Math 140B Final Exam

Review Sheet
For: UC San Diego
Spring 2024

Chapter Coverage: Principles of Mathematical Analysis by Walter Rudin, Chapter 5-8.

For this review sheet, definitions and theorems are assumed to be known (or just men-
tioned/referred to). This document compiles the theorems you need to know for the final
exam, as a strong grasp of the theorems is going to help a lot.

1 Important Things
1. (Chain Rule) Let f, g be real-valued functions with real-valued domains. Suppose f and
g are differentiable at x, then g ◦ f is differentiable at x and

(g ◦ f )′ (x) = g ′ (f (x)) · f ′ (x).

2. Let f : (a, b) → R. A function has a local maximum at p if ∀x ∈ Bδ (p), f (x) ≤ f (p) for
some small δ > 0. If f has a local maximum (or local minimum) at x, then f ′ (x) = 0.

3. (Mean Value Theorems) If f, g : [a, b] → R, differentiable on (a, b) and continuous on

[a, b], then there is some point x ∈ (a, b) such that

(f (b) − f (a))g ′ (x) = (g(b) − g(a))f ′ (x).

(a) (MVT) If g(x) = x, then ∃x ∈ (a, b) such that f (b) − f (a) = (b − a)f ′ (x).
(b) If f ′ (x) ≥ 0 (or f ′ (x) ≤ 0), then f is monotone increasing (or decreasing).

4. (Derivative IVT) Suppose f : [a, b] → R is diff. and λ ∈ R satisfies f ′ (a) < λ < f ′ (b).
Then there exists x ∈ (a, b) with f ′ (x) = λ.

5. (L’Hospital’s Rule) Suppose −∞ ≤ a ≤ b ≤ +∞, and f, g : (a, b) → R are differentiable,

f ′ (x)
and ∀x ∈ (a, b), g ′ (x) ̸= 0. If lim ′ = A and g(x) → 0, f (x) → 0, OR g(x) → +∞
x→a g (x)
f (x)
or g(x) → −∞ as x → a, then lim = A, where A is in the extended reals.
x→a g(x)

6. (Taylor’s Theorem/Higher-Order Mean Value Theorem) Suppose f : [a, b] → R

and for some n ∈ Z+ , f (n−1) is defined and continuous on [a, b]; also f (n) is continuous on
(a, b).
n−1 (k)
X f (α)
Pick α ∈ [a, b] and define P (t) = (t − α)k . Then for every β ∈ [a, b] \ {α},
k=0
k!
f (n) (x)
there is some x between α and β such that f (β) = P (β) + (β − α)n .
n!
1
 Z x
7. (Fundamental Theorem of Calculus) Let f ∈ Rba and define F (x) = f (t) dt for
a
a ≤ x ≤ b. Then,

(a) F is continuous
(b) If f is continuous at x0 , then F is differentiable at x0 and F ′ (x0 ) = f (x0 ).
Z b
′
(c) If F : [a, b] → R is differentiable and F = f , then f dx = F (b) − F (a).
a

8. (Riemann Integrability) Let f be bounded and α monotone increasing. (Rba is only defined
on bounded functions.) Then, f ∈ Rba ⇐⇒ ∀ε > 0, ∃P : U (P, f, α) − L(P, f, α) < ε.

(a) (Improves bound) If P ⊆ P ∗ then U (P ∗ , f, α)−L(P ∗ , f, α) ≤ U (P, f, α)−L(P, f, α).

(b) If f : [a, b] → R has finitely many discontinuities, and α : [a, b] → R is monotone
increasing, and is continuous at the points where f is discontinuous, then f ∈ Rba (α).
(c) If f ∈ Rba and ϕ is continuous, then ϕ ◦ f ∈ Rba (α).
(d) If f, g ∈ Rba (α), then f g ∈ Rba (α).
Z b
b
Rb
(e) |f | ∈ Ra (α) and f dα ≤ a |f | dα.
a

9. (Change of Variables) Let α : [a, b] → R be monotone increasing and let f : [a, b] → R

where f ∈ Rba . If ϕ : [A, B] → [a, b] is a strictly increasing continuous bijection, then
Z B Z b
f ◦ ϕ d(α ◦ ϕ) = f dα.
A a

The special case when α(x) = x and ϕ is differentiable with ϕ′ ∈ Rba is just
Z B Z b
′
f (ϕ(t))ϕ (t) dt = f (x) dx.
A a

10. (Integration by Parts) If F ′ = f ∈ Rba and G′ = g ∈ Rba , then F g, f G ∈ Rba and

Z b Z b
F g dx = F (b)G(b) − F (a)G(a) − f G dx.
a a

11. (Integration of Vectors) If f⃗ ∈ Rba , then |f⃗| ∈ Rba .

12. (Rectifiable Curves) A curve γ is a continuous mapping γ : [a, b] → Rk . The length

X n
of a curve is defined as Λ(γ) = supP (P, γ) where Λ(P, γ) = |γ(xi ) − γ(xi−1 )| for
i=1
P = {x0 , . . . , xn }.

(a) If Λ(γ) < ∞, it’s rectifiable.

Z b
′
(b) If γ is differentiable and γ is continuous, then γ is rectifiable and Λ(γ) = |γ ′ (t)| dt.
a

2
13. (Problems with Pointwise Convergence) That’s why we need to define uniform conver-
gence.

(a) Order of limits matter.

(b) (Loss of continuity) A convergent sequence of continuous functions might not con-
verge to a function that is continuous.
(c) (Loss of Integrability) A convergent sequence of Riemann-Integrable functions might
not converge to a function that is Riemann-integrable.
(d) (Nonconvergence of derivatives) If a convergent sequence of differentiable functions
{fn } converging to f might not satisfy fn′ → f .
(e) (Nonconvergence of integrals) Integration doesn’t commute with limits.

14. (Uniform Convergence) Let fn , f : E → C. We say that (fn ) converges uniformly to

f on E if ∀ε > 0, ∃N so that ∀n ≥ N and ∀x ∈ E, |fn (x) − f (x)| < ε. (fn is close to f )

(a) (Cauchy Criterion) (fn ) converges uniformly on E if and only if ∀ε > 0, ∃N such
that for all m, n ≥ N and ∀x ∈ E, |fn (x) − fm (x)| < ε. This is because C is a
completer metric space.
 
(b) (fn ) converges uniformly on E to f if and only if lim sup |fn (x) − f (x)| = 0.
n→∞ x∈E

∞
X
15. (Weierstrass M-Test) Let fn : E → C such that ∀x ∈ E, |fn (x)| ≤ Mn . If Mn < ∞,
n=1
∞
X
then fn converges uniformly on E.
n=1

16. (Uniform Convergence and Continuity)

(a) Suppose (fn ) → f uniformly on E and let x ∈ E ′ . Also, assume that for each n,
lim fn (t) = An ∈ C. Then, (An ) converges and lim f (t) = lim An .
t→x t→x n→∞

(b) If (fn ) is a sequence of contiuous functions which converge uniformly to f , then f is

continuous.
(c) Suppose (fn ) is a sequence of continuous functions over a compact metric space K
such that (fn ) → f pointwise. Then if f is continuous and fn (x) ≤ fn+1 (x) for all n
and x, then fn → f uniformly on K.

17. (Uniform Convergence and Integration)

(a) (Uniform Convergence of Integrable Functions preserves Integration) Let

α : [a, b] → R be monotone increasing. Suppose fn ∈ Rba (α) and fn → f uniformly
Z b Z b
b
on [a, b]. Then, f ∈ Ra (α) and f dα = lim fn (x) dα.
a n→∞ a
∞
X
(b) (Term-by-Term Convergence) If fn ∈ Rba and fn converges uniformly to f on
n=1
Z b ∞
X
[a, b], then f dα = fn dα.
a n=1

3
18. (Uniform Convergence and Differentiation) Assume fn : [a, b] → C are diff. and
there is x0 ∈ [a, b] such that (fn (x0 )) converges. If (fn′ ) converges uniformly on [a, b] then
(fn ) converges uniformly to some differentiable function f , and ∀x ∈ [a, b],

f ′ (x) = lim fn′ (x).

n→∞

19. The sequence of functions {fn } on E is said to be pointwise bounded if ∀x ∈ E,

(fn (x)) is a bounded sequence. It’s uniformly bounded if |fn (x)| < M for all n ∈ Z+
and x ∈ E.

20. (Equicontinuity) A family F of complex functions f is equicontinuous (uniformly, uni-

formly continuous) if for every ε > 0, there exists δ > 0 so that for all functions f ∈ F,

d(x, y) < δ =⇒ |f (x) − f (y)| < ε.

(a) If {fn } is a pointwise bounded sequence on a countable set E, then we can find a
subsequence {fnk } such that {fnk } converges for every x ∈ E.
(b) If K is compact and fn is a continuous function on K, and {fn } converges uniformly
on K, then {fn } is equicontinuous on K.
(c) (Arzela-Ascoli Theorem) If K is compact, fn is continuous on K and {fn } is
pointwise bounded and equicontinuous on K, then {fn } is uniformly bounded on K
and contains a uniformly convergent subsequence.

21. A family A of complex functions defined on E is said to be an algebra if it’s closed

under functional addition, multiplication, and scalar multiplication. If whenever fn → f
uniformly on E, f ∈ A, then A is uniformly closed.

(a) A is said to separate points on E if ∀x1 , x2 ∈ E where x1 ̸= x2 , there is a function

f ∈ A so that f (x1 ) ̸= f (x2 ).
(b) If for every x ∈ E, there is f ∈ A so that f (x) ̸= 0, then A vanishes at no point of
E.
(c) (Stone-Weierstrass Theorem) If A is an algebra of real continuous functions on a
compact set K which separates points and vanishes at no point of K, then the
uniform closure of A consists of all real continuous functions on K. Furthermore, if
A is an algebra of self-adjoint complex continuous functions which vanishes at no
point and separates points on K, then the uniform closure consists of all complex
continuous functions on K.
∞
X
22. (Power Series) The power series cn xn converges uniformly within the radius of con-
n=1
vergence and has derivatives of all orders.
∞
X f (n) (a)
(a) If |a| < R, then f (x) = (x − a)n at least for |x| < R − |a|.
n=0
n!
∞
X ∞
X
(b) If cn converges, then f (x) = cn xn has a limit as x → 1, in particular
n=0 n=0
∞
X
lim f (x) = f (1) = cn .
x→1−
n=0

4
 ∞
X ∞
X
(c) (Uniqueness) Suppose an , bn converge on (−R, R). Letting
n=0 n=0
( ∞ ∞
)
X X
E= x ∈ (−R, R) : an = bn ,
n=0 n=0

if E has a limit point in (−R, R), then E = (−R, R) and an = bn for all n.
∞
X zn
23. (Exponential and Logarithmic Function) Define ez = 1 + for all z ∈ C. The sum
n=1
n!
converges, hence is defined for all z ∈ C.

(a) It’s continuously bijective and strictly increasing on R.

(b) ea+b = ea · eb and ex ̸= 0 for all x ∈ C.
(c) The derivative of ex is ex .
(d) Define log(x) = m ∈ C such that em = x.
(e) The function log is continuous and increasing on (0, +∞).
Z y Z 1
1 1
(f) When y > 1, log(y) = dx. If 0 < y < 1, then − log(y) = dx.
1 x y x

24. (Trigonometric Functions) Define E(ix) = cos(x)+i sin(x), which is equivalent to defining
eix + e−ix eix − e−ix
cos(x) = and sin(x) = , and both are real-valued functions.
2 2i
(a) Geometric properties hold. For example, the Pythagorean sin2 (x) + cos2 (x) = 1.
(b) (sin x)′ = cos(x) and (cos x)′ = − sin(x).
(c) eix has period 2π, hence sin(x) and cos(x) are both 2π-periodic.
(d) The function eix is injective and continuous over [0, 2π).

25. (Weak Fundamental Theorem of Algebra) Suppose n is a positive integer and a0 , . . . , an

are complex coefficients with an ̸= 0. Then, P (z) = a0 + a1 z + · · · + an z n has a root
z ′ ∈ C, a complex number such that P (z ′ ) = 0.

26. (Precursor to Fourier Series) A trigonometric polynomial is a function of the form

N
X N
X
f (x) = a0 + (an cos(nx) + bn sin(nx)), and can be rewritten as f (x) = cn einx . A
n=1 n=−N
trigonometric series is the limit of a trigonometric polynomial as N → +∞.

(a) Define the inner product of two integrable functions f, g : [a, b] → C as ⟨f, g⟩ =
Z b
p
f (x)g(x) dx. Define the L2 -norm of a function f as ⟨f, f ⟩.
a
 
1 inx
(b) With this definition, √
2π
e form an orthonormal basis for continuous func-
n∈Z
tions on [−π, π].
(c) The fourier coefficients represent the coordinates of the projection of f against
the space formed by the orthonormal basis (einx ) above. Therefore, Z π the cn is the
1
n-th coefficient of the vector ⟨f, einx ⟩ · einx , which means cn = f (x)e−inx dx
2π −π

5
 ∞
X ∞
X
inx
27. (Fourier Series) If cn e is the Fourier series of f , denote it as f (x) ∼ cn einx .
n=−∞ n=−∞

(a) If f : [−π, π] → C is integrable, then as n → ∞, both

Z π Z π
f (x) cos(nx) dx → 0 and f (x) sin(nx) dx → 0.
−π −π

(b) Suppose f : R → C is 2π-periodic and f ∈ Rπ−π . Fix x0 ∈ R. Assume there are

δ > 0 and M > 0 such that ∀t ∈ (−δ, δ), |f (x0 + t) − f (x0 )| ≤ M |t|. Then the
N
X
partial fourier series SN (x0 ) = cm e−imx0 converges to f (x0 ) as N → ∞.
m=−N

(c) (Localization Theorem) If ∀t ∈ (x0 − δ, x0 + δ), f (t) = 0, then ∀t ∈ (x0 − δ, x0 + δ),

lim SN (t) = 0.
N →∞
(This shows that pointwise convergence on a fourier series can only be proven locally
(on small neighborhoods), but they may not converge pointwise outside of those
neighborhoods despite their fourier coefficients being convergent.)
(d) (Trigonometric Polynomial form of Stone-Weierstrass) If R → C is 2π-periodic and
continuous, then for any ε > 0 there is a (finite) trigonometric polynomial such that

|P (x) − f (x)| < ε.

(e) (Parseval’s Theorem) Suppose f, g : [−π, π] → C are Riemann integrable and

∞
X ∞
X
inx
f (x) ∼ cn e , g(x) ∼ γn einx ,
n=−∞ n=−∞

Then:
2
i. (Partial sum converges in Ls -norm)
Z π
1
lim ||f − SN (f )||2 = lim |f (x) − SN (f )(x)|2 dx = 0
N →∞ N →∞ 2π −π
ii. (Isometry between inner product with fourier coefficients)
Z π ∞
1 X
⟨f, g⟩ = f (x)g(x) dx = cn γn = ⟨⃗c, ⃗γ ⟩
2π −π n=−∞
Z π ∞
1 2
X
2
iii. (Isometry between norms) ||f ||2 = |f (x)| dx = |cn |2 = ||⃗c||22 .
2π −π n−∞