Week 5: Exp./Log.

/Taylor Series/Limits II

▶ The natural logarithm and exponent

▶ Exponential and logarithmic functions
▶ Growth and Decay
▶ Linear approximations
▶ Taylor polynomials
▶ Indeterminate form

1 / 57
The factorial of an integer number

For numbers n = 0, 1, 2, . . . define 0! = 1 and n! (“n-factorial’):

n! = n · (n − 1) · (n − 2) · . . . 2 · 1

▶ 1! = 1
▶ 3! = 3 · 2 · 1 = 6
▶ 4! = 4 · 3 · 2 · 1 = 24
▶ 100! = 100 · 99 · . . . · 2 · 1 ≈ 9.332621544394410 · 10157 has
158 digits!

2 / 57
The function e x

For all x ∈ R e x is defined as:

∞
x x2 x3 x4 X xk
e =1+x + + + + ... =
2 3! 4! k!
k=0

Although all terms are powers of x is e x not a polynomial!

A sum that does not stop (such as the one for e x ) is called a series
Is this really the definition of e x ? Yes!
Domain of f (x) = e x is Df = R and range is (0, ∞)
3 / 57
The function e x

Again: for all x ∈ R e x is defined as:

∞
x2 x3 x4 X xk
ex = 1 + x + + + + ... =
| 2 {z3! 4! } k=0 k!
series

This means
e = e 1 = 1 + 1 + 1/2 + 1/6 + 1/24 + · · · ≈ 2.718281828459046

4 / 57
The function e x

Again: for all x ∈ R e x is defined as:

∞
x2 x3 x4 X xk
ex = 1 + x + + + + ... =
| 2 {z3! 4! } k=0 k!
series

where the infinite sum is defined with the help of limits:

∞ n
x
X xk X xk
e = = lim
k! n→∞ k!
k=0
| {z } |k=1{z }
no polynomial in x polynomial of degree n in x

The fact that e x is this series for all x ∈ R is really special:

Often the limit only exists for x in a small interval!

5 / 57
The derivative of the function e x
d x
For the derivative it holds thate = e x because:
dx
!
d x d x2 x3 x4
e = 1+x + + + + ...
dx dx 2 3! 4!
!
d d d x2 d x3 d x4
= 1+ x+ + + + ...
dx dx dx 2 dx 3! dx 4!
!
x2 x3
= 0+1+x + + + ...
2 3!
= ex

with
d x k+1 xk xk
= (k + 1) · =
dx (k + 1)! (k + 1)! k!

6 / 57
The inverse of e x is the function ln(x)

Function e x is increasing, thus 1-to-1, 5

ex and ln(x)
thus it has an inverse, with name ln(x)
Inverse function ln(x) is thus also 1-to-1,
has domain (0, ∞) and range (−∞, ∞) 0

y = e x ⇐⇒ x = ln(y )
−5
−5 0 5
e ln(x) = x, ln(e x ) = x

The standard power-properties hold:

▶ e x e y = e x+y
1
▶ e −x = , etc
ex

7 / 57
The inverse of e x is the function ln(x)

For the derivative of the inverse ln(x) 5

ex and ln(x)
of e x it holds (set f (x) = e x , inverse
g (x) = ln(x)) (g ◦ f )(x) = x thus with
use of the chain rule 0

d d
(g ◦ f )(x) = x ⇐⇒ g ′ (f (x)) · f ′ (x) = 1 =⇒
dx dx
1 1
(g )′ (f (x)) = ⇐⇒ (g )′ (e x ) =
chain rule f ′ (x) d e x −5
−5 0 5
e x =e x
dx
1 1
⇐⇒ (g )′ (y ) = x ⇐⇒ (ln)′ (y ) =
y :=a y g =ln y

so
d 1
ln(x) =
dx x

8 / 57
Exponential and logarithmic functions
ax with domain (−∞, ∞) and range (0, ∞) is increasing thus
1-to-1
2x and log2(x) (1/2)x and log1/2(x)
4 4

3 3

2 2

1 1

0 0

−1 −1

−2 −2
−2 −1 0 1 2 3 4 −2 −1 0 1 2 3 4

=⇒ has inverse, we call loga (x),

with domain (0, ∞) and range (−∞, ∞)
this one you can easily sketch by mirroring at the line y = x

y = ax ⇐⇒ x = loga (y ) aloga (x) = x, loga (ax ) = x

9 / 57
Exponential and logarithmic functions
Remember:
e x ′ = e x =⇒
x ′ ′
ax ′ = e ln(a ) = e x ln(a) = e x ln(a) · ln(a) = ax · ln(a)

Remember:
(ln(x))′ = 1/x.
With function f (x) = ax and inverse of this g (x) = a log(x):

1 1
(g )′ (y ) = = x
chain rule f ′ (x) a · ln(a)
1
= =⇒
y :=e x y ln(a)
d 1
dx loga (x) = x
a ln(a)

for all x > 0.

10 / 57
Exponential and logarithmic functions

log10 used very often

loge = ln used often
log2 only used often in computer science
loga little used

loga (xy ) = loga (x) + loga (y )

loga ( yx ) = loga (x) − loga (y )
loga (x r ) = r · loga (x)
loga ( x1 ) = − loga (x)
loga (1) =0
loga (x) = loga (b) · logb (x)

11 / 57
Function x → f (x)g (x) and its derivative
g (x)
f (x)g (x) = e ln(f (x)) = e g (x) ln(f (x))

Derivative: Write “base number” f (x) as e-power:

 ′  ′
f (x)g (x) = e g (x) ln(f (x))
 
g (x) ln(f (x)) ′ 1 ′
= e g (x) · ln(f (x)) + g (x) · f (x)
f (x)
f ′ (x)
 
= f (x)g (x) g ′ (x) · ln(f (x)) + g (x)
f (x)
′
Thus (f g )′ = f g (g ′ ln(f ) + g ff )

Step by step: Hint: remember method, not formula

1. Write f (x)g (x) as e g (x) ln(f (x))
2. Find the derivative: e g (x) ln(f (x)) ·(derivative exponent)
3. Write e g (x) ln(f (x)) back to f (x)g (x)
12 / 57
Exponential function x x and its derivative
Derivative of x x

▶ Standard approach: Base number x is e-power:

x x = (e ln(x) )x = e x ln(x)

▶ or expression x x is e−power:
x)
x x = e ln(x = e x ln(x)

(rule ln(x r ) = r ln(x))

Derivative:
′  
x ′

x ln(x) x ln(x) 1
(x ) = e =e 1 · ln(x) + x · = x x (ln(x) + 1)
x

13 / 57
Logarithmic equations
Basic principle

▶ to get rid of e-power, take ln

▶ to get rid of ln, take e-power

Alternatively: see that you get:

ln(something) = ln(something else),

then you can conclude

something = something else

because ln is an increasing function (1-to-1)

Always check the obtained solutions! (as with equations

=⇒ argument of ln must always be > 0
inequalities with roots)
14 / 57
Logarithmic equations

Examples:

▶ ln(x) + ln(x + 2) = 1
▶ ln2 (x) + ln(x 2 ) − 3 = 0
√
▶ Solve y from e y = x6 for x ≥ 1

15 / 57
Logarithmic inequalities

Basic principle: ln(x) is an increasing function, thus:

ln(x) > ln(4) ⇒ x > 4
ln(x) < ln(4) ⇒ 0 < x < 4
Pay attention to the domain!

▶ ln(2x − 3) < 3 − ln(x)

⇒ e ln(2x−3) < e 3−ln(x) because e x is an increasing function
⇒ 2x − 3 < e 3 · e − ln(x)
−1
⇒ 2x − 3 < e 3 · e ln(x )
3
2x − 3 < ex ⇒ solve . . .
2x 2 −3x−e 3
(solve first for the equation, or x
< 0 etc)
But extra requirement for domain:
2x − 3 > 0 and x > 0 thus x > 23

16 / 57
Exercises: Exponential and logarithmic equations
▶ Write ln 98 in terms of ln(2) and ln(3)
▶ Simplify ln(sin(x)) − ln sin(x)
4
6
▶ Simplify e − ln(5x )
▶ Simplify e ln(5x)−ln(2y )
▶ Solve y from ln(y ) = 2x + 7
▶ Solve y from ln(y − 3) − ln(5) = 3x + ln(x)
▶ Solve e 2x = 9
3
▶ Derivative of y = cos(e −4x )
▶ Derivative of y = ln(x 8 )
▶ Derivative of y = ln x9
dy
▶ dx for x 3y = y 5x
▶ Solve ln(2x) + 3 ln((x + 32 )1/3 ) = 1

17 / 57
Computer algorithms for Mathematical Operations

1. Arithmetic operations
→ addition, multiplication as in schoolbook

2. Algebraic functions
→ polynomials ax n + bx n−1 + . . . + c = 0
3. Special functions
→ sin(x), cos(x), e x , ln(x), . . . ???

18 / 57
Approximate f (x) with 1st degree polynomial p1 (x)
First degree polynomial p approximates f at x = a:
p(x) = c0 + c1 (x − a) has two degrees of freedom.
2
x

0
−3 −2 −1 0 1 2 3
x

Thus we need 2 conditions: same function value, same derivative:

▶ p(a) = f (a) so it goes through the point; and
▶ p ′ (a) = f ′ (a) thus slope is the same

because from p(x) = f (a) + f ′ (a)(x − a) it follows that

|{z}
=:y
p(x) = f (a) + f ′ (a)(x − a) p(a) = f (a) + 0
p ′ (x) = 0 + f ′ (a) p ′ (a) = f ′ (a)
gives tangent line!!
19 / 57
▶ Give the linearization of f (x) = 3x 3 + 5x + 4 in x = −1
▶ f (x) = 5x 2 − 3x. Give an estimate for f (−1.9) with a
linearization in x = −2

20 / 57
Approximate f (x) with 2nd degree polynomial p2 (x)

Suppose we want to use more derivatives of f and p identically:

▶ p(a) = f (a) goes through the point; and

▶ p ′ (a) = f ′ (a) has the same slope
▶ p ′′ (a) = f ′′ (a) has the same curvature

This is polynomial of degree 2 (check the degree!)

(x − a)2 ′′
p2 (x) = f (a) + (x − a)f ′ (a) + f (a)
| {z } 2
tangent line p1 (x)

p(x) = . . . (above) . . . p(a) = f (a) + 0 + 0

fulfills 3 req: p ′ (x) = 0 + f ′ (a) + (x − a)f ′′ (a) p ′ (a) = f ′ (a) + 0
p ′′ (x) = f ′′ (a) p ′′ (a) = f ′′ (a)
This polynomial is unique.

21 / 57
Approximate f (x) with 3rd, 4th, . . . degree polynomial
Suppose we want to use more derivatives of f and p identical:

▶ p(a) = f (a) goes through the point; and

▶ p ′ (a) = f ′ (a) has the same slope
▶ p ′′ (a) = f ′′ (a) has same curvature
▶ p ′′′ (a) = f ′′′ (a) etc

This polynomial of degree 3 (check the degree!)

(x − a)2 ′′ (x − a)3 ′′′

p3 (x) = f (a) + (x − a)f ′ (a) + f (a) + f (a)
2 3!
fulfills the 4 requirements:
p(x) = . . . (see above) p(a) = f (a) + 0 + 0 + 0
(x − a)2 ′′′
p ′ (x) = 0 + f ′ (a) + (x − a)f ′′ (a) + f (a) p ′ (a) = f ′ (a) + 0 + 0
2
p ′′ (x) = f ′′ (a) + (x − a)f ′′′ (a) p ′′ (a) = f ′′ (a) + 0
p ′′′ (x) = f ′′′ (a) p ′′′ (a) = f ′′′ (a)
This polynomial is unique. The same for degree 4, 5, etc.
22 / 57
Approximate f (x) with 3rd, 4th, . . . degree polynomial
E.g.: Approximate the function f (x) = sin(x) at a = 0. That
yields the derivatives

f (x) = sin(x) f (0) = 0

f ′ (x) = cos(x) f ′ (0) = 1
f ′′ (x) = − sin(x) f ′′ (0) = 0
f ′′′ (x) = − cos(x) f ′′′ (0) = −1

and approximating polynomial of degree 3

(x − a)2 ′′ (x − a)3 ′′′
p3 (x) = f (a) + (x − a)f ′ (a) + f (a) + f (a)
2 2 3! 3
(x − 0) ′′ (x − 0) ′′′
= f (a) + (x − 0)f ′ (a) + f (0) + f (0)
2 3
2 3
3!
x x x
= 0+x ·1+ ·0+ · −1 = x −
2! 3! 6
Does p3 approximate the sine better than p1 (p1 is the tangent
line)?
23 / 57
Approximate f (x) with 3rd, 4th, . . . degree polynomial

E.g.: Remember: Take function f (x) = sin(x) and approximate f

2
p1
1.5

p5
1

0.5

0
sin

−0.5 p3
p
7

at a = 0 −1−1 0 1 2 3 4

f (x) = sin(x) f (0.03) = 0.029995500202496

p1 (x) = x p1 (0.03) = 0.030000000000000
p3 (x) = x − 16 x 3 p3 (0.03) = 0.029995500000000
p5 (x) = x − 16 x 3 + 1 5
120 x p5 (0.03) = 0.029995500202500
p7 (x) = x − 16 x 3 + 1 5
120 x − 1
5040 x
7 p7 (0.03) = 0.029995500202496

24 / 57
Approximate f (x) with 3rd, 4th, . . . degree polynomial

E.g.: Remember: Take function f (x) = sin(x) and approximate f

2
p1
1.5

p5
1

0.5

0
sin

−0.5 p3
p
7

at a = 0 −1−1 0 1 2 3 4

When we approximate f (0.01) with pn (0.01) the difference

f (x) − pn (x) is called the approximation error

25 / 57
Approximate f (x) with 3rd, 4th, . . . degree polynomial

E.g.: Remember: Take function f (x) = sin(x) and approximate f

at a = 0
2
p1
1.5

p5
1

0.5

0
sin

−0.5 p3
p
7
−1
−1 0 1 2 3 4

Although p1 , p3 , p5 , p7 only approximate sin at a = 0

the approximations at a ̸= 0 always get better!
The approximations of a higher degree also become better close to
a = 0 (later) but we can’t see that here.

26 / 57
Approximate f (x) with 3rd, 4th, . . . degree poly. (VI)

E.g.: Remember: Take function f (x) = sin(x) and approximate f

at a = 0
2
p1
1.5

p5
1

0.5

0
sin

−0.5 p3
p
7
−1
−1 0 1 2 3 4

Why are we doing such approximations with p1 , p3 , p5 , p7 , . . .?

Answer: Because computers can only add/subtract and multiply!
so computers use these kind of approximations to calculate sin(a)
for a given a!

27 / 57
Taylor polynomial pn approximates f around point a

The polynomial approximation pn of degree n of the function f in

point a:

1
pn (x) = f (a) + f ′ (a)(x − a) + f ′′ (a)(x − a)2
2!
1 ′′′ 1
+ f (a)(x − a)3 + · · · + f (n) (a)(x − a)n
3! n!

is “the” Taylor polynomial – and is a polynomial of degree ≤ n

If a = 0 the Taylor polynomial is also called Maclaurin polynomial

28 / 57
Taylor series: Matching all derivatives . . .
The function p∞ that approximates the function f in a and which
matches all derivatives f (0) (a) = f (a), f ′ (a), f ′′ (a), f (3) (a),
f (4) (a), . . . in a is:

1
p∞ (x) = f (a) + f ′ (a)(x − a) + f ′′ (a)(x − a)2
2!
1 ′′′ 1
+ f (a)(x − a)3 + · · · + f (n) (a)(x − a)n + · · ·
3! n!

is called Taylor series of f in a.

Attention: p∞ (x) is not a polynomial!
It can be shown for all the good functions that: p∞ (x) = f (x)
(really identical!) for x close to a.
This means only for x so that |x − a| < r and r is really small.
Only for e x , sin(x) and cos(x) it holds that p∞ (x) = f (x) for all
x ∈ R!
29 / 57
Match only n + 1 derivatives . . .
Seen: p∞ = f around x and also pn ̸= f for most f and n.
Generalization of the Mean Value Theorem:
For x around a there is a number c so that

1
p∞ (x) = pn (x) + f (n+1) (c)(x − a)n+1
| {z } (n + 1)!
f (x)

Written differently: Approximation error

1
f (x) − pn (x) = f (n+1) (c)(x − a)n+1
(n + 1)!

can be estimated. Computers approximate all functions e x , sin(x),

cos(x) etc by polynomials pn and do this so that the approximation
error cannot be seen.
But there are rounding errors
30 / 57
Taylor series in a = 0 for known functions

x2 x3
ex = 1+x + 2! + 3! + ···
x3 x5
sin(x) = x− 3! + 5! − ···
x2 x4
cos(x) = 1− 2! + 4! − ···
x2 x3
ln(1 + x) = x − 2 + 3 − ···
1
1−x = 1 + x + x2 + · · ·

31 / 57
Taylor series in a = 0 for known functions
p(x) = f (0) + f ′ (0)x + 12 f ′′ (0)x 2 + 1 ′′′
3! f (0)x
3 + ···

▶ of e x : all derivatives of e x are also e x

thus f (0) = f ′ (0) = f ′′ (0) = · · · = 1
1 2 1 3
thus e x = 1 + x + 2! x + 3! x + ···
▶ of sin(x): derivatives of sin(x):
′ ′ ′ ′ ′
sin(x) → cos(x) → − sin(x) → − cos(x) → sin(x) → . . .
thus f (0) = 0, f ′ (0) = 1, f ′′ (0) = 0, f ′′′ (0) = −1, f ′′′′ (0) = 0,
...
1 3 1 5
thus sin(x) = x − 3! x + 5! x − · · · (odd powers: odd function)
▶ of cos(x): derivative of cos(x):
′ ′ ′ ′ ′
cos(x) → − sin(x) → − cos(x) → sin(x) → cos(x) → . . .
thus f (0) = 1, f ′ (0) = 0, f ′′ (0) = −1, f ′′′ (0) = 0, f ′′′′ (0) = 1,
...
1 2 1 4
thus cos(x) = 1 − 2! x + 4! x − · · · (even powers: even function)
32 / 57
Ex: Taylor polynomial in x = 0 of sin(x)
1 3 1 5 1 7
sin(x) ≈ x − 3! x + 5! x − 7! x + ···
2
p1
1.5

p5
1

0.5

0
sin

−0.5 p3
p7
−1
−1 0 1 2 3 4

sin(x) with approximations:

p1 (x) = x
p3 (x) = x − 61 x 3
p5 (x) = x − 61 x 3 + 1 5
120 x
p7 (x) = x − 61 x 3 + 1 5
120 x − 1
5040 x
7

33 / 57
Some notation for every Taylor polynomial

n
1 2 1 X 1
ex ≈ 1+x + x + · · · + xn = xk
2! n! k!
k=0
n X (−1)k
1 3 1
sin(x) ≈ x − x + · · · + (−1)n x 2n+1 = x 2k+1
3! (2n + 1)! (2k + 1)!
k=0

34 / 57
Taylor polynomial/series of polynomials

E.g.: f (x) = x 2
Preparation: f ′ (x) = 2x, f ′′ (x) = 2, f ′′′ (x) = 0, etc.
1 ′′
p(x) = f (0) + f ′ (0)x + 2! 1 ′′′
f (0)x 2 + 3! f (0)x 3 + · · ·
1 1
=0+0·x + 2! · 2x 2 + 3! · 0 · x 3 + 0 + 0 + · · · = x 2!
We find that for all polynomials it holds that:
Taylor series of polynomial = polynomial itself
Evt Taylor polynomial can also exist ̸= polynomial itself!:
3
p
E.g.: Taylor polynomial of x 5 + 3x 3 + 1:
5

0th-degree: p0 (x) = 1 2
p3
1st-degree: p1 (x) = 1 p0
1
2nd-degree: p2 (x) = 1
3rd-degree: p3 (x) = 3x 3 + 1 0
4th-degree: p4 (x) = 3x 3 + 1
5th-degree: p5 (x) = x 5 + 3x 3 + 1 −1
−1 −0.5 0 0.5 1
6th-and higher degrees: also x 5 + 3x 3 + 1

35 / 57
Taylor series cut off and O-notation

Often a certain number of terms is more than enough in a Taylor

series

x3 x5
E.g.: sin(x) = x − 3! + 5! − ···

For small x (“x ≪ 1”, “x much smaller than 1”),

is x 5 is really very small ⇒ we often neglect

Then “cut off”, with notation:

3
sin(x) = x − x3! + O(x 5 ) if x → 0
O: pronounce “big-O”, or “order”
O(x 5 ) means: depends on x 5 ,

36 / 57
Taylor series cut off and O-notation

E.g.: 100x 2 + 5x 3 = O(x 2 ) if x → 0:

x 100x 2 + 5x 3 100x 2 5x 3
1 105 100 5
0.1 1.005 1 0.005
0.01 0.010005 0.01 0.000005
0.001 0.000100005 0.0001 0.000000005

because 100x 2 and 100x 2 + 5x 3 most resemble each other as

x → 0.
E.g.: 100x 2 + 5x 3 = 100x 2 + O(x 3 ) if x → 0
E.g.: − 21 x 4 = O(x 4 )
a minus sign you can integrate in O

37 / 57
Calculation with O-notation

x3 x5
E.g.: sin(x) = x − 3! + 5! − ···
x3
sin(x) = x − 3! + O(x 5 ) als x → 0
But you can also say:
sin(x) = x + O(x 3 ) if x → 0
then you cut the series off earlier, and you neglect more terms

E.g.: (x + O(x 3 ))2 = x 2 + O(x 4 ) because

(x +O(x 3 ))(x +O(x 3 )) = x 2 +O(x 4 )+O(x 4 )+O(x 6 ) = x 2 +O(x 4 )

x 2 +O(x 3 )
E.g.: x = x + O(x 2 )

38 / 57
Determination of Taylor series

▶ With definition (determine derivatives), or easier:

▶ Try using known Taylor series!
1
E.g.: Taylor series of 1+2x 2
around x = 0
1
E.g.: Taylor polynomial of degree 3 for f (x) = 8−x in x = 0

39 / 57
Determination of Taylor series

E.g.: Taylor series of e 3x around x = 0

E.g.: Taylor series of e x+3 around x = 0

40 / 57
Examples Taylor series
x
E.g.: Taylor series of ex around x = 0
x
WRONG:
1 + x + 12 x 2 + · · ·
because this is not a polynomial in x, but a rational function
GOOD: x e −x = x (1 − x + 12 x 2 − · · · )

3
E.g.: Taylor series of e x with 3rd degree around x = 0
1 + x 3 + 21 x 6 + · · · = 1 + x 3 + O(x 6 )
here only 2 terms remain

E.g.: Taylor series of cos(x 4 ) with 3rd degree around x = 0

1 − 21 x 8 + · · · = 1 + O(x 8 )
here only 1 term remains

41 / 57
Step-by-step: Taylor series in a ̸= 0

1. You need powers y = x − a

2. write x = y + a and fill that in into f (x)
3. Try to find a Taylor series with powers of y
(and not eg y − 1)
4. Replace y again by x − a

42 / 57
Linear, quadratic approximations: examples
√
Taylor
√ polynomial of degree 0, 1, 2 for f (x) = 1 + x in x = 0
Of 1 + x we do not know a standard Taylor series
=⇒ use the definition
f ′ (x) = 12 (1 + x)−1/2 , f ′′ (x) = − 41 (1 + x)−3/2
Degree 0: p(x) = f (0) = 1
Degree 1: p(x) = 1 + f ′ (0)x = 1 + 12 x
Degree 2: p(x) = 1 + 21 x + 21 · (− 41 )x 2
√ 1
Useful: 1+x ≈1+ x
2
√
Assume we want to approximate
√ 1.1
Linear approximation: 1.1 ≈ 1 + 21 · 0.1 = 1.05
√
Quadratic approximation: 1.1 ≈ 1 + 21 0.1 − 18 · (0.1)2 = 1.04875
√
True value: 1.1 = 1.04880 · · ·

43 / 57
Exercises: Taylor polynomial and Taylor series

▶ Find the Taylor polynomial of degree 0, 1, 2 and 3 of

f (x) = ln(x) in x = 1
3
▶ Taylor series of 6 + e 4x
▶ Find the 3rd-order Taylor polynomial of
f (x) = ln((1 + 2x)2 ) + e −x+1 in x = 0
▶ Find the 4th-order Taylor polynomial of
1
f (x) = ln(x) + sin(πx) + 1+(x−1) 2 in x = 1

44 / 57
Limits, with l’Hôpital and Taylor

sin(x)
Reminder: lim =1 why??
x→0 x

sin(x)/x

0.5

−0.5
−1 −0.5 0 0.5 1

This sort of limit we can determine with l’Hôpital

45 / 57
Limits with l’Hôpital
f (x)
Assume f (a) = 0 and g (a) = 0; the question is: lim =?
x→a g (x)
f (x)−f (a)
f (x) f (x) − f (a) x−a f ′ (x)
lim = lim = lim = lim
x→a g (x) x→a g (x) − g (a) x→a g (x)−g (a) x→a g ′ (x)
x−a
Therefore rule of l’Hôpital:

f (x) f ′ (x)
lim = lim ′
x→a g (x) x→a g (x)

only in the following 2 situations:

▶ if limit is 0/0 after filling in
▶ if limit is ∞/∞ after filling in (or −∞ instead of ∞; proof difficult)
sin(x) cos(x)
E.g.: lim = lim =1
x→0 x x→0 1
f (x) ′ f ′ (x)g (x)−f (x)g ′ (x)
Attention! Do not mix up l’Hôpital with quotient rule g (x)
= (g (x))2
46 / 57
Limits with l’Hôpital
x2 − 9 2x
With l’Hôpital this also works: lim = lim =6
x→3 x − 3 x→3 1

l’Hôpital can also be repeated if after taking the derivatives,

there is still 0/0 (or ∞/∞):
f (x) f ′ (x) f ′′ (x)
lim = lim ′ = lim ′′ etc (until no more 00 or
x→a g (x) x→a g (x) x→a g (x)
∞
∞ occurs )

sin(x) − x cos(x) − 1 − sin(x)

E.g.: lim = lim = lim =0
x→0 cos(x) − 1 x→0 − sin(x) x→0 − cos(x)

ex ∞
E.g.: lim (of type ∞ )
x→∞ x 2
ex ex
= lim = lim =∞
x→∞ 2x x→∞ 2

47 / 57
Limits with Taylor

You can also use Taylor for the limit, it’s an alternative to l’Hôpital
Taylor makes it hereby possible to divide by a factor
sin(x) x − 16 x 3 + · · · 1 − 16 x 2 + · · ·
E.g.: lim = lim = lim =1
x→0 x x→0 x x→0 1
With O-notation:
sin(x) x + O(x 3 ) 1 + O(x 2 )
lim = lim = lim =1
x→0 x x→0 x x→0 1

48 / 57
Limit with Taylor

One disadvantage: it is often not immediately clear to which order

you have to use Taylor, (Exam 24.10.05)

4
1
2
ln(1 + 2x 2 ) + 1
1+x 2
−1 1
2
(2x 2 − 4x2 + O(x 6 )) − x 2 + x 4 + O(x 6 )
lim = lim
x→0 sin(x 2 ) − x 2 cos(x 2 ) x→0 x 2 + O(x 6 ) − x 2 (1 − 12 x 4 + O(x 8 ))
O(x 6 )
= lim = ?? no luck! “constants are in the O”
x→0 O(x 6 ) + O(x 10 )
Therefore now Taylor with 6th powers:
4x 4 6
1
2
(2x 2 −2
+ 8x3 + O(x 8 )) − x 2 + x 4 − x 6 + O(x 8 )
lim 6
x→0 x 2 − x6 + O(x 10 ) − x 2 (1 − 12 x 4 + O(x 8 ))
( 43 − 1) x 6 + O(x 8 ) 1
3
+ O(x 2 ) 1.1
= lim = lim = =1
x→0 (− 16 + 12 ) x 6 + O(x 10 ) x→0 1
3
+ O(x 4 ) 3 3

Works also with l’Hôpital, but then you have to use it 6× . . .

49 / 57
Problem cases for limits . . .
What are the problem cases and why exactly?
We have seen earlier:

▶ ̸=0
0 : can be ∞, −∞ or can not exist;
consider left and right limits
▶ 00 : l’Hôpital, factor division or Taylor
▶ ∞∞ : divide by largest in denominator or l’Hôpital

We will now also consider:

▶ 0 · ∞: (which is stronger?)
▶ ∞ − ∞: (which is stronger?)
▶ 00 : 0x = 0 if x > 0, but x 0 = 1 for all x, thus problem
▶ ∞0 : ∞x = ∞ if x > 0, but x 0 = 1 if x < ∞: problem
▶ 1∞ : 1x = 1 if x ̸= ∞
but (1 + 10−10 )∞ = ∞ and (1 − 10−10 )∞ = 0
50 / 57
Case 0 · ∞ or 0 · (−∞)

0 ∞
0 · ∞: rewrite to 0 or ∞

E.g.: lim+ x ln(x) is of type 0 · (−∞):

x→0
(but in this case you may also say that it is a standard limit)

51 / 57
Case ∞ − ∞

∞ − ∞: Try to combine to 1 term

1 1
E.g.: lim+ −
x→0 x sin(x)
p
E.g.: lim x2 − x − x
x→∞

52 / 57
Cases 00 , ∞0 , 1∞ : use trick with e ln(··· )

3
lim (1 + )x :
x→∞ x
Is type “1∞ ”; remember these examples with all tricks

3
E.g.: lim (1 + )x
x→∞ x

53 / 57
Case ∞0

√
▶ lim x 1/ x
is of the form ∞0 thus
x→∞

√
▶ lim ( x)1/x is of the form ∞0 thus
x→∞

54 / 57
Step-by-step plan limits
1. Always first try filling in
if it is no problem case: done!
̸=0
2. 0 : consider lim+ and lim
x→a x→a−
if eg ∞ and ∞ then the limit is also ∞
if eg ∞ and −∞ the limit does not exist
0
3. 0: l’Hôpital (several times), or Taylor,
or dividing factor
∞
4. ∞: divide through biggest power in
denominator or l’Hôpital)
0 ∞
5. 0 · ∞: rewrite to 0 or ∞
6. ∞ − ∞: rewrite as 1 term
7. ∞0 , 1∞ , 00 : trick with e ln

Note: 0
∞, 0∞ and ∞∞ are not problem cases

55 / 57
Exercises: Limits Indeterminate Form
2x − 16
▶ lim
x→8 6x 2 − 384
3x 2
▶ lim (a) with l’Hôpital; (b) with Taylor
x→0 cos(x) − 1
2x(cos(2x) − 1)
▶ lim (a) with l’Hôpital; (b) with Taylor
x→0 sin(3x) − 3x
3
e x − cos(x 3 ) − 14 ln(1 + 4x 3 )
▶ lim
x→0 sin(x 2 ) + arctan(x 2 ) − 2x 2

x 3 − 4x + 15
▶ lim (a) with l’Hôpital; (b) with dividing
x→−3 x 2 − x − 12
factor
ex + x 2
▶ determine lim with l’Hôpital
x→∞ 2e x − x
▶ lim+ x 9/(1−x)
x→1
▶ lim (1 + 2x)19/(2 ln(x))
x→∞
56 / 57
Week 5: We have seen

▶ Exponential and logarithmic functions

▶ Taylor polynomials and series
▶ Limits: L’Hôpital

57 / 57