Skip to Content
Math4111Exam reviewsMath 4111 Final Review

Math 4111 Final Review

Weierstrass M-test

Let n=1fn(x)\sum_{n=1}^{\infty} f_n(x) be a series of functions.

The weierstrass M-test goes as follows:

  1. Mn0\exists M_n \geq 0 such that xE,fn(x)Mn\forall x\in E, |f_n(x)| \leq M_n.
  2. Mn\sum M_n converges.

Then n=1fn(x)\sum_{n=1}^{\infty} f_n(x) converges uniformly.

Example:

Ver.0

x[1,1)\forall x\in [-1,1),

n=1xnn\sum_{n=1}^{\infty} \frac{x^n}{n}

converges. (point-wise convergence on [1,1)[-1,1))

x[1,1)\forall x\in [-1,1),

xnn1n\left| \frac{x^n}{n} \right| \leq \frac{1}{n}

Since n=11n\sum_{n=1}^{\infty} \frac{1}{n} diverges, we don’t know if the series converges uniformly or not using the weierstrass M-test.

Ver.1

However, if we consider the series on [1,1][-1,1],

n=1xnn2\sum_{n=1}^{\infty} \frac{x^n}{n^2}

converges uniformly. Let Mn=1n2M_n = \frac{1}{n^2}. This satisfies the weierstrass M-test. And this series converges uniformly on [1,1][-1,1].

Ver.2

x[12,12]\forall x\in [-\frac{1}{2},\frac{1}{2}],

n=1xnn\sum_{n=1}^{\infty} \frac{x^n}{n}

converges uniformly. Since xnn=xnn(1/2)nn12n=Mn\left| \frac{x^n}{n} \right|=\frac{|x|^n}{n}\leq \frac{(1/2)^n}{n}\leq \frac{1}{2^n}=M_n, by geometric series test, n=1Mn\sum_{n=1}^{\infty} M_n converges.

M-test still not applicable here.

n=1xnn\sum_{n=1}^{\infty} \frac{x^n}{n}

converges uniformly on [12,12][-\frac{1}{2},\frac{1}{2}].

Comparison test:

For a series n=1an\sum_{n=1}^{\infty} a_n, if

  1. Mn\exists M_n such that anMn|a_n|\leq M_n
  2. n=1Mn\sum_{n=1}^{\infty} M_n converges

Then n=1an\sum_{n=1}^{\infty} a_n converges.

Proving continuity of a function

If f:EYf:E\to Y is continuous at pEp\in E, then for any ϵ>0\epsilon>0, there exists δ>0\delta>0 such that for any xEx\in E, if xp<δ|x-p|<\delta, then f(x)f(p)<ϵ|f(x)-f(p)|<\epsilon.

Example:

Let f(x)=2x+1f(x)=2x+1. For p=1p=1, prove that ff is continuous at pp.

Let ϵ>0\epsilon>0 be given. Let δ=ϵ2\delta=\frac{\epsilon}{2}. Then for any xRx\in \mathbb{R}, if x1<δ|x-1|<\delta, then

f(x)f(1)=2x+13=2x2=2x1<2δ=ϵ.|f(x)-f(1)|=|2x+1-3|=|2x-2|=2|x-1|<2\delta=\epsilon.

Therefore, ff is continuous at p=1p=1.

You can also use smaller δ\delta and we don’t need to find the “optimal” δ\delta.

Play of open covers

Example of non compact set:

Q\mathbb{Q} is not compact, we can construct an open cover Gn=(,2)(2+1n,)G_n=(-\infty,\sqrt{2})\cup (\sqrt{2}+\frac{1}{n},\infty).

Every unbounded set is not compact, we construct an open cover Gn=(n,n)G_n=(-n,n).

Every k-cell is compact.

Every finite set is compact.

Let pAp\in A and AA is compact. Then A\{p}A\backslash \{p\} is not compact, we can construct an open cover Gn=(inf(A)1,p)(p+1n,sup(A)+1)G_n=(\inf(A)-1,p)\cup (p+\frac{1}{n},\sup(A)+1).

If KK is closed in XX and XX is compact, then KK is compact.

Proof:

Let {Gα}αA\{G_\alpha\}_{\alpha\in A} be an open cover of KK.

AA is open in XX, if and only if X\AX\backslash A is closed in XX.

Since X\KX\backslash K is opened in XX, {Gα}αA{X\K}\{G_\alpha\}_{\alpha\in A}\cup \{X\backslash K\} is an open cover of XX.

Since XX is compact, there exists a finite subcover {Gα1,,Gαn,X\K}\{G_{\alpha_1},\cdots,G_{\alpha_n},X\backslash K\} of XX.

Since X\KX\backslash K is not in the subcover, {Gα1,,Gαn}\{G_{\alpha_1},\cdots,G_{\alpha_n}\} is a finite subcover of KK.

Therefore, KK is compact.

Cauchy criterion

In sequences

Def: A sequence {an}\{a_n\} is Cauchy if for any ϵ>0\epsilon>0, there exists NN such that for any m,nNm,n\geq N, aman<ϵ|a_m-a_n|<\epsilon.

Theorem: In R\mathbb{R}, every sequence is Cauchy if and only if it is convergent.

In series

Let sn=k=1naks_n=\sum_{k=1}^{n} a_k.

Def: A series n=1an\sum_{n=1}^{\infty} a_n converges if the sequence of partial sums {sn}\{s_n\} converges.

ϵ>0\forall \epsilon>0, there exists NN such that for any m>nNm>n\geq N,

smsn=k=n+1mak<ϵ.|s_m-s_n|=\left|\sum_{k=n+1}^{m} a_k\right|<\epsilon.

Comparison test

If anbn|a_n|\leq b_n and n=1bn\sum_{n=1}^{\infty} b_n converges, then n=1an\sum_{n=1}^{\infty} a_n converges.

Proof:

Since n=1bn\sum_{n=1}^{\infty} b_n converges, ϵ>0\forall \epsilon>0, there exists NN such that for any m>nNm>n\geq N,

k=n+1mbk<ϵ.\left|\sum_{k=n+1}^{m} b_k\right|<\epsilon.

By triangle inequality,

k=nmakk=n+1makk=n+1mbk<ϵ.\left|\sum_{k=n}^{m}a_k\right|\leq \sum_{k=n+1}^{m} |a_k|\leq \sum_{k=n+1}^{m} b_k<\epsilon.

Therefore, ϵ>0\forall \epsilon>0, there exists NN such that for any m>nNm>n\geq N,

smsn=k=n+1mak<ϵ.|s_m-s_n|=\left|\sum_{k=n+1}^{m} a_k\right|<\epsilon.

Therefore, {sn}\{s_n\} is Cauchy, and n=1an\sum_{n=1}^{\infty} a_n converges.

Last updated on