陶哲轩实分析11.1及习题-Analysis I 11.1

这一节开始讲Riemann积分了,整个Riemann积分这一章的处理是比较新的,用相对好理解的piecewise constant function来定义上积分和下积分,二者相等即为可积,和以往比较经典的达布和其实是一回事,但接受起来就更容易。第一节主要说分法partition。

Exercise 11.1.1. Prove Lemma 11.1.4.
Lemma 11.1.4. Let X be a subset of the real line. Then the following two statements are logically equivalent:
( a ) X is bounded and connected.
( b ) X is a bounded interval.

Solution: If X={\emptyset} then the statement is obvious, since X is bounded and connected and is an interval for trivial reasons.
Now suppose X\neq {\emptyset}, then (b) implies (a) is clear, by the definition of an interval. Now suppose (a) holds, then \exists M>0,X\subset [-M,M], since X is non-empty and bounded, we can say \sup X and \inf X both exist in \mathbf R. Now choose any x\in \mathbf R such that

\inf X<x<\sup X

By the definition of supremum and infimum, we can find a,b\in X, s.t.

\inf X\leq a<x<b\leq \sup X

Since X is connected, [a,b]\subset X, thus x\in X. So we have the relation (\inf X,\sup X )\subset X, also we can easily have X\subset [\inf X,\sup X ], then X is equal to (\inf X,\sup X ) plus possibly two points \inf X and \sup X, anyway X is an interval.

\blacksquare

Exercise 11.1.2. Prove Corollary 11.1.6.
Corollary 11.1.6. If I and J are bounded intervals, then the intersection I\cap J is also a bounded interval.

Solution: By Lemma 11.1.4, I and J are bounded and connected, we have M_1>0,M_2>0, s.t.

(I\subset [-M_1,M_1 ])\wedge (J\subset [-M_2,M_2 ])\implies I\cap J\subset [-\max (M_1,M_2 ),\max (M_1,M_2)]

Thus I\cap J is bounded.
Let \forall x,y\in I\cap J, then as x,y\in I, we have [x,y]\subseteq I, similarly we have [x,y]\subseteq J, so [x,y]\subseteq I\cap J, this means I\cap J is connected.
We now can conclude, using Lemma 11.1.4 again, that I\cap J is a bounded interval.

\blacksquare

Exercise 11.1.3. Let I be a bounded interval of the form I=(a,b) or I=[a,b) for some real numbers a<b. Let I_1,\dots,I_n be a partition of I. Prove that one of the intervals I_j in this partition is of the form I_j=(c,b) or I_j=[c,b) for some a\leq c\leq b.

Solution: First, if I_j is not of the form (c,b) or [c.b) for any a\leq c\leq b, then we can have some 0<{\varepsilon}_j<b-a s.t. b-{\varepsilon}\notin I_j, as I_j is an interval, it’s connected, thus we have

\sup I_j\leq b-{\varepsilon}_j<b

Now assume all intervals I_j,1\leq j\leq n is not of the form (c,b) or [c.b) for any a\leq c\leq b, then we have \sup I_j\leq b-{\varepsilon}_j<b,1\leq j\leq n. Then we can conclude

b-\dfrac{\min_{1\leq j\leq n} ({\varepsilon}_j)}{2}\notin I_i,\quad\forall 1\leq i\leq n

so we have b-\min_{1\leq j\leq n} ({\varepsilon}_j )/2\notin I, this is a contradiction since I_1,\dots ,I_n form a partition of I.
Thus there must exists at least one of I_j in the form of (c,b) or [c.b). By the definition of partition, such I_j can only appear once if I_1,\dots ,I_n form a partition of I.

\blacksquare

Exercise 11.1.4. Prove Lemma 11.1.18.
Lemma 11.1.18. Let I be a bounded interval, and let \mathbf P and \mathbf P' be two partitions of I. Then \mathbf P \# \mathbf P' is also a partition of I, and is both finer than \mathbf P and finer than \mathbf P'.

Solution:

\mathbf P\#\mathbf P' is a partition of I:
Choose \forall x\in I, then there is exactly one K\in \mathbf P and J\in \mathbf P' s.t. x\in K,x\in J, thus x\in K\cap J, and any other element belongs to \mathbf P\#\mathbf P' cannot contain x, otherwise K or J can’t be unique. This means \mathbf P\#\mathbf P' is a partition of I.

\mathbf P\#\mathbf P' is finer than \mathbf P or \mathbf P':
For every M\in \mathbf P\#\mathbf P', by definition we shall be able to find K\in \mathbf P and J\in \mathbf P' s.t.

M=K\cap J

This means M\subseteq K,K\in \mathbf P and M\subseteq J,J\in \mathbf P', thus we have \mathbf P\#\mathbf P' is finer than \mathbf P or \mathbf P'.

\blacksquare

陶哲轩实分析10.5及习题-Analysis I 10.5

洛必达法则,计算极为好使,但使用需小心。

Exercise 10.5.1. Prove Proposition 10.5.1.
Proposition 10.5.1 (L’Hôpital’s rule I). Let X be a subset of \mathbf R, let f:X\to\mathbf R and g:X\to\mathbf R be functions, and let x_0\in X be a limit point of X. Suppose that f(x_0)=g(x_0)=0, that f and g are both differentiable at x_0, but g'(x_0)\neq 0. Then there exists a \delta >0 such that g(x)\neq 0 for all x\in (X\cap (x_0-\delta,x_0+\delta))-\{x_0\}, and

\lim\limits_{x\to x_0;x\in (X\cap (x_0-\delta,x_0+\delta))-\{x_0\}}\dfrac{f(x)}{g(x)}=\dfrac{f'(x_0)}{g'(x_0)}

Solution: By Proposition 10.1.7, since g'(x_0 )\neq 0, we can find a {\delta}>0 such that for all x\in X and |x-x_0 |\leq {\delta}, we have

|g(x)-g(x_0 )-g'(x_0 )(x-x_0 )|\leq \dfrac{|g'(x_0 )|}{2} |x-x_0 |

Using triangle inequality and g(x_0 )=0, we further get

\dfrac{|g'(x_0 )|}{2} |x-x_0 |\leq |g(x)|

Thus if x\in \left(X\cap (x_0-{\delta},x_0+{\delta})\right)-\{x_0 \}, we shall have |g(x)|>0, or g(x)\neq 0. In this case we can safely use Proposition 9.3.14 to have

\begin{aligned}\lim\limits_{x\to x_0;x\in (X\cap (x_0-{\delta},x_0+{\delta}))-\{x_0 \} } \dfrac{f(x)}{g(x)}&=\lim\limits_{x\to x_0;x\in (X\cap (x_0-{\delta},x_0+{\delta}))-\{x_0 \} } \frac{f(x)-f(x_0)}{g(x)-g(x_0 )}\\&=\dfrac{\lim\limits_{x\to x_0;x\in (X\cap (x_0-{\delta},x_0+{\delta}))-\{x_0 \} }\dfrac{f(x)-f(x_0 )}{x-x_0 }}{\lim\limits_{x\to x_0;x\in (X\cap (x_0-{\delta},x_0+{\delta}))-\{x_0 \} }\dfrac{g(x)-g(x_0)}{x-x_0 }} \\&=\dfrac{f'(x_0)}{g'(x_0 ) }\end{aligned}

\blacksquare

Exercise 10.5.2. Explain why Exercise 1.2.12 does not contradict either of the propositions in this section.

Solution: In the first example, we have g(x)=1+x\nrightarrow 0 if x\to 0.
In the second example, notice that to use the L’Hôpital’s rule, we need first the limit

\lim\limits_{x\to a;x\in (a,b]}\dfrac{f'(x)}{g'(x)}

exist, then we can safely use the rule, but in this example, the limit

\lim\limits_{x\to 0+;x\in (0,{\infty}) }\dfrac{(x^2 \sin (x^{-4}))'}{x'}

does not exist, due to the calculations in the textbook. So it’s not safe to use the L’Hôpital’s rule.

\blacksquare

陶哲轩实分析10.4及习题-Analysis I 10.4

反函数定理(实数轴上的),第二册中会有在n维欧式空间中的反函数定理。

Exercise 10.4.1. Let n\geq 1 be a natural number, and let g:(0,\infty)\to(0,\infty) be the function g(x):=x^{1/n}.
( a ) Show that g is continuous on (0,\infty).
( b ) Show that g is differentiable on (0,\infty), and that g'(x)=\frac{1}{n}x^{\frac{1}{n}-1} for all x\in (0,\infty).

Solution:

( a ) Since f(x)=x^n,x\in (0,{\infty}) is a continuous and strictly monotone increasing function, by Proposition 9.8.3, f^{-1} (x)=x^{1/n}=g(x) is continuous on (0,{\infty}).

( b ) We already have f'(x)=nx^{n-1},x\in (0,{\infty}), since g=f^{-1}, f is differentiable on (0,{\infty}), and g is continuous by (a), we can use the inverse function theorem to say that g is differentiable at any x\in (0,{\infty}), further for x\in (0,{\infty}), we have x^{1/n} such that f(x^{1/n})=x, thus

g'(x)=\dfrac{1}{f'(x^{1/n})}=\dfrac{1}{n(x^{1/n})^{n-1}}=\dfrac{1}{n} x^{\frac{1-n}{n}}=\dfrac{1}{n} x^{1/n-1}

\blacksquare

Exercise 10.4.2. Let q be a rational number, and let f:(0,\infty)\to\mathbf R be the function f(x)=x^q.
( a ) Show that f is differentiable on (0,\infty) and that f'(x)=qx^{q-1}.
( b ) Show that \lim_{x\to1;x\in (0,\infty)}\frac{x^q-1}{x-1}=q for every rational number q.

Solution:

( a ) As q\in \mathbf Q, thus \exists m\in \mathbf Z,n\in \mathbf N^+, s.t. q=m/n. By Exercise 10.4.1, we have g(x)=x^{1/n} is differentiable on (0,{\infty}). And f(x)=[g(x)]^m, use Theorem 10.1.13 (d) and possibly (g), we can conclude f is differentiable on (0,{\infty}). Use Theorem 10.1.15, we can say

\begin{aligned}f'(x)&=m[g(x)]^{m-1} g'(x)=mx^{\frac{m-1}{n}} \left(\dfrac{1}{n} x^{1/n-1} \right)\\&=\dfrac{m}{n} x^{\frac{m-1}{n}+\frac{1-n}{n}}=qx^{\frac{m-n}{n}}=qx^{q-1}\end{aligned}

( b ) We have f(1)=1^q=1, thus by part(a), f'(1) exists, so

\lim\limits_{x\to 1;x\in (0,{\infty})-\{1\}}\dfrac{x^q-1}{x-1}=\lim\limits_{x\to 1;x\in (0,{\infty})-\{1\}}\dfrac{f(x)-f(1)}{x-1}=f'(1)=q1^{q-1}=q

during which the second equality comes from Definition 10.1.1
Since the function (x^q-1)/(x-1) is undefined at x=1, we conclude

\lim\limits_{x\to 1;x\in (0,{\infty})-\{1\}}\dfrac{x^q-1}{x-1}=\lim\limits_{x\to 1;x\in (0,{\infty})-\{1\}}\dfrac{x^q-1}{x-1}=q

\blacksquare

Exercise 10.4.3. Let \alpha be a real number, and let f:(0,\infty)\to\mathbf R be the function f(x)=x^{\alpha}.
( a ) Show that \lim_{x\to 1;x\in (0,\infty)\backslash \{1\}}\frac{f(x)-f(1)}{x-1}=\alpha.
( b ) Show that f is differentiable on (0,\infty) and that f'(x)={\alpha}x^{\alpha -1}.

Solution:

( a ) Let {\varepsilon}>0 be an arbitrary small number, then as {\alpha}\in \mathbf R, we can find p,q\in \mathbf Q, s.t.

{\alpha}-{\varepsilon}<p<{\alpha}<q<{\alpha}+{\varepsilon}

If x>1, then x^p<x^{\alpha}<x^q, thus we have

\dfrac{x^p-1}{x-1}<\dfrac{x^{\alpha}-1}{x-1}<\dfrac{x^q-1}{x-1}

Use Exercise 10.4.2 we know that

\lim\limits_{x\to 1;x\in (1,{\infty})}\dfrac{x^p-1}{x-1}=p, \quad  \lim\limits_{x\to 1;x\in (1,{\infty}) }\dfrac{x^q-1}{x-1}=q

Thus we can say that

{\alpha}-{\varepsilon}<p\leq \lim\limits_{x\to 1;x\in (1,{\infty}) }\dfrac{x^{\alpha}-1}{x-1}\leq q<{\alpha}+{\varepsilon}

Also if x<1, then x^p>x^{\alpha}>x^q, but x-1<0, thus

\dfrac{x^p-1}{x-1}<\dfrac{x^{\alpha}-1}{x-1}<\dfrac{x^q-1}{x-1}

and similarly we have

{\alpha}-{\varepsilon}<p\leq \lim\limits_{x\to 1;x\in (1,{\infty}) }\dfrac{x^{\alpha}-1}{x-1}\leq q<{\alpha}+{\varepsilon}

So we can conclude

{\alpha}-{\varepsilon}<\lim\limits_{x\to 1;x\in (0,{\infty})-\{1\}}\dfrac{x^{\alpha}-1}{x-1}<{\alpha}+{\varepsilon} \implies \lim\limits_{x\to 1;x\in (0,{\infty})-\{1\}}\dfrac{x^{\alpha}-1}{x-1}={\alpha}

( b ) For \forall c\in (0,{\infty}), by Theorem 10.1.13 g(x)=(x/c)^{\alpha} is differentiable on (0,{\infty}), thus

\begin{aligned}&\lim\limits_{x/c\to 1;x/c\in (0,{\infty})-\{1\} }\dfrac{(x/c)^{\alpha}-1}{x/c-1}={\alpha} \implies \lim\limits_{x\to c;x\in (0,{\infty})-\{c\}}\dfrac{(x^{\alpha}-c^{\alpha})/c^{\alpha}}{(x-c)/c}={\alpha} \\&\implies \lim\limits_{x\to c;x\in (0,{\infty})-{c} }\dfrac{x^{\alpha}-c^{\alpha}}{x-c}={\alpha}c^{{\alpha}-1}=\lim\limits_{x\to c;x\in (0,{\infty})-{c} }\dfrac{f(x)-f(c)}{x-c}=f'(c)\end{aligned}

Thus we can conclude f is differentiable on (0,{\infty}), and f'(x)={\alpha}x^{{\alpha}-1}.

\blacksquare