Midterm Revision
-
The way to prove a series is uniform convergence
- Theorem (Cauchy Criterium for uniform convergence of series)
- Theorem (Weierstrass M-test)
- Definition
-
If we need to check whether it is uniform convergent in a interval, we may use
-
Assume that it converges absolutely at x = x_0Then it converges uniformly on the closed interval [-c, c], where c = |x_0|.
Then it converges uniformly on the closed interval [-c, c], where c = |x_0|. 2. Theorem (Abel's theorem)
-
-
The way to find the radius of convergence of a series
- D'Alembert Criterion (test)
- Root test (Cauchy)
- Remember to check the endpoints of intervals (Use comparison test)
-
Taylor Series
\[ T_{N,x_0}(x)=f\left(x_{0}\right)+\frac{f^{\prime}\left(x_{0}\right)}{1!}\left(x- x_{0}\right)+\frac{f''\left(x_{0}\right)}{2!}\left(x-x_{0}\right)^{2}+\cdots+\frac{f^{\left(N\right)}\left(x_{0}\right)}{N!} \left(x-x_{0}\right)^{N} \]\(f(x) - T_{N,a}(x) = \frac{f^{(N+1)}(c)}{(N+1)!} (x-a)^{N+1} \quad \text{for some } c \text{ between } a \text{ and } x\) (Bounded error)
\(\sin(x)=\sum_{k=0}^{\infty}\frac{(-1)^{k}x^{2k+1}}{(2k+1)!}\)
\(\frac{1}{1-x}=\sum_{n=0}^{\infty}x^{n}\)
\(e^{x}=\sum_{n=0}^{\infty}\frac{x^{n}}{n!}\)
For \(\ln(x)\), we know \(\ln(x)^{\prime}=\frac{1}{x}=\sum_{n=0}^{\infty}\left(1-x\right)^{n}=\sum_{n=0}^{\infty} \left(-1\right)^{n}\left(x-1\right)^{n}\)
Then \(\ln(x)=\int\sum_{n=0}^{\infty}\left(-1\right)^{n}\left(x-1\right)^{n}=\sum_{n=0} ^{\infty}\left(-1\right)^{n}\int\left(x-1\right)^{n}=\sum_{n=0}^{\infty}\left(-1\right )^{n}\frac{\left(x-1\right)^{n+1}}{n+1}\) (centered at \(x=1\))
\(\ln(x+1)=\sum_{n=0}^{\infty}\left(-1\right)^{n}\frac{\left(x\right)^{n+1}}{n+1}\)
-
The way to prove \(f\) is integrable
- Definition: \(U(f)=L(f)\)
- Show \(U(f, P_{\epsilon}) - L(f, P_{\epsilon}) < \epsilon\)
- Show $\lim\limits_{n \to +\infty}L(f, P_{n}) = \lim\limits_{n \to +\infty}U(f, P_{n}) $
- Show \(f\) is continuous on a closed interval
- Theorem (Integrable Limit theorem)
-
Theorem (Fundamental Theorem of Calculus)
Theorem
-
Theorem (Decomposition of real rational functions)
Theorem (Integration by parts)
Theorem (Substitution)
-
The length of its graph in \([a, b]\) is \(\displaystyle \int_a^b \sqrt{1 + (f'(x))^2} \, dx\)
-
\(\displaystyle\int_{0}^{1} \frac{1}{\sqrt{x}} dx\) where \(I = (0, 1]\)
Then consider \(\displaystyle\lim_{c \to 0^+} \int_{c}^{1} \frac{1}{\sqrt{x}} dx\), we have \(\displaystyle\int_{c}^{1} \frac{1}{\sqrt{x}} dx = \displaystyle\int_{c}^{1} x^{-1/2} dx = \frac{x^{1/2}}{1/2}\Big|_c^1 = 2\sqrt{x}\Big|_c^1 = 2 - 2\sqrt{c}\)
Then \(\displaystyle\lim_{c \to 0^+} 2 - 2\sqrt{c} = 2\), thus \(\displaystyle\int_{0}^{1}\frac{1}{\sqrt{x}}dx\) converges to \(2\)
-
\(\int \sec(x) dx\) \(=\) \(\ln(|\sec(x)+\tan(x)|)+c\)
Remember \(\sec(x) = \frac{1}{\cos(x)}\) and \(\arctan(x)'=\frac{1}{1+x^{2}}\)
We want to use integration by parts. How do we choose \(u\) and \(v'\)?
We pick \(u\) in the following order:
(1) inverse of trigonometric functions \(\arcsin, \arccos,...\)
(2) logarithmic function \(\ln(x)\)
(3) algebraic function \(x^n, \sqrt{x}, ...\)
(4) trigonometric function \(\sin(x), \cos(x), ...\)
(5) exponential functions \(e^x, 2^x, ...\)
\(\int \sec^3(x) dx = \int \sec(x) \sec^2(x) dx\)
Let \(u = \sec(x) \implies du = \tan(x)\sec(x) dx\)
\(v' = \sec^2(x) \implies v = \tan(x)\)Now, by parts
\(\int \sec^3(x) dx = \sec(x)\tan(x) - \int \tan(x) (\tan(x)\sec(x)) dx\)
\(= \sec(x)\tan(x) - \int \tan^2(x) \sec(x) dx\)
\(= \sec(x)\tan(x) - \int (\sec^3(x) - \sec(x)) dx\)
\(= \sec(x)\tan(x) - \int \sec^3(x) dx + \int \sec(x) dx\)So \(\int \sec^{3}(x) dx = \frac{1}{2}(\sec(x)\tan(x) + \ln|\sec(x) + \tan(x)|) + C\)
Because \(\tan^2(x)\sec(x) = \frac{\sin^2(x)}{\cos^2(x)} \sec(x) = \frac{(1-\cos^2(x))}{\cos^2(x)} \sec(x) = (\sec^2(x) - 1) \sec(x)\)
-
(20 points)
Prove that the function \(f(x) = \begin{cases} \frac{1}{q}, & \text{if } x = \frac{p}{q} \text{ for } p \in \mathbb{Z}, q \in \mathbb{N}, \text{ and } p, q \text{ are mutually prime} \\ 0, & \text{if } x \text{ is irrational}. \end{cases}\) is integrable on \([0,1]\).Consider \(P_{n}=\{0,\frac{1}{n},\frac{2}{n},\ldots,1\}\), then \(m_{i}=\inf\{f(x):x\in[x_{i-1},x_{i}]\}=0\) since irrational number is dense in any compact set.
Then \(M_{i}=\sup\{f(x):x\in[x_{i-1},x_{i}]\}\) and \(U(f,P_{n})-L\left(f,P_{n}\right)=\sum_{i=1}^{n}\left(M_{i}-m_{i}\right)\left(x_{i} -x_{i-1}\right)=\sum_{i=1}^{n}M_{i}\left(x_{i}-x_{i-1}\right)\)
Let's \(B_{N}=\{1,\frac{1}{2},\frac{1}{3},\frac{2}{3},...,\frac{N-1}{N}\}\) with \(|B_N|=m\), then consider \(P_{n}=\{x_0=0,x_1,...,x_{n-1},x_n=1\}\)
We know \(B_{N}\subset [0,1]\), then \(\sum_{i=1}^{n}M_{i}\left(x_{i}-x_{i-1}\right)=\sum_{\left\lbrack x_{i-1},x_{i}\right\rbrack\cap B_{N}\neq\emptyset}M_{i}\left(x_{i}-x_{i-1}\right)+\sum_{\left\lbrack x_{i-1},x_{i}\right\rbrack\cap B_{N}=\emptyset}M_{i}^{\prime}\left(x_{i}-x_{i-1}\right)\)
Since \(M_{i}=\sup\{f(x):x\in[x_{i-1},x_{i}]\cap B_{N}\}=\frac{1}{i-1}\) and \(M_{i}^{\prime}=\sup\{f(x):x\in[x_{i-1},x_{i}]\setminus B_{N}\}\leq\frac{1}{N+1}\)
And we know there are at most \(2m\) subintervals such that \(\left\lbrack x_{i-1},x_{i}\right\rbrack\cap B_{N}\neq\emptyset\) because one point in \(B_N\) can be in two intervals
Then \(\sum_{i=1}^{n}M_{i}\left(x_{i}-x_{i-1}\right)\leq2m\cdot1\cdot\max\left(x_{i}-x_{i-1} \right)+\frac{1}{N+1}\sum\left(x_{i}-x_{i-1}\right)\leq2m\cdot1\cdot\max\left(x_{i} -x_{i-1}\right)+\frac{1}{N+1}\left(1-0\right)\leq2m\cdot1\cdot\max\left(x_{i}-x_{i-1} \right)+\frac{1}{N+1}\)
Then let \(\max\left(x_{i}-x_{i-1}\right)<\frac{\varepsilon}{4m}\) and \(\frac{1}{N+1}<\frac{\varepsilon}{2}\), then we have \(\sum_{i=1}^{n}M_{i}\left(x_{i}-x_{i-1}\right)<\varepsilon\)
Thus \(U(f,P_{n})-L\left(f,P_{n}\right)<\varepsilon\), then \(f\) is integrable
泰勒展开精度问题 不四舍五入