
>
0
\varepsilon_{n} > 0
εn>0,根据上确界的定义,对所有
n
n
n,均有
β
−
ε
n
<
f
(
x
n
)
≤
β
\beta -\varepsilon_{n} <f\left (x_{n}\right )\le \beta
β−εn<f(xn)≤β,
引用Bolzano-Weierstrass定理,
x
n
∈
[
a
,
b
]
x_{n} \in \left [ a,b \right ]
xn∈[a,b]必有收敛子列
{
x
n
k
}
\left \{ x_{n_{k} } \right \}
{xnk},
ε
n
\varepsilon_{n}
εn也必有收敛于0的子列
{
ε
n
k
}
>
0
\left \{ \varepsilon _{n_{k} } \right \}>0
{εnk}>0,满足
β
−
ε
n
k
<
f
(
x
n
k
)
≤
β
\beta -\varepsilon _{n_{k} }< f\left ( x_{n_{k}}\right )\le \beta
β−εnk<f(xnk)≤β,
令
k
→
∞
k\to \infty
k→∞,设
lim
k
→
∞
x
n
k
=
ξ
∈
[
a
,
b
]
\lim _{k\to \infty } x_{n_{k}} = \xi \in \left [ a,b \right ]
limk→∞xnk=ξ∈[a,b],
根据数列极限的局部保不等号性(数列极限保序性的逆命题),
β
−
0
=
β
≤
lim
x
n
k
→
ξ
f
(
x
n
k
)
≤
β
⇔
lim
x
n
k
→
ξ
f
(
x
n
k
)
=
β
\beta -0=\beta \le \lim _{x_{n_{k}} \to \xi } f\left (x_{n_{k}}\right ) \le \beta \Leftrightarrow \lim _{x_{n_{k}} \to \xi } f\left (x_{n_{k}}\right ) = \beta
β−0=β≤limxnk→ξf(xnk)≤β⇔limxnk→ξf(xnk)=β,
根据连续函数定义,
lim
x
n
k
→
ξ
f
(
x
n
k
)
=
f
(
ξ
)
=
β
\lim _{x_{n_{k}} \to \xi } f\left ( x_{n_{k}}\right ) =f\left ( \xi \right )=\beta
limxnk→ξf(xnk)=f(ξ)=β,
由此,
∃
ξ
∈
[
a
,
b
]
\exists \xi \in \left [ a,b \right ]
∃ξ∈[a,b]:
∀
x
∈
[
a
,
b
]
\forall x\in \left [ a,b \right ]
∀x∈[a,b],
f
(
x
)
≥
f
(
ξ
)
f\left ( x \right )\ge f\left ( \xi \right )
f(x)≥f(ξ),
同理可得
∃
η
∈
[
a
,
b
]
\exists \eta \in \left [ a,b \right ]
∃η∈[a,b]:
∀
x
∈
[
a
,
b
]
\forall x\in \left [ a,b \right ]
∀x∈[a,b],
f
(
x
)
≤
f
(
η
)
f\left ( x \right )\le f\left ( \eta \right )
f(x)≤f(η)。
定理5.3 零点存在定理
f ( x ) f\left ( x \right ) f(x)在 [ a , b ] \left [ a,b \right ] [a,b]上连续,且 f ( a ) f ( b ) < 0 f\left ( a \right )f\left ( b \right )<0 f(a)f(b)<0,则 ∃ ξ ∈ ( a , b ) \exists \xi \in \left ( a,b \right ) ∃ξ∈(a,b): f ( ξ ) = 0 f\left ( \xi \right )=0 f(ξ)=0。
f
(
x
)
f\left ( x \right )
f(x)在
[
a
0
,
b
0
]
\left [ a_{0},b_{0} \right ]
[a0,b0]上连续,不妨设
f
(
a
0
)
>
0
f\left ( a_{0} \right ) >0
f(a0)>0,
f
(
b
0
)
<
0
f\left ( b_{0} \right ) <0
f(b0)<0,
循环以下流程:
- 将区间 [ a n , b n ] ( i ∈ N + ) \left [ a_{n},b_{n} \right ](i \in \mathbb{N}^{+} ) [an,bn](i∈N+)二等分为 [ a n , a n + b n 2 ] \left [ a_{n},\frac{a_{n}+b_{n} }{2} \right ] [an,2an+bn]和 [ a n + b n 2 , b n ] \left [ \frac{a_{n}+b_{n} }{2},b_{n} \right ] [2an+bn,bn];
- 进行如下判定:
- 若 a n + b n 2 < 0 \frac{a_{n}+b_{n} }{2}<0 2an+bn<0,选取 [ a n , a n + b n 2 ] \left [ a_{n},\frac{a_{n}+b_{n} }{2} \right ] [an,2an+bn]作为下一次循环的 [ a n + 1 , b n + 1 ] \left [ a_{n+1},b_{n+1} \right ] [an+1,bn+1];
- 若 a n + b n 2 > 0 \frac{a_{n}+b_{n} }{2}>0 2an+bn>0,选取 [ a n + b n 2 , b n ] \left [ \frac{a_{n}+b_{n} }{2},b_{n} \right ] [2an+bn,bn]作为下一次循环的 [ a n + 1 , b n + 1 ] \left [ a_{n+1},b_{n+1} \right ] [an+1,bn+1];
- 若 a n + b n 2 = 0 \frac{a_{n}+b_{n} }{2}=0 2an+bn=0, ξ = a n + b n 2 \xi=\frac{a_{n}+b_{n} }{2} ξ=2an+bn即 f ( x ) f\left ( x \right ) f(x)的零点。
- 循环上述流程。
可由此构造闭区间套
{
[
a
n
,
b
n
]
∣
f
(
a
n
)
>
0
>
f
(
b
n
)
}
\left \{ \left [ a_{n},b_{n} \right ]|f\left ( a_{n} \right )>0>f\left ( b_{n} \right ) \right \}
{[an,bn]∣f(an)>0>f(bn)}。
由闭区间套定理,
∃
ξ
:
∀
n
,
ξ
∈
⋂
n
=
0
∞
[
a
n
,
b
n
]
\exists \xi:\forall n,\xi \in \bigcap_{n=0}^{\infty }\left [ a_{n},b_{n} \right ]
∃ξ:∀n,ξ∈⋂n=0∞[an,bn],且
lim
n
→
∞
a
n
=
lim
n
→
∞
b
n
=
ξ
\lim_{n \to \infty} a_{n}=\lim_{n \to \infty} b_{n}=\xi
limn→∞an=limn→∞bn=ξ,
由Heine定理,因为
lim
x
→
ξ
f
(
x
)
=
f
(
ξ
)
\lim _{x\to \xi }f\left ( x \right ) =f\left ( \xi \right )
limx→ξf(x)=f(ξ),所以对任意满足
x
n
≠
ξ
x_{n} \ne \xi
xn=ξ,
lim
n
→
∞
x
n
=
ξ
\lim_{n \to \infty} x_{n}=\xi
limn→∞xn=ξ条件的数列
{
x
n
}
\left \{ x_{n} \right \}
{xn},都有
lim
n
→
∞
f
(
x
n
)
=
f
(
ξ
)
\lim_{n \to \infty}f\left ( x_{n} \right )=f\left ( \xi \right )
limn→∞f(xn)=f(ξ),
因为
f
(
x
)
f\left ( x \right )
f(x)在闭区间套
{
[
a
n
,
b
n
]
∣
f
(
a
n
)
>
0
>
f
(
b
n
)
}
\left \{ \left [ a_{n},b_{n} \right ]|f\left ( a_{n} \right )>0>f\left ( b_{n} \right ) \right \}
{[an,bn]∣f(an)>0>f(bn)}中的每个闭区间
[
a
n
,
b
n
]
\left [ a_{n},b_{n} \right ]
[an,bn]上连续,且所以当
n
n
n充分大时,
lim
n
→
∞
f
(
a
n
)
=
f
(
ξ
)
\lim_{n \to \infty } f\left ( a_{n} \right ) =f\left ( \xi \right )
limn→∞f(an)=f(ξ),
同理可得,
lim
n
→
∞
f
(
b
n
)
=
f
(
ξ
)
\lim_{n \to \infty } f\left ( b_{n} \right ) =f\left ( \xi \right )
limn→∞f(bn)=f(ξ),
根据函数极限的保号性,
0
≤
lim
n
→
∞
f
(
a
n
)
=
f
(
ξ
)
=
lim
n
→
∞
f
(
b
n
)
≤
0
⇔
f
(
ξ
)
=
0
0\le \lim_{n \to \infty} f\left ( a_{n} \right )=f\left ( \xi \right ) =\lim_{n \to \infty} f\left ( b_{n} \right ) \le 0\Leftrightarrow f\left ( \xi \right )=0
0≤limn→∞f(an)=f(ξ)=limn→∞f(bn)≤0⇔f(ξ)=0。
定理5.4 介值定理
f ( x ) f\left ( x \right ) f(x)在 [ a , b ] \left [ a,b \right ] [a,b]上连续,且 m ≤ f ( x ) ≤ M ( m , M ∈ R ) m\le f\left ( x \right ) \le M(m,M\in \mathbb{R} ) m≤f(x)≤M(m,M∈R),则 f ( x ) f\left ( x \right ) f(x)可取到最小值 m m m与最大值 M M M之间的任何值。
由最值定理,
∃
ξ
\exists \xi
∃ξ,
η
∈
[
a
,
b
]
\eta \in \left [ a,b \right ]
η∈[a,b]:
f
(
ξ
)
=
m
=
min
f
(
x
)
f\left ( \xi \right )=m=\min f\left ( x \right )
f(ξ)=m=minf(x),
f
(
η
)
=
M
=
max
f
(
x
)
f\left ( \eta \right )=M=\max f\left ( x \right )
f(η)=M=maxf(x),
不妨设
ξ
<
η
\xi <\eta
ξ<η,
任取
k
∈
(
m
,
M
)
k \in \left ( m ,M \right )
k∈(m,M),设
g
(
x
)
=
f
(
x
)
−
k
g\left ( x \right )=f\left ( x \right )-k
g(x)=f(x)−k,则有
g
(
ξ
)
<
0
g\left ( \xi \right )<0
g(ξ)<0,
g
(
η
)
>
0
g\left ( \eta \right )>0
g(η)>0,
由零点存在定理,
∃
γ
∈
(
ξ
,
η
)
\exists \gamma \in \left ( \xi ,\eta \right )
∃γ∈(ξ,η):
g
(
γ
)
=
0
⇔
f
(
γ
)
=
k
∈
(
m
,
M
)
g\left ( \gamma \right )=0\Leftrightarrow f\left ( \gamma \right ) =k\in \left ( m,M \right )
g(γ)=0⇔f(γ)=k∈(m,M)。
一致连续
定义5.1 一致连续
若函数 f ( x ) f\left ( x \right ) f(x)在区间 X X X上有定义,且 ∀ ε > 0 \forall \varepsilon >0 ∀ε>0, ∃ δ > 0 \exists \delta >0 ∃δ>0: ∀ x ′ \forall x^{\prime} ∀x′, x ′ ′ x^{\prime\prime } x′′, ∣ x ′ − x ′ ′ ∣ < δ \left | x^{\prime}-x^{\prime\prime } \right |<\delta ∣x′−x′′∣<δ, ∣ f ( x ′ ) − f ( x ′ ′ ) ∣ < ε \left | f\left ( x ^{\prime }\right )- f\left ( x^{\prime \prime } \right ) \right |<\varepsilon ∣f(x′)−f(x′′)∣<ε,则称函数 f ( x ) f\left ( x \right ) f(x)在区间 X X X上一致连续。
定理5.5
设 f ( x ) f\left ( x \right ) f(x)在区间 X X X上连续,则设 f ( x ) f\left ( x \right ) f(x)在区间 X X X上一致连续的充要条件为:对任意区间 X X X上的数列 { x n ′ } \left \{ x_{n} ^{\prime }\right \} {xn′}, { x n ′ ′ } \left \{ x_{n} ^{\prime \prime } \right \} {xn′′},有 lim n → ∞ ( x n ′ − x n ′ ′ ) = 0 ⇒ lim n → ∞ ( f ( x n ′ ) − f ( x n ′ ′ ) ) = 0 \lim _{n\to \infty }\left ( x_{n}^{\prime} -x_{n} ^{\prime \prime } \right )=0\Rightarrow \lim _{n\to \infty }\left (f \left ( x_{n} ^{\prime }\right ) -f\left ( x_{n} ^{\prime \prime } \right ) \right )=0 limn→∞(xn′−xn′′)=0⇒limn→∞(f(xn′)−f(xn′′))=0。
<1>必要性
根据数列极限定义,
∀
δ
>
0
\forall \delta >0
∀δ>0,
∃
N
\exists N
∃N:
∀
n
>
N
\forall n>N
∀n>N,
∣
x
n
′
−
x
n
′
′
∣
<
δ
\left | x_{n}^{\prime} -x_{n} ^{\prime \prime } \right |< \delta
∣xn′−xn′′∣<δ,
又根据函数一致连续定义,
∀
ε
>
0
\forall \varepsilon >0
∀ε>0,
∃
δ
>
0
(
∣
x
n
′
−
x
n
′
′
∣
<
δ
)
\exists \delta >0 \left ( \left |x_{n}^{\prime} -x_{n} ^{\prime \prime } \right | <\delta\right )
∃δ>0(∣xn′−xn′′∣<δ):
∣
f
(
x
n
′
)
−
f
(
x
n
′
′
)
∣
<
ε
\left | f \left ( x_{n} ^{\prime }\right ) -f\left ( x_{n} ^{\prime \prime } \right ) \right |<\varepsilon
∣f(xn′)−f(xn′′)∣<ε。
<2>充分性
构造逆否命题:
(
p
→
q
)
⇔
(
¬
q
→
¬
p
)
p
:
∀
{
x
n
′
}
,
{
x
n
′
′
}
∈
X
,
lim
n
→
∞
(
x
n
′
−
x
n
′
′
)
=
0
→
lim
n
→
∞
(
f
(
x
n
′
)
−
f
(
x
n
′
′
)
)
=
0
¬
p
:
∃
{
x
n
′
}
,
{
x
n
′
′
}
∈
X
:
lim
n
→
∞
(
x
n
′
−
x
n
′
′
)
=
0
↛
lim
n
→
∞
(
f
(
x
n
′
)
−
f
(
x
n
′
′
)
)
=
0
q
:
f
(
x
)
在区间
X
上一致连续
¬
q
:
f
(
x
)
在区间
X
上非一致连续
\begin{array}{l} &&\left ( p\to q \right ) \Leftrightarrow \left ( \neg q\to \neg p \right ) \\ p&:& \forall \left \{ x_{n} ^{\prime }\right \},\left \{ x_{n} ^{\prime \prime } \right \}\in X,\lim _{n\to \infty }\left ( x_{n}^{\prime} -x_{n} ^{\prime \prime } \right )=0 \to \lim _{n\to \infty }\left (f \left ( x_{n} ^{\prime }\right ) -f\left ( x_{n} ^{\prime \prime } \right ) \right )=0 \\ \neg p&:&\exists \left \{ x_{n} ^{\prime }\right \},\left \{ x_{n} ^{\prime \prime } \right \}\in X:\lim _{n\to \infty }\left ( x_{n} ^{\prime} -x_{n} ^{\prime \prime } \right )=0 \not \to \lim _{n\to \infty }\left (f \left ( x_{n} ^{\prime }\right ) -f\left ( x_{n} ^{\prime \prime } \right ) \right )=0\\ q&:&f\left ( x \right )在区间X上一致连续\\ \neg q&:&f\left ( x \right )在区间X上非一致连续\\ \end{array}
p¬pq¬q::::(p→q)⇔(¬q→¬p)∀{xn′},{xn′′}∈X,limn→∞(xn′−xn′′)=0→limn→∞(f(xn′)−f(xn′′))=0∃{xn′},{xn′′}∈X:limn→∞(xn′−xn′′)=0→limn→∞(f(xn′)−f(xn′′))=0f(x)在区间X上一致连续f(x)在区间X上非一致连续
而对逆否命题
¬
q
→
¬
p
\neg q\to \neg p
¬q→¬p可以通过构造无穷小量数列
δ
n
=
1
n
\delta _{n} =\frac{1}{n}
δn=n1证明。
定理5.6 Cantor定理
函数 f ( x ) f\left ( x \right ) f(x)在闭区间 [ a , b ] \left [ a,b \right ] [a,b]上连续,则函数 f ( x ) f\left ( x \right ) f(x)在闭区间 [ a , b ] \left [ a,b \right ] [a,b]上一致连续。
反证法,设函数
f
(
x
)
f\left ( x \right )
f(x)在闭区间
[
a
,
b
]
\left [ a,b \right ]
[a,b]上连续,但是函数
f
(
x
)
f\left ( x \right )
f(x)在闭区间
[
a
,
b
]
\left [ a,b \right ]
[a,b]上非一致连续。
根据一致连续函数的定义,
∃
x
n
′
,
x
n
′
′
∈
[
a
,
b
]
:
∀
ε
0
>
0
\exists x_{n}^{\prime} ,x_{n} ^{\prime \prime }\in \left [ a,b \right ]:\forall \varepsilon_{0} >0
∃xn′,xn′′∈[a,b]:∀ε0>0,
∣
x
n
′
−
x
n
′
′
∣
<
δ
=
1
n
\left | x_{n}^{\prime} -x_{n}^{ \prime \prime } \right |< \delta=\frac{1}{n}
∣xn′−xn′′∣<δ=n1,
∣
f
(
x
n
′
)
−
f
(
x
n
′
′
)
∣
≥
ε
0
\left | f\left ( x_{n}^{\prime } \right ) -f\left ( x_{n}^{\prime \prime } \right ) \right |\ge \varepsilon_{0}
∣f(xn′)−f(xn′′)∣≥ε0。
引用列紧性定理,有界数列
{
x
n
′
}
\left \{ x_{n} ^{\prime } \right \}
{xn′}必有收敛子列
{
x
n
k
′
}
\left \{ x^{\prime} _{n_{k} } \right \}
{xnk′},
设
lim
k
→
∞
x
n
k
′
=
ξ
∈
[
a
,
b
]
\lim_{k \to \infty}x_{n_{k} }^{\prime } =\xi \in \left [ a,b \right ]
limk→∞xnk′=ξ∈[a,b],
又由
∣
x
n
k
′
−
x
n
k
′
′
∣
<
δ
k
=
1
n
k
\left | x_{n_{k}}^{\prime} -x_{n_{k}}^{\prime \prime } \right |< \delta_{k}=\frac{1}{n_{k}}
xnk′−xnk′′
<δk=nk1可知
lim
k
→
∞
(
x
n
k
′
−
x
n
k
′
′
)
=
0
\lim _{k\to \infty } \left ( x_{n_{k}}^{\prime} -x_{n_{k}}^{\prime \prime } \right )=0
limk→∞(xnk′−xnk′′)=0,所以
lim
k
→
∞
x
n
k
′
′
=
ξ
\lim _{k\to \infty } x_{n_{k}}^{\prime \prime }=\xi
limk→∞xnk′′=ξ,
根据连续函数的定义,
lim
k
→
∞
f
(
x
n
k
′
)
=
lim
k
→
∞
f
(
x
n
k
′
′
)
=
f
(
ξ
)
\lim _{k\to \infty } f\left ( x_{n_{k}}^{\prime } \right )= \lim _{k\to \infty } f\left ( x_{n_{k}}^{\prime \prime } \right )=f\left ( \xi \right )
limk→∞f(xnk′)=limk→∞f(xnk′′)=f(ξ),推出
lim
k
→
∞
(
f
(
x
n
k
′
)
−
f
(
x
n
k
′
′
)
)
=
0
\lim _{k\to \infty } \left ( f\left ( x_{n_{k}}^{\prime } \right )- f\left ( x_{n_{k}}^{\prime \prime } \right ) \right ) =0
limk→∞(f(xnk′)−f(xnk′′))=0,找到了一个与假设矛盾的特例,所以函数
f
(
x
)
f\left ( x \right )
f(x)在闭区间
[
a
,
b
]
\left [ a,b \right ]
[a,b]上一致连续。
定理5.7
f ( x ) f\left ( x \right ) f(x)在开区间 ( a , b ) \left ( a,b \right ) (a,b)上连续,则 f ( x ) f\left ( x \right ) f(x)在开区间 ( a , b ) \left ( a,b \right ) (a,b)上一致连续的充要条件是 f ( a + ) f\left ( a+ \right ) f(a+), f ( b − ) f\left ( b- \right ) f(b−)存在。
<1>充分性
构造函数
f
~
(
x
)
=
{
A
=
f
(
a
+
)
a
f
(
x
)
(
a
,
b
)
B
=
f
(
b
−
)
b
\tilde{f}\left ( x \right )= \left\{\begin{matrix} A=f\left ( a+ \right ) & a \\ f\left ( x \right ) & \left ( a,b \right ) \\ B=f\left ( b- \right ) & b \end{matrix}\right.
f~(x)=⎩
⎨
⎧A=f(a+)f(x)B=f(b−)a(a,b)b
使得函数
f
~
(
x
)
\tilde{f}\left ( x \right )
f~(x)在闭区间
[
a
,
b
]
\left [ a,b \right ]
[a,b]上连续。
由Cantor定理,
f
~
(
x
)
\tilde{f}\left ( x \right )
f~(x)在闭区间
[
a
,
b
]
\left [ a,b \right ]
[a,b]上一致连续,推出
f
(
x
)
f\left ( x \right )
f(x)在开区间
(
a
,
b
)
\left ( a,b \right )
(a,b)上一致连续。
<2>必要性
不妨以开区间
(
a
,
b
)
\left ( a,b \right )
(a,b)的左端点
a
a
a处为例,
根据一致连续的定义,
∀
ε
>
0
\forall \varepsilon > 0
∀ε>0,
∃
δ
>
0
\exists \delta >0
∃δ>0:
∀
x
′
\forall x^{\prime }
∀x′,
x
′
′
x^{\prime \prime }
x′′,
∣
x
′
−
x
′
′
∣
<
δ
\left | x^{\prime}-x^{\prime\prime } \right |<\delta
∣x′−x′′∣<δ,
∣
f
(
x
′
)
−
f
(
x
′
′
)
∣
<
ε
\left | f \left ( x^{\prime } \right ) -f\left ( x^{\prime \prime } \right ) \right |<\varepsilon
∣f(x′)−f(x′′)∣<ε,
在开区间
(
a
,
b
)
\left ( a,b \right )
(a,b)中取数列
{
x
n
}
\left \{ x_{n} \right \}
{xn},并且
lim
n
→
∞
x
n
=
a
\lim_{n \to \infty } x_{n}=a
limn→∞xn=a,
x
n
≠
a
x_{n}\neq a
xn=a,则有
∃
δ
>
0
\exists \delta >0
∃δ>0:
∀
x
n
′
\forall x_{n} ^{\prime }
∀xn′,
x
n
′
′
x_{n}^{\prime \prime }
xn′′,
∣
x
n
′
−
x
n
′
′
∣
<
δ
\left | x_{n}^{\prime}-x_{n}^{\prime\prime } \right |<\delta
∣xn′−xn′′∣<δ,
∣
f
(
x
n
′
)
−
f
(
x
n
′
′
)
∣
<
ε
\left | f \left ( x_{n}^{\prime } \right ) -f \left ( x_{n}^{\prime \prime } \right ) \right |<\varepsilon
∣f(xn′)−f(xn′′)∣<ε,
根据Cauchy收敛准则,基本列
{
f
(
x
n
)
}
\left \{ f \left ( x_{n} \right ) \right \}
{f(xn)}在左端点
a
a
a的右邻域
(
a
,
a
+
δ
)
\left ( a,a+\delta \right )
(a,a+δ)内收敛,
引用Heine定理,
f
(
x
)
f\left ( x \right )
f(x)在开区间
(
a
,
b
)
\left ( a,b \right )
(a,b)的左端点
a
a
a的右邻域
(
a
,
a
+
δ
)
\left ( a,a+\delta \right )
(a,a+δ)内收敛,开区间
(
a
,
b
)
\left ( a,b \right )
(a,b)的左端点
a
a
a处的右极限
f
(
a
+
)
f\left ( a+ \right )
f(a+)存在。
同理
f
(
x
)
f \left ( x \right )
f(x)在开区间
(
a
,
b
)
\left ( a,b \right )
(a,b)的右端点
b
b
b处的左极限
f
(
b
−
)
f\left ( b- \right )
f(b−)存在。
837
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
