Sei sulla pagina 1di 2

Bolzano’s Theorem

Bolzano’s theorem states the following:


Let I ⊂ R be an interval and f : I → R a continuous function on I.
Then f (I) is an interval.
We start with a couple of simple lemmas.
Lemma 1. For all a, b ∈ R, a < b, we have
[a, b] = {x ∈ R : x = (1 − λ)a + λb, 0 ≤ λ ≤ 1}.
Proof. Let S = {x ∈ R : x = (1 − λ)a + λb, 0 ≤ λ ≤ 1}. If x ∈ S, there
exists 0 ≤ λ ≤ 1, such that x = (1 − λ)a + λb = a + λ(b − a). Note that
since 0 ≤ λ(b − a) ≤ b − a, we have a ≤ a + λ(b − a) ≤ a + (b − a) = b.
Thus, x ∈ [a, b]. We proved that S ⊂ [a, b] (*). Conversely, if x ∈ [a, b],
x−a b−x
let λ = . Then 0 ≤ λ ≤ 1, 1 − λ = , and
b−a b−a
b−x x−a (b − a)x
(1 − λ)a + λb = a+ b= = x.
b−a b−a b−a
Thus, x ∈ S. We proved [a, b] ⊂ S (**). Combine (*) and (**) to
complete the proof. 

Lemma 2. Let I ⊂ R. The following are equivalent:


(1) I is an interval;
(2) ∀ x, y ∈ I such that x < y, [x, y] ⊂ I.
Proof. (1) ⇒ (2) Since I is an interval, it has to be of the form:
[a, b], [a, b), (a, b], (a, b), (−∞, a), (−∞, a], (a, ∞), [a, ∞), or R = (−∞, ∞),
where a, b ∈ R and a < b. We treat just one of these cases, the other
ones are treated similarly. Say, I = (a, b], and let x, y ∈ I, x < y.
Then, for all z ∈ [x, y], we have a < x ≤ z ≤ y ≤ b, that is z ∈ (a, b].
Thus [x, y] ⊂ I.
(2) ⇒ (1) Let a = inf I and b = sup I. We will show that (a, b) ⊂
I ⊂ [a, b]. This proves immediately that I is an interval. To treat all
cases at once, we make the convention that if I is unbounded below
(or above), then a = −∞ (respectively, b = +∞), and in this case we
replace the [ part with ( at the infinite end point.
Let therefore x ∈ (a, b) be fixed, that is a < x < b. By the properties
of infimum and supremum, we can find x1 , y1 ∈ I such that a ≤ x1 <
x < y1 ≤ b. By Lemma 1, we can then write x = (1 − λ)x1 + λy1 , with
0 < λ < 1. But we know by assumption (2) that [x1 , y1 ] ⊂ I, therefore
x ∈ I, proving that (a, b) ⊂ I. Furthermore, since a ≤ x ≤ b for all
x ∈ I, I ⊂ [a, b]. 
1
2

We are now ready to prove Bolzano’s theorem.


Proof. Let y1 , y2 ∈ J = f (I) := {y ∈ R : ∃x ∈ I, y = f (x)}, and
y1 < y2 be arbitrary.
We will prove that [y1 , y2 ] ⊂ J. Combining this fact with Lemma 2,
we will be able to conclude that J is also an interval.
Let then y1 < λ < y2 . We want to show that λ ∈ J. Since y1 , y2 ∈ J,
there exists a, b ∈ I such that y1 = f (a), y2 = f (b). Note that a 6= b,
and, without loss of generality, let us assume that a < b. Consider the
set
A = {x ∈ [a, b] : f (x) ≤ λ}.
A is clearly bounded. Let c = sup A. By 3.15.1, we know that there
exists a sequence xn ∈ A such that xn → c. Since xn ∈ A, we have
f (xn ) ≤ λ for all n ≥ 1. But, since f is continuous at c, f (c) =
limn→∞ f (xn ) and from 3.9.2 we can conclude that f (c) ≤ λ. Since
λ < f (b), we have c < b.
Obviously, f (x) ≥ λ for all x ∈ (c, b], since c = sup A. Pick then
any sequence zn ∈ (c, b) such that zn → c. But then f (zn ) ≥ λ for all
n ≥ 1. Thus, we can repeat our reasoning above to get that f (c) ≥ λ.
We have obtained f (c) ≤ λ and f (c) ≥ λ, meaning that λ = f (c) ∈ J.
This proves our claim [y1 , y2 ] ⊂ J, and hence finishing the proof. 

An immediate corollary of Bolzano’s theorem is the following version


of 4.6.1 in the lecture notes:
Let I ⊂ R be an interval and let f : I → R be a continuous function.
If a, b ∈ I, a < b, are such that f (a)f (b) < 0, then there exists c ∈ (a, b)
such that f (c) = 0.
Proof. Since J = f ([a, b]) is an interval with one end-point positive and
the other negative, we know that 0 ∈ J. Therefore, there must exist
c ∈ [a, b] such that 0 = f (c). Clearly, c 6= a and c 6= b, since f (a) and
f (b) are not zero by hypothesis. 

Potrebbero piacerti anche