Analysis I

Cheat Sheet · v1.0 · 2020

Flavio Schneider

D-INFK
ETH Zürich
 1 Algebra (vii) |xn | = |x|n 2 Functions · Horizontal aysmptote (if domain is unlimited at
±∞) if:
 
|x|
(viii)  x = |y|
 
1.1 Exponential Properties y 2.1 Domain limx→+∞ f (x) = k (right y = k)
(ix) |a − b| = b − a, if a ≤ b limx→−∞ f (x) = h (left y = h).
(i) x0 = 1 · Fractions denominator 6= 0.
(x) |a + b| ≤ |a| + |b| · Oblique aysmptote (if domain is unlimited at
(ii) xn xm = xn+m
(xi) |a| − |b| ≤ |a − b| · Logarithms if the base is a number, the argument ±∞) if:
xn 1
(iii) xm
= xn−m = xm−n
must be > 0, if the base depends on a variable, the f (x)
limx→+∞ x = m ∧ limx→+∞ [f (x) − mx] = q
(iv) (xn )m = xnm base must be > 0∧ 6= 1. (right at y = mx + q)
1.5 Factorization
 n f (x)
x xn · Roots with even index, the argument must be limx→−∞ x = m ∧ limx→−∞ [f (x) − mx] = q
(v) y
= yn (i) x2 − a2 = (x + a)(x − a) ≥ 0, for roots with odd index the domain is R. (left at y = mx + q).
(vi) x−n = 1 (ii) x2 + 2ax + a2 = (x + a)2
xn · Arccos/Arcsin the agrument must be ∈ [−1, 1].
(vii) 1
= xn (iii) x − 2ax + a2 = (x − a)2
2
2.6 Monotonicity
x−n For other trig functions we use trig properties to
 −n (iv) x2 + (a + b)x + ab = (x + a)(x + b) change them to cos and sin.
x y
n yn A funciton f is:
(viii) y
= x = xn (v) x3 + 3ax2 + 3a2 x + a3 = (x + a)3
 1 n √ · Exponential base > 0. · Monotonically increasing if:
(ix) x
n
m = xm
1
= (xn ) m = m
xn (vi) x3 − 3ax2 + 3a2 x − a3 = (x − a)3
∀x, y : x ≤ y ⇒ f (x) ≤ f (y)
(vii) x3 + a3 = (x + a)(x2 − ax + a2 )
2.2 Parity · Monotonically decreasing if:
(viii) x3 − a3 = (x − a)(x2 + ax + a2 )
1.2 Logarithm Properties ∀x, y : x ≤ y ⇒ f (x) ≥ f (y)
(ix) x2n − a2n = (xn − an )(xn + an ) We consider the partiy of the function only if
(i) logn (0) = Undefined Dom(f ) is mirrored on the origin: · Strictly increasing if:
(ii) logn (1) = 0 (Dom(f ) = [−2, 2] ∨ (−∞, ∞) ∨ (−∞, −1] ∪ [1, ∞]). ∀x, y : x < y ⇒ f (x) < f (y)
1.6 Complete The Square
(iii) logn (n) = 1 · Strictly decreasing if:
ax2 + bx + c = 0 ⇒ a(x + d)2 + e = 0 · Even function (with respect to the y axis) if:
∀x, y : x < y ⇒ f (x) > f (y)
(iv) logn (nx ) = x f (−x) = f (x).
b
(v) nlogn (x) = x · d= 2a · Odd function (with respect to the origin) if: 2.7 Max, Min
(vi) logn (xr ) = r logn (x) 6= logrn (x) = (logn (x))r 2 f (−x) = −f (x).
b
(vii) logn (xy) = logn (x) + logn (y)
· e=c− 4a Calculate f 0 (x) = 0, then all the solutions xi are
  · In every other case the function is neither even nor our candidates, where for a small  > 0:
(viii) logn x y
= logn (x) − logn (y) 1.7 Quadratic Formula
odd.
· Max if: f 0 (xi − ) > 0 ∧ f 0 (xi + ) < 0.
(ix) − logn (x) = logn x1

√
ax2 + bx + c = 0 ⇒ x=
−b± b2 −4ac
2.3 Axis Intercept · Min if: f 0 (xi − ) < 0 ∧ f 0 (xi + ) > 0.
log(x) 2a
(x) log(n)
= logn (x) · Inflection if (use sign table):
· If b2 − 4ac > 0 ⇒ Two real unequal solutions. · X intercept can be many; is calculated by solving
g(x) f 0 (xi − ) < 0 ∧ f 0 (xi + ) < 0, or
f (x) = 0. If f (x) = h(x) we solve just g(x) = 0. f 0 (xi − ) > 0 ∧ f 0 (xi + ) > 0
1.3 Radical Properties · If b2 − 4ac = 0 ⇒ Two repeated real solutions. The points are then (xi , 0).
√
n
1 If f 0 (x) > 0, then f is strictly increasing.
(i) x=x n
· If b2 − 4ac < 0 ⇒ Two complex solutions.
√ √ √ · Y intercept can be just one; is calculated by If f 0 (x) < 0, then f is strictly decreasing.
(ii) n xy = n x n y setting x = 0, the point is then (0, f (0)). If If f 0 (x) = 0 f is constant.
p √ √ x=0∈ / Dom(f ) there is no Y intercept.
(iii) m m x = mn x
√
q
(iv) n x
nx
= n
√
2.8 Convexity
y y
√ 2.4 Sign · Convex (∪) if: f 00 (x) > 0
n n
(v) x = x, if n is odd
√ The sign can only change when there is a x inter- · Concave (∩) if: f 00 (x) < 0
(vi) n xn = |x|, if n is even cept (if the function is continous), thus if we solve
f (x) ≥ 0 we get both the X intercepts and where
2.9 Inflection Points
1.4 Absolute Value Properties the function is positive.
( Calculate f 00 (x) = 0, then all the solutions xi are
x if x ≥ 0 our candidates (except where f (x) is not defined),
(i) |x| = 2.5 Asymptotes/Holes
−x if x < 0 where for a small  > 0:
(ii) |x| ≥ 0 · Hole at point (x0 , fsemplified (x0 )) if plugging the · Increasing Inflection if:
(iii) | − x| = |x| critical point x0 in the numerator of f gives 00 . f 00 (xi − ) < 0 ∧ f 00 (xi + ) > 0
(iv) |ca| = c|a|, if c > 0 · Vertical asymptote at a critical point x0 if: · Decreasing Inflection if:
(v) |xy| = |x||y| limx→x− f (x) = ±∞ (left at x = x0 ) f 00 (xi − ) > 0 ∧ f 00 (xi + ) < 0
0
(vi) |x2 | = x2 limx→x+ f (x) = ±∞ (right at x = x0 ). · Otherwise nothing happens on xi .
0
Flavio Schneider Analysis I · Cheat Sheet Page 1
3 Trigonometry 3.5 Reciprocal Identities 3.12 Half-Angle Identities 3.19 Degrees
1 q
(i) cot(x) = x 1−cos(x)
3.1 Unit Circle


tan(x) (i) sin 2
=± 2
y
1
(ii) csc(x) = sin(x)
q
1+cos(x)
x


(ii) cos 2
=± 2
(0, 1)
1
(iii) sec(x) = cos(x) q 
− 12 ,
√
2
3
 
1
√
2, 2
3


x

1−cos(x)
(iii) tan 2
=± 2

−
√
2
√ 
2
√
2
√ 
2
2 , 2 π 2 , 2
2

3.6 Quotient Identities (iv) tan x

=
1−cos(x)  √ 
2π
3
π
3 √ 
2 sin(x) − 3 1
2 , 2
3π
4 90◦
π
4
3 1
2 , 2

sin(x) sin(x)
120◦ 60◦
(i) tan(x) = cos(x) (v) tan x


= 5π
6
π
6
2 1+cos(x) 150◦ 30◦
cos(x)
(ii) cot(x) = sin(x) (−1, 0) (1, 0)
3.13 Sum-to-Product Formulas π 180◦ 0◦ ◦
360 2π x
   
3.7 Sum Identities (i) sin(x) + sin(y) = 2 sin x+y
2
cos x−y
2 210◦ 330◦
7π 11π
(i) sin(x + y) = sin(x) cos(y) + cos(x) sin(y)
    6 6
x−y x+y
(ii) sin(x) − sin(y) = 2 sin cos 2 2

−
√
3 1
 5π
240 ◦

270◦
300 ◦
7π
√
3 1

2 , −2 2 , −2
(ii) cos(x + y) = cos(x) cos(y) − sin(x) sin(y)    
4
4π 5π
4

3 3

(iii) tan(x + y) =
tan(x)+tan(y) (iii) cos(x) + cos(y) = 2 cos x+y2
cos x−y 2

−
√
2
√ 
2 ,− 2
2
3π
2
√
2
√ 
2 ,− 2
2

1−tan(x) tan(y)      √   √ 
− 12 , − 3 1 3

(iv) cos(x) − cos(y) = −2 sin x+y 2

cos x−y 2
2 2, − 2

(0, −1)
3.8 Difference Identities
(i) sin(x − y) = sin(x) cos(y) − cos(x) sin(y) 3.14 Product-to-Sum Formulas
(ii) cos(x − y) = cos(x) cos(y) + sin(x) sin(y) (i) sin(x) sin(y) = 1
[cos(x − y) − cos(x + y)]
2
tan(x)−tan(y)
(iii) tan(x − y) = 1+tan(x) tan(y) (ii) 1
cos(x) cos(y) = 2 [cos(x − y) + cos(x + y)]
3.2 Domain and Range
(iii) sin(x) cos(y) = 12 [sin(x + y) + sin(x − y)]
· sin : R −→ [−1, 1] 3.9 Double Angle Identities
(iv) cos(x) sin(y) = 12 [sin(x + y) − sin(x − y)]
· cos : R −→ [−1, 1] (i) sin(2x) = 2 sin(x) cos(x)
 
· tan : x ∈ R  x 6= π

+ kπ −→ R (ii) cos(2x) = cos2 (x) − sin2 (x) 3.15 Tangent expression
2
cos(2x)+1
(iii) cos(2x) = 2 cos2 (x) − 1 ⇒ cos2 (x) = h i
· cot : {x ∈ R | x 6= kπ} −→ R 2
If u = tan( x2 ) : dx = 2
du
1−cos(2x) 1+u2
(iv) cos(2x) = 1 − 2 sin2 (x) ⇒ sin2 (x) = 2
· csc : {x ∈ R | x 6= kπ} −→ R \ (−1, 1)
2 tan(x) 1−u2
(v) tan(2x) = (i) cos(x) = 1+u2
· sec : x ∈ R  x 6= π2 + kπ −→ R \ (−1, 1) 1−tan2 (x)
 
2u
(ii) sin(x) = 1+u 2
· sin−1 : [−1, 1] −→ − π2 , π2
 
3.10 Co-Function Identities (iii) 2u
tan(x) = 1−u2
π
· cos−1 : [−1, 1] −→ [0, π]


(i) sin 2
− x = cos(x)
π
3.16 Hyperbolic Functions


· tan−1 : R −→ − π2 , π2
  (ii) cos 2
− x = sin(x)
π


(iii) tan − x = cot(x) ex −e−x
2 (i) sinh(x) = 2
3.3 Pythagorean Identities (iv) cot π


− x = tan(x)
2 ex +e−x
(ii) cosh(x) =
(i) sin2 (x) + cos2 (x) = 1 π 2


(v) csc 2
− x = sec(x)
ex −e−x
(ii) tan2 (x) + 1 = sec2 (x) π

(iii) tanh(x) =
(vi) sec 2
− x = csc(x) ex +e−x
(iii) 1 + cot2 (x) = csc2 (x)
3.11 Even-Odd Identities 3.17 Laws of Sines
3.4 Periodicity Identities sin(α) sin(β) sin(γ)
(i) sin(−x) = − sin(x) (i) a
= b
= c
(i) sin(x ± 2π) = sin(x)
(ii) cos(−x) = cos(x)
(ii) cos(x ± 2π) = cos(x)
(iii) tan(−x) = − tan(x) 3.18 Laws of Cosines
(iii) tan(x ± π) = tan(x)
(iv) cot(x ± π) = cot(x) (iv) cot(−x) = − cot(x) (i) a2 = b2 + c2 − 2bc cos(α)
(v) csc(x ± 2π) = csc(x) (v) csc(−x) = − csc(x) (ii) b2 = a2 + c2 − 2ac cos(β)
(vi) sec(x ± 2π) = sec(x) (vi) sec(−x) = sec(x) (iii) c2 = a2 + b2 − 2ab cos(γ)
Flavio Schneider Analysis I · Cheat Sheet Page 2
4 Limits, Sup and Inf 4.5 E-Log Trick 4.7 Strategy 5 Continuity
g
lim f = lim e g ln(f ) Given a limit lim f (x):
x→x0
x→x0 x→x0
Definition 4.1 (Limits). Let f (x) be a func- Definition 5.1 (Pointwise Continous).
tion defined on an open interval that contains 1. Is f (x0 ) solvable normally (polynomials and A function f : [a, b] → R is pointwise conti-
x = x0 , except possibly x = x0 , then the limit Theorem 1: L’Hospital’s Rule radicals) ? nous at x0 ∈ [a, b] if limx→x0 f (x) = f (x0 ),
limx→x0 f (x) = L exists if and only if for all 2. Try to decompose the limit with the properties or:
f (x) 0 ±∞
 there is a δ such that: If by plugging x0 in g(x)
we get 0
or ±∞
, and go back to step 1 for each piece.
∀x0 , ∃δ ∀x : (|x−x0 | < δ ⇒ |f (x)−f (x0 )| < )
then:
0 < |x − x0 | < δ ⇒ |f (x) − L| <  ∀x 3. If it contains a radical expession try with the
f (x) f 0 (x) root trick, pay attention that if it’s not a square
Sequence Definition: lim = lim 0 = L ⇔ L 6= ±∞ root you can try with the third root factoriza-
x→x0 g(x) x→x0 g (x)
Definition 5.2 (Uniformly Continous).
lim f (x) = L ⇔ ∀(xn ) where lim xn = x0 , then tion, but for bigger roots it’s probabily another
x→x0 n→∞ method. If the root contains the entire limit it A function f : [a, b] → R is uniformly con-
lim f (xn ) = L can be put out (PR6). tinous if it’s continous at every point in it’s
n→∞
Theorem 2: Limit Squeeze Theorem domain ∀x0 ∈ [a, b] : limx→x0 f (x) = f (x0 ),
4. If conatins a trigonometric function try with or:
4.1 Limit Properties the Squeeze Theorem, if the trig function con-
Let lim f (x), if g(x) ≤ f (x) ≤ h(x), ∀x, and tains another function go with the composed
Assume that limx→x0 f (x) and limx→x0 g(x) exists
x→x0 ∀∃δ ∀x0 , x : (|x−x0 | < δ ⇒ |f (x)−f (x0 )| < )
lim g(x) = lim h(x) = L, then: function decomposition. If it simplifies well
ant that c ∈ R, then: x→x0 x→x0
with the series definition of cos, sin, or tan try
(i) lim [cf (x)] = c lim f (x) lim f (x) = L to simplify the sum and solve each piece.
x→x0 x→x0 x→x0 Definition 5.3 (Lipschitz Continous).
5. If it’s a composed function try the chain rule. A function f : [a, b] → R is Lipschitz conti-
(ii) lim [f (x) ± g(x)] = lim f (x) ± lim g(x)
x→x0 x→x0 x→x0 nous if:
6. If it’s raised to an unusual power try the E-
(iii) lim [f (x)g(x)] = lim f (x) lim g(x)
x→x0 x→x0 x→x0
4.6 Important Limits Log trick.
0 ±∞ ∃L∀x, x0 : |f (x) − f (x0 )| ≤ L|x − x0 |
lim f (x)
 x n 7. If you get or use l’Hopital.

f (x)

x→x0 · lim 1+ = ex 0 ±∞
(iv) lim = , lim g(x) 6= 0 n→∞ n 8. If you get ±∞ · 0 or 0 · ±∞ tranform the func-
x→x0 g(x) lim g(x)
x→x0
x→x0
an − 1 tion into a fraction so that you get 00 or or ±∞
5.1 Properties
n · lim = ln(a) ±∞
then use l’Hopital. Let f and g be continous, then also f ± g, f · g,

n→0 n
(v) lim [f (x)]n = lim f (x) f
⇔ g 6= 0 and f ◦ g are continous.
x→x0 x→x0 · lim ln(n) = ∞ g
n→∞
hp
n
i r 4.8 Supremum and Infimium (i) Polynomials: All polinomials P (x) are uni-
(vi) lim f (x) = n lim f (x)
x→x0 x→x0 loga (1 + n) 1 formly continous in their entire domain.
· lim =
(vii) lim x = x0 n→∞ n ln(a) Definition 4.2. (ii) Bijective: If f : [a, b] → R is continous and
x→x0
loga (1 + n) 1 · The Supremum of a set S denoted sup(S) = monotone, then it’s bijective and f −1 is also
· lim = continous.
n→0 n ln(a) u is a number u that satisfies the condition
4.2 Chain Rule
that u is an upper bound of S and for any
Let f and g be continous, and given · lim sin(n) = 1 upper bound v of S, u ≤ v. Theorem 3: Intermediate Value
limx→x0 f (g(x)) of composed function we can n→0 n
· The Infimium of a set S denoted inf (S) = u
solve limx→x0 g(x) = y0 , then: 1 − cos(n) is a number u that satisfies the condition that Let f be a continous function on [a, b] and let
· lim =0 s be a number with f (a) < s < f (b), then
n→0 n u is an lower bound of S and for any lower
lim f (g(x)) = lim f (y) bound v of S, u ≥ v. there exists at least one solution to f (x) = s.
x→x0 y→y0
1 − cos(n) 1
· lim =
n→0 n2 2 · If the supremum doesn’t exists we can write:
4.3 Exponential Rule Theorem 4: Extreme Value
tan(n) sup(S) = ∞.
Let f and g be continous, where limx→x0 f (x) = · n→0
lim
n
=1
· If the infimium doesn’t exist we can write: Let f : I → R be a continous function on
f (x0 ) > 0 and limx→x0 g(x) = g(x0 ) (where both
n! inf (S) = −∞. I = [a, b] then there exist two numbers c ∈ I
limits exists), then: · lim n = 0
n→∞ n · To prove that the minimum doesn’t exist: and d ∈ I such that:
lim f (x)g(x) = f (x0 )g(x0 ) en − 1 ∀∃n0 ∈ N : f (x) ≤ inf (a) +  ∀x ≥ n0 .
x→x0 · lim =1 ∀x ∈ I : m = f (c) ≤ f (x) ≤ f (d) = M
n→0 n · To prove that the maximum doesn’t exist:
√
n ∀∃n0 ∈ N : f (x) ≤ sup(a) −  ∀x ≥ n0 . Where m is a lower bound and M an upper
4.4 Root Trick · lim n! = ∞ bound.
√ n→∞
p p f +g √
lim f − g = lim f −g· √ · lim n n = 1
x→x0 x→x0 f +g n→∞
Flavio Schneider Analysis I · Cheat Sheet Page 3
 6 Derivatives · d
dx
(sinh(x)) = cosh(x) Theorem 7: Mean Value Theorem
d
· dx
(cosh(x)) = sinh(x) Let f : [a, b] −→ R be continous and differen-
Definition 6.1 (Derivative). The derivative
of f (x) with respect to x is: d 1 2 tiable on ]a, b[, then exists c ∈ ]a, b[ with:
· dx
(tanh(x)) = cosh(x)
= 1 − tanh (x) f (b) = f (a) + f 0 (c)(b − a), or:
df f (x + h) − f (x)
= f 0 (x) = lim · d
(sinh−1 (x)) = √ 1
f (b) − f (a)
dx h→0 h dx x2 +1 f 0 (c) =
f (x0 ) − f (x) b−a
= lim · d
dx
(cosh−1 (x)) = √ 1
x0 →x x0 − x x2 −1

· d
dx
(tanh−1 (x)) = √ 1
1−x2
6.1 Properties
d
(i) dx (c) = 0
(ii) (cf )0 = cf 0 (x) 6.3 Differentiable
(iii) (f ± g)0 = f 0 (x) + g 0 (x) Theorem 5: Differentiable
(iv) (f g)0 = f 0 g + f g 0
 0 0
g0
A function f is differentiable at a point x0 iff:
(v) fg = f g−f g2
d f (x) − f (x0 ) f (x) − f (x0 )
(vi) dx
([f (x)]n ) = n[f (x)]n−1 f 0 (x) lim = lim
x→x+
0
x − x0 x→x0− x − x0
d
(vii) dx
(f (g(x)) = f 0 (g(x))g 0 (x)
1
(viii) [f ]0 (x) = f 0 (f −1
−1
(x))
(i) Tangent Line: The function f has a tangent
point at a if and only if f is differentiable at a.
6.2 Common Derivatives The equation of the tangent line is:
d
· dx
(x) =1
d y = f 0 (a)(x − a) + f (a)
· dx
(|x|) = sign(x)
d
· dx
(ex ) = ex
(ii) Continous: If f (x) is differentiable at a, then
d
· dx
(ax ) = ax ln(a) f is continous at a. The converse is not true
d 1 (e.g. f (x) = |x|, a = 0).
· ( )
dx x
= − x12
d √ 1
(iii) Classes: If f : [a, b] → R is differentiable k
· dx
x= 2
√
x times we say that f ∈ C k ([a, b]) where C is
d f 0 (x) 1
called classification function. If f is differen-
· dx
(ln(f (x))) = f (x)
= x
if f (x) = x tiable infinite times we say that f is smooth
d 1 (f ∈ C ∞ ([a, b])).
· dx
(ln |x|) = x
, x 6= 0
d 1
· dx
(loga (x)) = xln(a)
, x>0

· d
(sin(x)) = cos(x) Theorem 6: Inverse Function Theorem
dx
d Let f : [a, b] −→ R be continous, differen-
· dx
(cos(x)) = − sin(x)
d
tiable and stricly increasing where ∀x ∈ [a, b] :
· dx
(tan(x)) = sec2 (x) = tan2 (x) + 1 f 0 (x) > 0 and
d
· dx
(cot(x)) = − csc2 (x)
c= inf f (x) < sup f (x) = d
d a<x<b a<x<b
· dx
(sec(x)) = sec(x) tan(x)
· d
(csc(x)) = − csc(x) cot(x) then:
dx

· d
(sin−1 (x)) = √ 1 · f : ]a, b[ → ]c, d[ is bijective.
dx 1−x2
· f −1 : ]c, d[ → ]a, b[ is differentiable with
d 1
· dx
(cos−1 (x)) = −√ 1
[f −1 ]0 (x) = f 0 (f −1
1−x2 (x))
d 1
· dx
(tan−1 (x)) = 1+x2
Flavio Schneider Analysis I · Cheat Sheet Page 4
7 Integrals 7.1 Properties ax+b ax ad−bc
R R
· csc(x) cot(x)dx = − csc(x) · cx+d
dx = c
− c2
ln |cx + d|
Ra  
(i)  1−cos(x)  1 1
a f (x)dx = 0
R 1 R
· sin(x)
dx = ln  sin(x)  · (x+a)2
dx = − x+a
Definition 7.1 (Riemann-Integral). Given: (ii)
Rb Rb
a cf (x) = c a f (x) 1 1
  R
Rb Rb Rb ·
R 1  1+sin(x) 
dx = ln  cos(x)  · ax+b
dx = a
ln |ax + b|
· A continous function f (x) : [a, b] −→ R cos(x)
(iii) a f (x) + g(x)dx = a f (x)dx ± a g(x)dx 1 1 x
tan−1
R

R 1 · a2 +x2
dx =
· A partition P ..= {a = x0 , ..., xn−1 , xn = b}
Rb Rb · dx = − cot(x) a a
(iv) a f (x)dx = − a f (x)dx sin2 (x)  
where Ii = [xi−1 , xi ] 1 2
tan−1 √2ax+b
R
= √
Rb Rc Rb 1 · dx
R
(v) a f (x)dx = a f (x)dx + c f (x)dx
· cos2 (x)
dx = tan(x) ax2 +bx+c
· A set of points ξ ..= {ξ1 , ..., ξn } where 4ac−b2 4ac−b2
Rb
ξi ∈ Ii = [xi−1 , xi ]. (vi) a cdx = c(b − a) − cos(x)
 
1
R
· dx = 1 1
ln  x−a
R
· dx =
 
1+sin(x) 1+sin(x) (x−a)(x−b) a−b x−b 
(vii) If f (x) ≥ g(x), then:
Then the Riemann-Sum is defined as: Rb Rb R 1 sin(x) x 1
 2 
ln a + x2 
R
a f (x) ≥ a g(x) · dx = · dx =
1+cos(x) 1+cos(x) a2 +x2 2
n (viii) If m ≤ f (x) ≤ M , then: cos(x) x2
X
1 x
R
= x − a tan−1
R

S(f, P, ξ) ..= f (ξi ) · (xi − xi−1 ) · dx = · dx
m(b − a) ≤ ab f (x)dx ≤ M (b − a)
R
1−sin(x) 1−sin(x) a2 +x2 a
i=1
− sin(x) x3
R  R
1 1 2
If  ab f (x)dx ≤ ab |f (x)|dx − 12 a2 ln a2 + x2 
R R  
(ix)
  · 1−cos(x)
dx = 1−cos(x) · a2 +x2
dx = 2
x
Where the Riemann-Integral is:
x a
R
sin(ax)dx = − a1 cos(ax) · + ln |a + x|
R
· (x+a)2
dx = a+x
b n
7.2 Common Integrals (+C)
Z X
..
f (x)dx = lim f (ξi ) · (xi − xi−1 ) 1
R x 1
 
ln ax2 + bx + c
R
· cos(ax)dx = a
sin(ax) · = −
a n→∞ ax2 +bx+c 2a
i=1 Basic 
− a1
R
R · tan(ax)dx = ln(cos(ax)) √ b
tan−1 √2ax+b
· kdx = kx a 4ac−b2 4ac−b2
(i) Over Sum: n+1 ·
R
x sin(ax)dx = − a1 x cos(ax) + 1
sin(ax)
a2 Square Roots
n
x dx = xn+1 , n 6= −1
R n
X ·
S(f, P ) ..= lim sup f (x) · (xi − xi−1 ) 1 1 R√
R
· x cos(ax)dx = x sin(ax) + cos(ax) 3
n→∞
i=1 x∈Ii ·
R 1 −1
= (n−1)x
a a2 · x − adx = 23 (x − a) 2
xn n−1 R
Infimum of the over sum: · sinh(x)dx = cosh(x) R√ 
2b
√
1X
n
·
R −1 R 1
x dx = x dx = ln |x| R · ax + bdx = 3a + 2x
3
ax + b
inf S(f, P ) ..= (b − a) · lim sup f (x) · cosh(x)dx = sinh(x) R√ √ √
P n→∞ n ax
i=1 x∈Ii x2 + adx = 21 x x2 + a + a2 ln |x + x2 + a|
R x
· a dx = ln(a) ·
R
tanh(x)dx = ln(cosh(x)) ·
(ii) Under Sum: R√ √ 2
a2 − x2 dx = 12 x a2 − x2 + a2 sin−1 x
R x
e dx = ex
R

n
X · · coth(x)dx = ln | sinh(x)| · a
S(f, P ) ..= lim inf f (x) · (xi − xi−1 ) R √ R √
· loga (x)dx = x loga (x) − x loga (e) −1 −1 3 5
R
n→∞ x∈Ii · sinh (x)dx = x sinh (x) − x2 +1 · x x − adx = 23 a(x − a) 2 + 25 (x − a) 2
i=1
Supremum of the under sum: Trigonometirc √
cosh−1 (x)dx = x cosh−1 (x) − R √
R
· x2 − 1 3
1X
n R · x x2 ± a2 dx = 13 (x2 ± a2 ) 2
supS(f, P ) ..= (b − a) · lim · sin(x)dx = − cos(x)
inf f (x) tanh−1 (x)dx = x tanh−1 (x) + 1
ln(1 − x2 )
R
n→∞ n x∈Ii
· 2
R 3
2 5
P · (ax + b) 2 dx = 5a (ax + b) 2
R
i=1 · cos(x)dx = sin(x)
coth−1 (x)dx = x coth−1 (x) + 1
ln(x2 − 1)
R

(i − 1)(b − a) i(b − a)
 R · 2
 √ 
· tan(x)dx = − ln | cos(x)| = ln | sec(x)| √ 1
R
Ii = a + ,a + · dx = ln x + x2 ± a2 
 
n n R Logarithmic 2 2
x ±a
· cot(x)dx = ln | sin(x)|
(iii) Inequality: 1
= sin−1 x
R

√
R
R · ln(ax)dx = x ln(ax) − x · dx a
sup S(f, P ) ≤ inf S(f, P ) · sec(x)dx = ln | sec(x) + tan(x)| a2 −x2
p∈P (I) p∈P (I) 2
· x ln(ax)dx = x4 (2 ln(ax) − 1)
R √
√1
R R
· csc(x)dx = − ln | csc(x) + cot(x)| · x±a
dx =2 x±a
(iv) Monotone: A monotone function f : I → R √ R ln(ax) √
sin−1 (x)dx = x sin−1 (x) + 1 − x2 dx = 12 (ln(ax)2
R
is Riemann-Integrable over I. · · x ·
R
√ x
dx = x2 ± a2
√ x2 ±a2
cos−1 (x)dx = x cos−1 (x) − 1 − x2
R
(v) Continous: A continous function f : I → R · Exponential
is Riemann-Integrable over I. √ Other
e dx = a1 eax
R ax
tan−1 (x)dx = x tan−1 (x) − 12 ln(1 + x2 ) ·
R
·
x sin(ax)dx = − a1 x cos(ax) + a12 sin(ax)
R
√ ·
· xe dx = (x − 1)ex
R x
cot−1 (x)dx = x cot−1 (x) + 12 ln(1 + x2 )
R
·
x cos(ax)dx = a1 x sin(ax) + a12 cos(ax)
R
Theorem 8: Riemann-Integrable   ·
sin2 (x)dx = 12 (x − sin(x) cos(x)
R
· · xeax dx = x − a12 eax
R
1
R bx bx (b sin(ax) − a cos(ax))
a · e sin(ax)dx = a2 +b
A function f is Riemann-Integrable iff: 2e
cos2 (x)dx = 21 (x + sin(x) cos(x)
R
· Rational Functions 1
R bx bx (a sin(ax) + b cos(ax))
supS(f, P ) = inf S(f, P ) √ · e cos(ax)dx = a2 +b 2e
tan2 (x)dx = tan(x) − x
R R 1
P1 P2 · · √ =2 x
x
cot2 (x)dx = − cot(x) − x
R
· (x+a)n+1
(x + a)n dx =
R
More formally: · , n 6= −1
sec2 (x)dx = tan(x) n+1
R
·
∀∃P : |S(f, P ) − S(f, P )| <  (x+a)n+1 ((n+1)x−a)
csc2 (x)dx = − cot(x) x(x + a)n dx =
R R
· · (n+1)(n+2)
Flavio Schneider Analysis I · Cheat Sheet Page 5
 √
7.3 U-Substitution (ii) b2 x2 − a2 =⇒ x = ab sec(u), with property · Absolute
R∞ Converpgence: R∞
The substitution, u = g(x), du = g 0 (x)dx is:
tan2 (x) = 1 − sec2 (x). a |f (x)|dx converges ⇒ a f (x) converges
√
Z b Z g(b) (iii) a2 + b2 x2 =⇒ x = ab tan(u), with property
f (g(x))g 0 (x)dx = f (u)du sec2 (x) = 1 + tan2 (x).
a g(a) Definition 7.2 (Antiderivative).
Let f : [a, b] → R be a function where
7.4 Integration By Parts 7.7 Rational Functions
Z b Z b Given an integral
R P (x)
dx: f (x) = F 0 (x) ∀x ∈ [a, b]
Q(x)
0 0
f (x)g (x)dx = [f (x)g(x)]ba − f (x)g(x)dx
a a · For deg(P (x)) ≥ deg(Q(x)), then apply a polyno- then F is called the antiderivative of f .
· u = f (x), v = g(x) mial division so that we get an equivalent integral
R R(x) R R(x)
A(x) + Q(x) dx where Q(x) dx is easier to solve.
· du = f 0 (x)dx, dv = g 0 (x)dx Theorem 9: Mean Value Theorem
R R · For deg(P (x)) < deg(Q(x)), then factor Q(x) as
[ udv = uv − vdu]. As a rule of thumb use the completely as possible and find the partial fraction (Integration) Let f be continous on [a, b], then
following order, u should be the function that comes decomposition (P.F.D) of the rational expression. there exists a c such that:
first beween: Logarthmic ↔ Inverse trig. → Alge- Z b
braic (Axn ) → Trigonimetric → Exponential (kx ). 1. Q(x) = (ax + b)(cx2 + dx + e), then the P.F.D. 1
f (c) = f (x)dx
is: b−a a
A B
7.5 Trig-Function Trick ax+b
+ cx2 +dx+e
(f (c) = favg )
For sinn (x) cosm (x)dx evaluate the following:
R 2. Q(x) = (ax + b)n , then the P.F.D. is:
A1 A2 An
ax+b
+ (ax+b) 2 + · · · + (ax+b)n
(i) Deg(n) odd: strip one sin out and convert
Theorem 10: Fundamental Theorem
the rest to cos with sin2 (x) = 1 − cos2 (x), then
use substitution on u = cos(x). 7.8 Improper Integrals
Part 1: Suppose that f is continous on [a, b]
and F is defined as: F (x) ..= ax f (t)dt, then F
R
(ii) Deg(m) odd: strip one cos out and convert Convergent, if lim = k with k finite.
the rest to sin with cos2 (x) = 1 − sin2 (x), then Divergent, if lim = ±∞ ∨ D.N.E. is differentiable on ]a, b[ and for all x ∈ ]a, b[:
use substitution on u = sin(x).
(iii) Deg(n) and Deg(m) both odd: use either Infinite Limit: F 0 (x) = f (x)
(i) or (ii).
(i) a∞ f (x)dx = limt→∞ at f (x)dx
R R
Part 2: Suppose that f is continous on [a, b]
(iv) Deg(n) and Deg(m) both even: use dou- and F is the antiderivative of f , then:
(ii) −∞ f (x)dx = limt→−∞ tb f (x)dx
Rb R
ble angle and/or half angle trig identities to
R∞
f (x)dx + c∞ f (x)dx
Rc R b
reduce the integral.
Z
(iii) −∞ f (x)dx = −∞
provided that both integrals are convergent. f (x)dx = F (b) − F (a)
For tann (x) secm (x)dx evaluate the following:
R
a

Discontinous Integrand:
(i) Deg(n) odd: strip one tan and one sec out,
and convert the rest to sec using tan2 (x) = (i) Discontinuity at a:
7.9 Derivative of Integrals
sec2 (x) − 1, then use substitution on u =
Rb Rb
a f (x)dx = limt→a+ t f (x)dx
If we have to evaluate the derivative of an integral
sec(x). where F (x) = ag(x) f (t)dt, then by the first part
R
(ii) Discontinuity at b:
(ii) Deg(m) even: strip 2 sec out and convert Rb Rt of the Fundamental Theorem of Calculus (and the
the rest to tan with sec2 (x) = 1+tan2 (x), then a f (x)dx = limt→b− a f (x)dx
Chain Rule) we have: F 0 (x) = f (g(x)) · g 0 (x).
use substitution on u = tan(x). (iii) Discontinuity at a and b (a < c < b): R g(x)
If F (x) = h(x) f (t)dt = ag(x) f (t)dt− ah(x) f (t)dt,
R R
Rc Rb
(iii) Deg(n) odd and Deg(m) even: use either a f (x)dx + c f (x)dx, if both convergent.
then F 0 (x) = f (g(x))g 0 (x) − f (h(x))h0 (x).
(i) or (ii). Convergence Tests:
(iv) Deg(n) even and Deg(m) odd: Deal with
each integral differently. · Comparison Test: If f (x) ≥ g(x) ≥ 0 on [a, ∞[,
then:
If a∞ f (x)dx converges ⇒ a∞ g(x)dx converges.
R R
7.6 Root-Trig Substitution Trick R∞ R∞
If a g(x)dx diverges ⇒ a f (x)dx diverges.
Useful: If a > 0 ⇒ a∞ x1p dx converges if p > 1
R
If the integrals is one of the following roots use the
given substitution and formula to convert it to an and diverges if p ≤ 1.
integral involving trig functions. · Limit Comparison Test: If f, g are continous
√ f (x)
(i) a2 − b2 x2 =⇒ x = ab sin(u), with property on [a, ∞[ with limx→∞ g(x) = L 6= ∞, then:
∞ R∞
cos2 (x) = 1 − sin2 (x).
R
a |f (x)|dx converges ⇔ a |g(x)|dx converges
Flavio Schneider Analysis I · Cheat Sheet Page 6
 8 Sequences
Definition 8.1 (Sequence). A sequence is
set of numbers in a specific order, more for-
mally: (an )∞n=1 is a function f : N → R where
f (n) = an .
Definition 8.2 (Convergence).
A sequence (an ) is convergent to a value L
if limn→∞ an = L, or:
∀ > 0∃N ∈ N∀n ∈ N : (n > N ⇒ |an −L| < )
If the limit doesn’t exist (±∞ or doesn’t con-
verge) we say that (an ) is divergent.
Inutition: If for any small number  there is, we can
find a number N () and L such that all points of an
after N are at most at distance  from L, the series
converges.
8.1 Convergence Criteria
(i) Linearity: If (an ) converges to a, (bn ) con-
verges to b and k ∈ N, then (kan + bn ) con-
verges to ka + b.
(ii) Multiplication: If (an ) converges to a and
(bn ) converges to b then (an · bn ) converges to
a · b.
(iii) Division: If (an ) converges to a and (bn ) con-
verges to b 6= 0 then ( ab n ) converges to ab .
n tion g on ]a, b[, then f is differentiable on ]a, b[
(iv) Uniqueness: If (an ) is convergent to a, then:
limn→∞ an = a is unique.
(v) Subsequence: If (an ) converges to a, then:
any subsequence (ank ) is also convergent to a.
(vi) Squeeze Theorem: If we have 3 convergent
sequences limn→∞ an = limn→∞ cn = L, and
limn→∞ bn = b where an ≤ bn ≤ cn , then
b = L.
(vii) Absolute: If (an ) is convergent to a, then:
|an | also converges and limn→∞ |an | = |a|.
8.2 Divergence Criteria 8.6 Accumulation Points 9 Sequences of Functions
(i) (an ) is divergent if it has two subsequence that
converge to different limits. Definition 8.6.
Definition 9.1. A sequence of a function
A number a is an accumulation point of (an )
(ii) (an ) is divergent if it has a divergent subse- (fn ) is a list of functions (f1 , f2 , ...) such that
if there exists a subsequence (ank ) that con-
quence. each fn maps a given subset of R into R:
verges to a, or:
(iii) (an ) is divergent if it’s unbounded.
(fn )n∈N , fn : I ⊆ R −→ R
∀ > 0∃K ∈ N : (k ≥ K ⇒ |ank − a| < )
8.3 Monotonicity
(i) Convergence: If (an ) converges to L, then L 9.1 Convergence
is the only accumulation point of (an ).
Definition 8.3.
(ii) Boundedness: If (an ) is bounded, then it has Definition 9.2. A sequence of a function
· A sequence is increasing if: at least one accumulation point. (fn ) can converge to a function f (x) in two
∀n : an < an+1 (iii) Divercence: If an diverges, then it has no different ways:
· A sequence is decreasing if: accumulation point.
· Pointwise if ∀x ∈ I:
∀n : an > an+1
· A sequence is monotonic if it’s either in- 8.7 Strategy lim fn (x) = f (x)
n→∞
creasing or decreasing.
· Convergence: Treat (an ) like a function and cal-
culate the limit, if it exists it’s convergent. If · Uniformly if ∀x ∈ I:
Lemma: Every sequence has a monotonic subse- it’s a recursive sequence use the Monotone Con-
quence. vergence Criteria by first proving that it’s both lim sup |fn (x) − f (x)| = 0, or:
n→∞ x∈I
monotonic increasing/decreasing and then that it’s
8.4 Boundedness bounded above if increasing and bounded below if ∀ > 0 ∃N ∈ N : n ≥ N ⇒ |fn (x)−f (x)| < 
decreasing. To find the limit let limn→∞ an =
Definition 8.4. limn→∞ an+1 = L and solve L = a∞ by plugging
(i) Convergence: If (fn ) converges uniformly,
L inside of an .
· A sequence is bounded above if: then it also converges pointwise.
∃M > 0∀n ∈ N : an ≤ M · Monotonicity: To prove that the sequence (ii) Continuity: If (fn ) converges uniformly, then
is monotonic pick a candidate between increas- f is continous.
· A sequence is bounded below if: ing/decreasing and solve the inequality with
∃m > 0∀n ∈ N : m ≤ an an , an+1 to prove your candidate. If the sequence (iii) Differentiability: If (fn ) converges pointwise
to f , and fn0 converges uniformly to the func-
· A sequence is bounded if it’s either bounded is recursive prove your candidate by induction.
above or below. · Boundedness: Try to change n in an to make and f 0 = g, or: limn→∞ fn0 = (limn→∞ fn )0 =
the sequence as small as possible to find a lower f0
8.5 Cauchy Sequence bound m, and similarly as big as possible to find
(iv) Integrability: If a sequence of integrable
an upper bound M. Give the result in terms of
function fn converges uniformly to f on [a, b],
m ≤ an ≤ M . If it’s a recursive sequence pick a
Definition 8.5. then f is integrable and:
candidate of upper/lower bound and prove it by
limn→∞ ab fn (x)dx = ab limn→∞ (fn (x)) dx =
R R
A sequence (an ) is Cauchy if: induction. Rb
a f (x)dx
∀ > 0∃N ∈ N : ∀m, n ≥ N ⇒ |an − am | < 
Inutition: If for any small number  there is, we can
find a number N such that all points of an after N
(viii) Ratio Test: Let (an ) be a sequence where
 are at most at distance  from each other, the series
a
∀n ∈ N : an > 0, then if limn→∞  n+1 =a is Cauchy.

an 
and a < 1, limn→∞ an = 0.
(ix) Boundedness: If (an ) converges, then (an ) is
bounded.
(x) Monotone Convergence: (an ) is monotone
then it’s convergent ⇔ (an ) is bounded.
If (an ) is increasing and bounded ⇒
limn→∞ an = sup{an : n ∈ N}
If (an ) is decreasing and bounded ⇒
limn→∞ an = inf {an : n ∈ N}
Flavio Schneider
(i) Cauchy Convergence Criterion:
(an ) is convergent ⇔ it’s Cauchy.
(ii) Cauchy Bounded: If (an ) is cauchy, then it’s
also bounded.
(iii) Linearity: If (an ) is Cauchy, (bn ) is Cauchy
and k ∈ N, then (kan + bn ) is also Cauchy.
The advantage is that we don’t have to find a limit
L to prove that the sequence converges.
Analysis I · Cheat Sheet Page 7
 P∞
10 Series (iii) Unsorted Property: P∞ If i=0 an converges If limn→∞ bn = 0 and bn is decreasing ⇒ the
absolutely, so does Definition 10.5 (Geometric Series).
i=0 bn where bn is a bi- series converges.
jection of the elements in an . A geometric series is a type of power series
Definition 10.1. of the form:
(iv) Sum Property:P ∞ 10.6 Convergence Strategy
P
i=0 (an +bP n ) converges ab-
· A partial sum is the sum ofPthe first n num- solutely if both ∞ i=0 an and
∞
i=0 bn are ab-
∞ ∞
bers of (an )∞ .. n X X 1. Divergence Test: If it’s easy to see that the
n=1 , or: sn = i=1 ai solute convergent. arn = arn−1
limit is not 0.
· An infinite series is the sum of all terms of n=0 n=1
an infinite sequence (an )∞ 2. P-Test/GeometricP Series: If it’s of the form
n=1 , or: 10.3 Common Series · Convergence: converges to a
if |r| < 1
P 1 P n
1−r np
, ar , or arn+1 .
∞ N (i) ex =
P∞ xn and diverges otherwise.
X X n=0 n! 3. Comparison Test: If it’s similar to a p-series
lim sn = ai ..= lim ai P∞ 2n+1 · Partial Sum: the nth partial sum of a ge- or geometric series.
n→∞ N →∞ n x
i=1 i=1 (ii) sin(x) = n=0 (−1) (2n+1)! a(1−r n )
metric series is sn = (1−r) . 4. Limit Comparison Test: If it’s a rational ex-
P∞ 2n
(iii) cos(x) = n x pression with polynomials with positive terms.
n=0 (−1) (2n)!
10.1 Convergence 5. Root Test: If can be written as an = (bn )n .
(iv) tan−1 (x) = ∞
P n x2n+1 10.5 Convergence Tests
n=0 (−1) 2n+1 P∞ 6. Ratio Test: If it contains factorials or cn .
(i) Divergence Test: Let n=1 an be a series
Definition 10.2 (Convergence). 1
= ∞ n
P
(v) 1−x n=0 x 7. Alternating Series Test: If can be written as
An infinite series is called convergent if: with limn→∞ an 6= 0 or undefined, then the
P∞ n+1
series diverges. an = (−1)n+c bn , if c ∈
/ {0, 1} we have to ma-
(vi) ln(1 + x) = n=0 (−1)n xn+1
∞ n P∞ 1 nipulate it to make it 0 or 1 (e.g.: (−1)n+2 =
X X (ii) P-Test: The series n=1 np is convergent
ak converges ⇔ lim ak exists (−1)n (−1)2 = (−1)n ).
n→∞ if p > 1 and divergent if p ≤ 1.
k=1 k=1 10.4 Common Sums 8. Integral Test: If f (n) = an is easy to inte-
(iii) Comparison Test: Let (an ), (bn ) be ulti-
⇔ (sn ) converges Pn n(n+1) n2 +n grate and f is positive and decreasing (ev. use
(i) i=1 i = 2
= 2 mately positive such that ∃N ∈ N ∀n ≥ N :
0≤ derivative).
Pn 2 1 2n3 +3n2 +n Pan ≤ bn , then:
If ∞
P∞
(ii) i=1 i = n(n + 1)(2n + 1) = If ∞
P∞
P
(i) Linearity: P i=0 ai = a, i=0 bi = b, and
∞
6 6 n=1 bn is convergent then n=1 an is also
c ∈ R, then: i=0 (cai + bi ) = ca + b (iii)
Pn 3 1 2
n (n + 1)2 convergent. 10.7 Value Calculation
P∞ P∞ i=1 i = 4 P∞ P∞
(ii) Comparison: If i=0 ai = a, i=0 bi = b,
If n=1 an is divergent then n=1 bn is also To calculate the value of a series there are two ways:
and ∀n ∈ N an ≤ bn , then a ≤ b. divergent.
P∞
(iv) Limit Comparison Test: Let (an ), (bn ) be 1. Find the series representation as a Geomet-
(iii) Start
P∞Convergence: i=0 ai is convergent
Definition 10.4 (Power Series). positive sequences and assume limn→∞ ab n = ric Power Series, and calcualte its convergence
⇔ i=N ai ∀N ∈ N is convergent.
A power series f is a series of the form: P∞ n value. Some tricks are: multiply the series by a
(iv) Bounded Convergence: If (an ) is ulti- L, Pthen: If 0 < L < ∞: n=1 an converges
∞ number, strip out the first terms (how many are
mately positive and (sn ) is abounded above, ∞ ⇔ n=1 P bn converges.
P∞ X ∞ P∞ necessary), subtract the starting series to the
then i=0 ai converges. Otherwise the series f (x) = an (x − c)n If L = 0: n=1 bn converges ⇒ n=1 an con- obtained one to balance the multiplied term.
diverges to infinity. n=0 verges. P∞ P∞ By repeating this process we might be able to
(v) Unbounded Divergence: If (an ) is un- If L = ∞: n=1 bn diverges ⇒ n=1 an di-
· Convergence: the series converges abso- get to a geometric series.
bounded and limn→∞ an = L 6= 0, then if verges.
lutely for 0 ≤ |x − c| < R, and diverges P∞ 2. If the series converges absolutley we can rear-
L > 0 the series diverges to +∞ and if L < 0 (v) Root Test: Let n=1 an be a series with
otherwise. To calculate the radius of con- 1 range the sum such that they cancel each other,
the series diverges to −∞. vergence  we use (an ) ultimately and limn→∞ |an | = L ≥ 0, n
 the ratio (or1 root) test: to do so we have to find the partial fraction de-
a then: composition of the series so that there are sub-
limn→∞  n+1  = limn→∞ |an | n = L then

a If 0 ≤ L < 1 the series converges absolutely.
10.2 Absolute Convergence n
tracting terms. Subsequently we will evaluate
  If 1 < L ≤ ∞ the series diverges. enough terms to find a repeating pattern (fac-
1  an  1 If L = 1, this test is inconclusive.
Definition 10.3 (Absolute Convergence). R= = lim  =
1 toring a constant out might help) such that they
L n→∞ an+1 
limn→∞ |an | n
P∞
An absolute convergent series
P∞ (vi) Ratio Test: Let n=1 an be a se- cancel out indefinitely. P
Then we will rewrite the
i=0 an is a
convergent series where also: . ries with  (an  ) ultimately positive and series as a partial sum N i=0 with all the terms
a that do not cancel (at the beginning and end
limn→∞  n+1 = L, then:

∞ an 
(i) Continuity: A Power Series f (x) is continous of the infinite series) and evaluate the limit to
If L < 1 the series converges absolutely.
X
|an | converges find its value.
on {x : |x − c| < R}. If 1 < L ≤ ∞ the series diverges.
i=0
(ii) Differentiability: A Power Series f (x) is dif- If L = 1 the test is inconclusive.
If
P
an is convergent but
P
|an | is divergent, ferentiable in its radius of convergence R and: (vii) Integral Test: If f (n) = an with f (x) conti- 11 Other
it’s called conditionally convergent. nous,eventually positive and
∞ R∞ P decreasing, then: 11.1 Length of a curve
0
X
n−1 k f (x)dx converges ⇔ ∞ n=k an converges.
P∞ f (x) = n · an (x − c)
i=0 |an | converges so does
(i) P
Theorem: If P∞
∞ n=0
(viii) Alternating Series Test: Let n=1 an Given a parametric curve where x = f (t) and
i=0 an .
n
be a series where an = (−1) bn or an = y = g(t) defined on an interval t ∈ [a, b] then the
(ii) Inequality:  ∞
P  P∞
n=0 an ≤

n=0 |an | (−1)n+1 bn , then: length of the curve is evaluated as follows:
Flavio Schneider Analysis I · Cheat Sheet Page 8
 11.4 Taylor Approximation 11.5 Integral Series Approximation 11.7 Approximating Definite Inte-
grals
s 2 2 P∞
b 
Given a convergent infinite series n=1 f (n) it’s
Z
dx dy
L= + dt Definition 11.1 (Taylor Series). usually hard to find it’s value. With this method n−1 
a dt dt Z b b−a X i(b − a)

A Taylor Series is the rappresentation of we can approximate the sum if f (n) is continous, f (x)dx = lim f a+
a function as an infinte power series where positive and decreasing. a n→∞ n i=0 n
We assume that the curve is traced exactly once as
t increases from a to b, and that the curve is traced f is differentiable any times at a point x0
∞ k ∞
out from left to right as t increases. (f ∈ C ∞ (x0 )) of the form: X X X
11.8 Theorems
S= f (n) = f (n) + f (n)
∞ n=1 n=1 n=k+1
X f (n) (x0 ) Theorem 12: Polynomial Roots
T∞ (f )(x; x0 ) = ·(x − x0 )n
| {z } | {z }
11.2 Bisection Method n=0 |
n! Sk Partial Sum Rk Remainder
{z }
an
A polynomial Pn of degree d has:
The Bisection method is used to approximate so- Since we can calculate an approximation Sk ≈ S,
lution to f (x) = 0 in an interval [a, b] where Rk will tell us the difference from the actual value · From 0 to n distinct real roots if d is even.
Where an is the Taylor coefficient.
f (a) · f (b) < 0 (xa is positive and xb negative or of S (Rk = S − Sk ). Using integrals we can find · From 1 to n distinct real roots if d is odd.
vice-versa). We can use the first n terms of a Taylor Series to upper and lower bound for Rk :
· Always n complex roots (Fundamental The-
approximate the value of a function f (x) around x0 Z ∞ Z ∞
b−a Rk ≥ f (x)dx, Rk ≤ f (x)dx orem of Algebra).
1. Calculate the midpoint c ← and evaluate with Tn (f )(x; x0 ).
2 k+1 k
f (c).
f (x) ≈ f (x0 ) + f 0 (x0 )(x − x0 ) = T1 (f )(x; x0 )
Then the value of the infinte series will be:
2. If |f (x)| is small enough, stop and return c. Z ∞ Z ∞ Theorem 13: Archimedean Property
This is a rough approximation (n = 1) of f (x) at
3. If f (a) · f (c) > 0 let a ← c otherwise let b ← c the point x0 with a polinomial of deg = 1, the value Sk + f (x)dx ≤ S ≤ Sk + f (x)dx
k+1 k
∀x ∈ R ∃nx ∈ N : x < nx
and restart from step 1. and derivative will be the same. If we derivate us-
ing the power rule, the first term will cancel leaving Thus to calculate the approximation:
This method works by keeping two points a and b just the derivative, to get a better approximation Theorem 14: Density Theorem
1. Choose a value for k, the higher the better the
with opposed sign and always shrinking the distance we add more terms so that also higher derivatives
approximation since we evaluate more terms
between them and the solution of f (x) = 0 which will get the same values, the factorial/exponent are ∀x, y ∈ R ∃z ∈ Q : x < y ⇒ x < z < y
where the intergal would find an unprecise
must exist by the Intermediate Value Theorem. used to get the correct derivative when the power
bound.
rule is applied multiple times, and (x − x0 ) will just
2. Evaluate Sk = kn=1 f (n).
P
shift the function if x0 is not centered at 0. Theorem 15: Function Implication
11.3 Newton method 3. Evaluate both improper intergrals L =
f (2) (x− x0 )2 R∞ R∞
f (x) ≈ f (x0 ) + f 0 (x0 )(x − x0 ) + k+1 f (x)dx and U = k f (x)dx.
Given a function f , the following implications
The Newton method is used to approximate solu- 2! hold:
tions to f (x) = 0, pay attention, not always this 4. Then S will be between L+Sk and U +Sk , eval-
f (n) (x − x0 )n
method converges, and it could also converge to a + ··· + = Tn (f )(x; x0 ) uate the mean to get an average term Sapprox = diff. ⇒ continous ⇒ r. integrable ⇒ bounded
n! 2Sk +L+U
wrong value. 2
.
Remainder: None of the properties on the right implies one
1. If an interval I = [a, b] is given we start by f (x) = Tn (f )(x; x0 ) + Rn (f )(x; x0 ) 11.6 Prove Bijectivity of the properties on the left.
making a random guess for the approximation
by taking b−a as our x0 . Rn (f )(x; x0 ) ..= |f (x) − Tn (f )(x; x0 )| To prove that a function f : X → Y is bijective we
2
have to prove that it’s both: 11.9 Extra
The remainder Rn quantifies how good is the esti-
2. We evaluate the next (n + 1)st guess with the Arithmetic Geometric Series
f (x ) mate of the Taylor Series with respect to the actual · Surjective: (∀x ∈ X ∃y ∈ Y : F (x) = y) we have
following formula xn+1 = xn − f 0 (xn ) provided value of the function f (x). to prove that f is continous and either one of the ∞
n
0
X
that f (xn ) exists. following is true: nq n−1 = 1 + 2q + 3q 2 + · · · + nq n−1
Theorem 11: Taylor’s Theorem n=1
3. To get n decimal places of precision we repeat 1. limx→inf (X) f (x) = inf (Y ) and
point 2. until the last n digits are unchanged 1 − (n + 1)q n + nq n+1
If f : I → R is differentiable n + 1 times limx→sup(X) f (x) = sup(Y ) =
for two consecutive cycles. f ∈ C (n+1) (I) in an interval I containing the 2. limx→inf (X) f (x) = sup(Y ) and (1 − q)2
center x0 ∈ I, then for each x ∈ I there exists limx→sup(X) f (x) = inf (Y )
The Newton method works by recursively finding a ξ ∈ ]x, x0 [ such that:
the intersection between the original function f and Then by the Intermediate Value Theorem f (x)
the tangent line where the current guess lies (g). f (n+1) (ξ) n+1 covers the entire domain and thus it’s surjective.
Subsequently it uses this line’s intercept with the x Rn (f )(x; x0 ) = (x − x0 )
(n + 1)! · Injective: (F (x) = F (y) ⇒ x = y) we have to
axis, to find the next guess.
show that f (x) is strictly increasing if when we
Lagrange Error Bound: proved surjectivity we used 1) or strictly decreas-
 (n+1)  (x−x0 )n+1
g(x) = f (xn ) + f 0 (xn )(x − xn ) ⇒ g(x) = 0 |Rn (f )(x; x0 )| ≤ sup f (ξ) (n+1)! , ing if we used 2) which can be proved with the first
x0 <ξ<a
|{z}
derivative (> 0 or < 0).
| {z } | {z }
yn slope xn

Flavio Schneider Analysis I · Cheat Sheet Page 9

 Continous Piecewise Function

2
x − ax + b x ≤ −1

f (x) = (a + b)x −1 < x < 1
 2

x + ax − b x ≥ 1

Then to have continuity both must be true:

!
f (−1) = 1 + a + b = lim f (x)
x→−1+

= lim (a + b)x = −(a + b)

x→−1+
!
f (1) = 1 + a − b = lim f (x)
x→1−
= lim (a + b)x = a + b
x→1−

Function Length If f is differentiable on [a, b],

then the graph of the function has a length:
Z b q
L= 1 + (f 0 (x))2 dx
a

Flavio Schneider Analysis I · Cheat Sheet Page 10