Je-O’s MA2108 Cheatsheet 2.

5 Completeness of real numbers

• Supremum Property of R: For ∅ = ̸ S ⊂ R, if it has an upper bound (lower
•
∞
P
n=1
an is called absolutely convergent if
∞
P
n=1
|an | is convergent.

bound), then it has supremum (infimum). ∞

• If
P
an is convergent but not absolutely convergent, we call it conditionally
1. Preliminaries • Archimedean Property of R: For any x, y > 0, there exists n ∈ N such that
nx > y.
n=1
convergent.

1.1 Functions • For any c > 0, there exists a unique nc ∈ N such that nc − 1 ≤ c ≤ nc .
•
∞
P
an is convergent if it is absolutely convergent.
• n=1
n o
f : A → B is called injective if for any x1 ̸= x2 , f (x1 ) ̸= f (x2 ). • inf 1 : n ∈ N = 0
n
• f : A → B is called surjective if f (A) = B. • Comparison Test: Suppose {an } and {bn } satisfy that there exist K ∈ N and
• Q is dense in R, namely, for any open interval I = (a, b), there exists r ∈ Q such ∞P
• f : A → B is called bijective if it is both injective and surjective. that r ∈ I. λ > 0 such that for all n ≥ K, 0 ≤ an ≤ λbn . Then bn is convergent implies
n=1
• For S ⊂ R such that sup S (inf S) exists and sup S ∈ S (inf S ∈ S), we call sup S ∞
1.2 Mathematical Induction
P
(inf S) as max S (min S). an is convergent.
n=1
• Well-ordering Principle of N: For any nonempty subset S ⊆ N, S has a minimal ∞ ∞
3. Sequences and Series
P P
element, that is, there exists m ∈ S such that for any s ∈ S, m ≤ s. Then an is divergent implies bn is divergent.
n=1 n=1
• Principle of Strong Induction: Prove P (n) for every n ∈ N by proving P (1) and •
for any k′ ≤ k, P (k′ ) holds, then P (k + 1) holds. 3.1 Nested Intervals Limit Comparison Test: Suppose {an } and {bn } are both positive sequences and
lim
an
= λ.
• Let {In = [an , bn ] : n ∈ N} be a sequence of closed intervals such that for any n→∞ bn
n ∈ N, In+1 ⊆ In . Then there exists x ∈ R such that ∞ ∞
1.3 Finite and Infinite Sets ∞
If λ > 0, then
P
n=1
an convergent iff
P
n=1
bn convergent.
• A nonempty set S is finite if there exists a maximal n ∈ N, called the cardinality
\
x ∈ In ∞
P ∞
P ∞
P
of S, denoted by |S|, such that S has n elements. n=1 If λ = 0, then bn convergent implies an convergent, an divergent
n=1 n=1 n=1
• For n ∈ N, the union of n finite sets is also finite. ∞

• An infinite set S is countable if there exists a bijective map f : N → S. S is

3.2 Sequences implies
P
n=1
bn divergent.
• A sequence {an : n ∈ N} is a function a : N → R.
countable iff elements of S can be listed as S = {s1 , s2 , . . . , sn , . . .}. ∞
• We say a sequence {an : n ∈ N} approaches A as n → ∞ if for any ϵ > 0, there • 1 is convergent.
P
For p-series with p > 1,
∞ exists N ∈ N (depending on ϵ) such that for any n ≥ N , |an − A| < ϵ. np
• Suppose set S is infinite and S =
S
Fi where Fi are finite sets, then S is n=1
i=1 • If a sequence is convergent, then its limit is unique. • The Cauchy Condensation Test: Suppose {an } is decreasing and every term is
countable. ∞ ∞
• positive. Define bk = 2k a k . Then
P P
For a sequence {an : n ∈ N}, the following statements are equivalent: an is convergent iff bk is convergent.
• The union of countably many countable sets is also countable. 1. limn→∞ an = A
2 n=1 k=1
2. limn→∞ (an − A) = 0 • Root Test: Let {an } be a sequence.
2. Real Numbers •
3. limn→∞ |an − A| = 0
A convergent sequence is bounded, namely, there exists C > 0 such that |an | ≤ C If
P∞
1
there exists 0 ≤ r < 1 and K ∈ N such that |an | n ≤ r for all n ≥ K, then
a n is absolutely convergent.
2.1 Algebraic properties for any n ∈ N. n=1
1 P∞
• Wrt +, A1 (commutative law), A2 (associative law), A3 (existence of a zero • Suppose {an }, {bn } are convergent, and an ≥ bn , then If there exists K ∈ N such that |an | n ≥ 1 for all n ≥ K, then n=1 an is
element), A4 (existence of negative elements). lim a ≥ lim bn divergent.
n→∞ n n→∞
• Wrt ×, M1 (commutative law), M2 (associative law), M3 (existence of a unit • Ratio Test: Let {an } be a sequence with non-zero terms.
• Squeeze Theorem: Let sequences {an }, {bn }, {cn } satisfy that an ≤ bn ≤ cn . If there exists 0 < r < 1 and K ∈ N such that |an+1 /an | ≤ r for all n ≥ K, then
element), M4 (existence of reciprocals).
Suppose limn→∞ an = limn→∞ cn = A. Then limn→∞ bn = A. P∞
• Moreover, it satisfies, D1 (distributive law of multiplication over addition). n=1 an is absolutely convergent.
•
P∞
Given a sequence {an : n ∈ N}, we say that {bk : k ∈ N} is a subsequence of {an } if If there exists K ∈ N such that |an+1 /an | ≥ 1 for all n ≥ K, then n=1 an is
• Any nonempty set F with operators (+, ×) satisfying (A1-4, M1-4, D1) is called a for any k ∈ N, bk = an , where {nk : k ∈ N} satisfies that for any k ∈ N, nk ∈ N divergent.
k
field. and nk < nk+1 .

2.2 Important inequalities

• Let {bk = an } be a subsequence of {an }. Suppose limn→∞ an = A. Then
k
4. Limits of Functions
limk→∞ bk = A.
• Bernoulli’s Inequality: For x > −1, n ∈ N, we have
• Monotone Convergence Theorem: Suppose sequence {an : n ∈ N} is increasing
4.1 Definitions
n • Let A ⊂ R be a nonempty subset of R. x ∈ R is called a cluster point of A if for
(1 + x) ≥ 1 + nx (decreasing) and bounded from above (below). Then {an } is convergent.
any ϵ > 0, A ∩ (Vϵ (x) \ {x}) ̸= ∅.
• AM-GM Inequality: For a1 , a2 , . . . , an ≥ 0, • Monotone Subsequence Theorem: Any sequence {an } admits a monotone
• x ∈ R is a cluster point of A iff there exists a sequence {an } such that an ∈ A,
a1 + · · · + an √ subsequence {an }.
≥ n a1 · · · an k an ̸= x, and limn→∞ an = x.
n • The Bolzano-Weierstrass Theorem: Any bounded sequence {an } admits a • Limits of Function: Let f : A → R be a function on A and c be a cluster point of
convergent subsequence {an }. A. Suppose L ∈ R satisfies that for any ϵ > 0, there exists δ > 0 such that
• GM-HM Inequality: For a1 , a2 , . . . , an > 0, k
√ n |f (x) − L| < ϵ for all x ∈ A ∩ (Vδ (c) \ {c}). Then limx→c f (x) = L.
n a1 · · · an ≥ • We say that {an } is a Cauchy sequence if for any ϵ > 0, there exists N ∈ N such
1 + ··· + 1 that for any m, n ≥ N , |am − an | < ϵ. • If f : A → R does not have a limit at a cluster point c of A, then f is divergent at c.
a1 an
• Cauchy Convergence Criterion: A sequence is convergent iff it is a Cauchy • Let f : A → R be a function on A and c be a cluster point of A. Then f is bounded
• Triangle Inequality: For a, b ∈ R, sequence. in a neighborhood of c if there exist δ > 0 and C > 0 such that |f (x)| ≤ C, for all
||a| − |b|| ≤ |a + b| ≤ |a| + |b| x ∈ A ∩ Vδ (c).
• For {an } and {bn },
If f : A → R has a limit at a cluster point c of A, then f is bounded in a
• Cauchy-Schwartz Inequality: For a1 , . . . , an , b1 , . . . , bn ∈ R, if an ≥ bn and lim bn = +∞, then lim an = +∞
n→∞ n→∞ neighborhood of c.
 2    if an ≥ bn and lim an = −∞, then lim bn = −∞
n n n n→∞ n→∞ • Order Preserving Theorem: Let f, g : A → R be functions and c be a cluster
X X 2 X 2
ai bi  ≤  ai   bi  point of A. Suppose limx→c f (x) = L, limx→c g(x) = M , and there exists δ > 0
3.3 Infinite Series

i=1 i=1 i=1 such that f (x) ≥ g(x) for all x ∈ A ∩ (Vδ (c) \ {c}). Then, L ≥ M .
∞
equality holds iff ∃k ∈ R such that ai = kbi for i = 1, . . . , n. •
P
Suppose ai is convergent. Then lim a = 0.
n→∞ n • Squeeze Theorem: Let f, g, h : A → R be functions and c be a cluster point of A.
i=1 Suppose limx→c f (x) = limx→c g(x) = L, and there exists δ > 0 such that
2.3 ϵ-Neighborhood • Cauchy Criterion:
∞
P
ai is convergent iff for any ϵ > 0, there exists N ∈ N such
f (x) ≤ h(x) ≤ g(x) for all x ∈ A ⋒ (Vδ (c) \ {c}). Then, limx→c h(x) = L.
• For x ∈ R and ϵ ≥ 0, let us define the ϵ-neighborhood of x, denoted by Vϵ (x), as i=1 • Sequential Criterion of Limits: Let f : A → R be a function and c be a cluster
Vϵ (x) = (x − ϵ, x + ϵ). that for any m > n ≥ N , |an+1 + · · · + am | < ϵ. point of A. The following two statements are equivalent:
(1) limx→c f (x) = L
• Let a ∈ R. Suppose x ∈ Vϵ (a) for any ϵ > 0. Then x = a. • Monotone Convergence Theorem: (2) For any {an } ⊂ A \ {c} s.t. limn→∞ an = c, limn→∞ f (an ) = L
n
• Let I = (a, b) be an open interval. Then for any x ∈ I, there exists ϵ > 0 such that
P
Suppose ai ≥ 0 for any i ∈ N, and there exists C > 0 such that ai ≤ C for any • limx→c f (x) = L iff limx→c− f (x) = limx→c+ f (x) = L.
Vϵ (x) ⊂ I. i=1
∞
P • Infinite Limits: f : A → R, c : cluster point of A.
n ∈ N. Then ai is convergent. We say that limx→c f (x) = ∞ if for any L > 0, there exists δ > 0 such that
2.4 Bounds i=1
∞ ∞ |f (x)| > L for all x ∈ A ∩ (Vδ (c) \ {c}).
• Given a nonempty subset S ⊂ R, u ∈ R (l ∈ R) is called an upper bound (lower
P P
Suppose ai ≥ 0, and ai is divergent. Then ai = +∞. • Limits at infinity: f : A → R, there exists c ∈ R such that (c, +∞) ⊂ A.
bound) of S if for any s ∈ S, s ≤ u (s ≥ l). i=1 i=1
We say that limx→+∞ f (x) = L if for any ϵ > 0, there exists M > c such that
• ∅ =
̸ S ⊂ R is called bounded from above (below) if it has upper bounds (lower • The Leibniz Test: Suppose an ≥ 0 is decreasing and lim a = 0. Then the
n→∞ n |f (x) − L| < ϵ for all x > M .
bounds). following series is convergent.
• Sign Preserving Theorem: Let f : A → R be a function and c be a cluster point
∞
• sup S (inf S) is the supremum/least upper bound (infimum/greatest lower bound) X n−1 of A. Suppose limx→c f (x) = L > 0. Then there exists δ > 0 and C > 0 such that
(−1) an
of S. f (x) ≥ C for all x ∈ A ∩ (Vδ (c) \ {c}).
n=1
5. Continuous Functions 5.4 Monotone Functions on Intervals 6.3 Compactness
• One-sided Limits for Monotone Functions Exist: Let I ⊂ R be an interval and • K ⊆ R is called compact if for any open covering of K: K ⊆α∈A Uα , where every
5.1 Definition f : I → R be an increasing function on I, namely, f (x) ≤ f (y) for all x < y. Let
c ∈ I such that c is not an endpoint of I. Then limx→c− f (x) and limx→c+ f (x)
Uα is open, there exists a finite subcovering K ⊆n
i=1 Uαi .
• Let f : A → R be a function and c ∈ A. If for any ϵ > 0, there exists δ > 0 such both exist. • Heine-Borel Theorem: K ⊆ R is compact iff K is closed and bounded.
that |f (x) − f (c)| < ϵ for all x ∈ A ∩ Vδ (c), then we say f is continuous at c.
• Jump: Let I ⊂ R be an interval and f : I → R be a monotone function on I. • Let K ⊆ R be a compact set and f : K → R be a continuous function. Then f (K) is
• If c is not a cluster point of A, then f is always continuous at c. If c is a cluster Suppose c ∈ I is not an endpoint of I. The jump of f at c, compact.
point of A, then f is continuous at c iff limx→c f (x) = f (c). jf (c) = limx→c+ f (x) − limx→c− f (x).

• Sequential Criterion for Continuity: Let f : A → R be a function and c ∈ A. f

• Let I ⊂ R be an interval and f : I → R be a monotone function on I. Suppose c ∈ I 6.4 Connectedness
is not an endpoint of I. Then f is continuous at c iff jf (c) = 0. • A subset S ⊆ R is called disconnected if there exist open sets A, B such that
is continuous at c iff for any sequence {an } ⊂ A with limn→∞ an = c,
S ⊂ A ∪ B, A ∩ S, B ∩ S ̸= ∅, and A ∩ B ∩ S = ∅. Otherwise U is called connected.
limn→∞ f (an ) = f (c). • Let I ⊂ R be an interval and f : I → R be a monotone function on I. Then
discontinuous points of f in I is finite or countable. • S ⊆ R is connected iff S is an interval, namely, for any s1 < s2 ∈ S, [s1 , s2 ] ⊆ S.
• Composition of Continuous Functions: Let f : A → R and g : B → R be
• Continuous Inverse Theorem: Let f : I → R be a strictly increasing and • Intermediate Value Theorem: Let f : A → R be a continuous function and S ⊆ A
functions such that f (A) ⊆ B. Suppose f is continuous at c ∈ A and g is
continuous at f (c), then g ◦ f is continuous at c. continuous function defined on an interval I. Then the inverse function f −1 of f is be a connected set. Then f (S) is connected.
also strictly increasing and continuous function defined on an interval J. Moreover,

5.2 Continuous Functions on Intervals

J = f (I). 6.5 Metric Space
• Given a nonempty set S, d : S × S → R is called a metric on S if it satisfies:
• f : A → R is called bounded if there exists C > 0 such that |f (x)| ≤ C for all 6. Basic Topology - d(x, y) ≥ 0, equality holds iff x = y
- d(x, y) = d(y, x)
x ∈ A.

• Boundedness Theorem: Suppose f : [a, b] → R is a continuous function on [a, b].

6.1 Open and Closed Sets - Triangle inequality: For any x, y, z ∈ S, d(x, z) ≤ d(x, y) + d(y, z)
S with a metric d is called a metric space.
• A subset O ⊆ R is called open if for any x ∈ O, there exists δ > 0 such that
Then f is bounded.
Vδ (x) ⊆ O. • A metric d on S naturally defines a topology on S: U ⊆ S is open iff for any x ∈ U ,
• Maximum-Minimum Theorem: Suppose f : [a, b] → R is a continuous function on • A subset C ⊆ R is called closed if R \ C is open. there exists δ > 0 such that Vδ (x) ⊆ U , where Vδ (x) = {y ∈ S : d(x, y) < δ}.
[a, b]. Then f has an absolute maximum and absolute minimum. • In a metric space (S, d), we can define sequential convergence: Given a sequence
• ∅, R are both closed and open.
{an } ⊆ S and s ∈ S, we say that an approaches s as n → ∞, denoted by
• Bolzano’s Intermediate Value Theorem: Suppose f : [a, b] → R is a continuous • Usually, a ”topological” space is a nonempty set S with a collection of subsets limn→∞ an = s, if limn→∞ d(an , s) = 0.
function on [a, b]. Let m = min{f (a), f (b)} and M = max{f (a), f (b)}. Then for called ”open sets”. Closed subsets are defined to be complements of open sets.
any L ∈ [m, M ], there exists c ∈ [a, b] such that f (c) = L. • In a metric space (S, d), we can define Cauchy sequences: Given a sequence
• Any union of open sets is open. Any finite intersection of open sets is open.
{an } ⊆ S is called Cauchy if for any ϵ > 0, there exists N ∈ N such that for any
• Location of Roots Theorem: Suppose f : [a, b] → R is a continuous function on • Any intersection of closed sets is closed. Any finite union of closed sets is closed. m, n ≥ N , d(am , an ) < ϵ.
[a, b]. Let m = min{f (a), f (b)} and M = max{f (a), f (b)}. Suppose m ≤ 0 and
• Examples: • A metric space (S, d) is called complete if any Cauchy sequence is convergent in S.
M ≥ 0. Then there exists c ∈ [a, b] such that f (c) = 0.
Open sets: (a, b), (a, +∞), (−∞, a), union of open intervals (R, | · |) is complete, while (Q, | · |) is not.
• Suppose f : [a, b] → R is a continuous function on [a, b]. Then the range f ([a, b]) is Closed sets: [a, b], finite sets, Z
Not open, not closed: Q, {1/n : n ∈ N}
a closed interval.
• Infinite intersection of open sets might not be S
open: In = (−1/n, 1 + 1/n). Infinite
7. Useful Facts
• Preservation of Intervals Theorem: Let I be an interval. Suppose f : I → R is a union of closed sets might not be closed: Q = ∞ n=1 Fn , where F is finite. 7.1 Limits
continuous function on I. Then the range f (I) is an interval.
• Characterization of Closed Sets: F ⊆ R is closed iff for any convergent sequence
• Trigonometric Functions
{an } ⊆ F , limn→∞ an ∈ F .
limx→+∞ sin x = limx→+∞ tan x = 1
5.3 Lipschitz Function • F ⊆ R is closed iff it contains all its cluster points. x x
limx→+∞ 1−cos x = 1
• f : A → R is called Lipschitz if there exists K > 0 such that • Characterization of Open Sets: U ⊆ R is open iff it is a disjoint union of x2 2
|f (x) − f (y)| ≤ K|x − y| for all x, y ∈ A. K is called the Lipschitz constant for f . countably many open intervals. limx→+∞ tan x−sin x = 1
x3 2
• If f : A → R is a Lipschitz function on A, then f is uniformly continuous on A. • Induced Topology: Let A ⊆ R be a nonempty subset of R. B ⊆ A is called open in
A if there exists open subset U ⊆ R such that B = U ∩ A. • Logarithmic Functions
limx→+∞ x(log(x + 1) − log(x)) = 1
• Uniformly continuous functions preserve Cauchy sequences: Let f : A → R be
a uniformly continuous function. Then for any Cauchy sequence {an } from A, 6.2 Global Continuity Theorem limx→0 (1 + x)1/x = e
{f (an )} is also Cauchy. log(1+x)
• f : A → R is continuous on A iff for any open set U ⊆ R, f −1 (U ) is open in A. limx→0
x
= 1
• Continuous Extension Theorem: f : (a, b) → R is uniformly continuous on (a, b) x
• f : A → R is continuous on A iff for any open interval I ⊆ R, f −1 (I) is open in A. limx→0 e −1 = 1
iff f can be extended to a continuous function on [a, b], i.e., both limx→a+ f (x) x
(1+x)a −1
and limx→b− f (x) exist. • f : R → R is continuous on R iff for any open interval I ⊆ R, f −1 (I) is an open set. limx→0 = a
x