Lecture 32
Lecture 32
I Indeterminate forms
I L’Hôpital’s Rule for the form 0/0
I L’Hôpital’s Rule for the form ∞/∞
“Indeterminate form” refers to types, or forms, of a limit where we cannot deduce whether the limit exists by
breaking down the expression into parts and looking at their individual limits.
sin(x) 2
E.g., lim x
, lim 2x − x 2 , lim x −1 . Evaluate these limits.
x→0 x→+∞ x→1 x−1
Clearly, “indeterminate form” does not mean that the limit does not exist. We just have to work harder to
determine whether they do or not.
There are seven standard types of indeterminate forms:
0 ∞
, , 0 · ∞, ∞ − ∞, 1∞ , ∞0 , 00
0 ∞
Question
What types are the examples above?
Can you think of other examples to illustrate why those forms are “indeterminate.”
Changing forms
You can often rewrite an expression to change the type of an indeterminate form into another one, or to make it
determinate.
Example
sin(x)
lim x
is of the form 0/0.
x→0+
1/x
It can also be expressed as lim , which is of the form ∞/∞.
x→0+ csc(x)
It can also be expressed as lim sin(x) · x1 , which is of the form 0 · ∞.
x→0+
Exercise
Find the following limits by reducing them to a determinate form:
1. lim 2x − x 2
x→∞
x 2 +x
2. lim 2
x→∞ 2x −1
L’Hôpital’s Rule for the form 0/0 at a point
Theorem 1
Let I be an open interval containing the point a and let f and g be two functions that are differentiable at each
x ∈ I \ {a}. If
1. g 0 (x) 6= 0 for all x ∈ I \ {a},
2. lim f (x) = 0,
x→a
f 0 (x)
4. lim exists,
x→a g 0 (x)
f (x) f 0 (x)
then lim = lim 0 .
x→a g (x) x→a g (x)
The formula is easy to remember but pay attention to all the hypotheses!
Before looking at the proof and some examples of its use, here are some cases where you should not use
it:
x 2 + 1/x sin(1/x)
lim , lim
x→1 x x→0+ 1/x
Proof and remarks
Non-rigorous explanation: To get an intuitive idea for why L’Hôpital’s Rule works, consider the linear
approximations of f and g based at a.
Proof.
In addition to the hypotheses, we may further assume that f (a) = g (a) = 0, making f and g continuous at a.
Recall that the limit at a does not depend on f (a) or g (a).
1. Show that g (x) 6= 0 for all x ∈ I \ {a}.
2. Let x ∈ I \ {a} be arbitrary. Apply the Generalized Mean Value Theorem: there exists cx between a and x
such that
f 0 (cx ) [g (x) − g (a)] = g 0 (cx ) [f (x) − f (a)] .
Remarks:
1. The theorem also works for one-sided limits.
f 0 (x)
2. The theorem extends to cases where lim = ±∞.
x→a g 0 (x)
L’Hôpital’s Rule for the form 0/0 at ∞
Theorem 2
Let A be a real number and let f and g be two functions that are differentiable at each x > A. If
1. g 0 (x) 6= 0 for all x > A,
2. lim f (x) = 0,
x→+∞
f 0 (x)
4. lim exists,
x→+∞ g 0 (x)
f (x) f 0 (x)
then lim = lim .
x→+∞ g (x) x→+∞ g 0 (x)
A similar statement holds for limits at −∞ and the theorem extends to cases where
f 0 (x)
lim = ±∞.
x→+∞ g 0 (x)
Proof outline
1. Consider a change of variables u = 1/x, defining F (u) = f (1/u) and G (u) = g (1/u).
F 0 (u)
2. Apply the Chain Rule and evaluate lim 0 .
u→0+ G (u)
F (u)
3. Check hypotheses and apply Theorem 1 to find lim .
u→0+ G (u)
Exercises with L’Hôpital’s Rule - Part 1
Theorem
Let I be an open interval containing the point a and let f and g be two functions that are differentiable at each
x ∈ I \ {a}. If
1. g 0 (x) 6= 0 for all x ∈ I \ {a},
2. lim f (x) = +∞,
x→a
f 0 (x)
4. lim exists,
x→a g 0 (x)
f (x) f 0 (x)
then lim = lim 0 .
x→a g (x) x→a g (x)
Proof: Exercise.
One-sided version and infinite limit versions also hold.
L’Hôpital’s Rule for the form ∞/∞ at ∞
Theorem 3
Let A be a real number and let f and g be two functions that are differentiable at each x > A. If
1. g 0 (x) 6= 0 for all x > A,
2. lim f (x) = +∞,
x→+∞
f 0 (x)
4. lim exists,
x→+∞ g 0 (x)
f (x) f 0 (x)
then lim = lim .
x→+∞ g (x) x→+∞ g 0 (x)
f 0 (x)
Let L = lim .
x→+∞ g 0 (x)
Since lim f (x) = +∞ and lim g (x) = +∞, there exists M > A such that f (M) > 0 and
x→+∞ x→+∞
g (M) > 0.
Also, there exists M2 > M such that f (x) > f (M) and g (x) > g (M) for all x > M2 .
Let x > M2 > M be arbitrary. Apply the Generalized Mean Value Theorem on [M, x]: there exists cx ∈ (M, x)
such that
f 0 (cx ) [g (x) − g (M)] = g 0 (cx ) [f (x) − f (M)] .