3.4: Properties of Continuous Functions
( \newcommand{\kernel}{\mathrm{null}\,}\)
Recall from Definition 2.6.3 that a subset A of R is compact if and only if every sequence {an} in A has a subsequence {ank} that converges to a point a∈A.
Let D be a nonempty compact subset of R and let f:D→R be a continuous function. Then fD is a compact subset of R. In particular, f(D) is closed and bounded.
- Proof
-
Take any sequence {yn} in f(D). Then for each n, there exists an∈D such that yn=f(an). Since D is compact, there exists a subsequence {ank} of {an} and a point a∈D such that
limk→∞ank=a∈D.
It now follows from Theorem 3.3.3 that
limk→∞ynk=limk→∞f(ank)=f(a)∈f(D).
Therefore, f(D) is compact.
The final conclusion follows from Theorem 2.6.5 ◻
We say that the function f:D→R has an absolute minimum at ˉx∈D if
f(x)≥f(ˉx) for every x∈D.
Similarly, we say that f has an absolute maximum at ˉx if
f(x)≤f(ˉx) for every x∈D.
Figure 3.2: Absolute maximum and absolute minimum of f on [a,b].
Suppose f:D→R is continuous and D is a compact set. Then f has an absolute minimum and an asolute maximum on D.
- Proof
-
Since D is compact, A=f(D) is closed and bounded (see Theorem 2.6.5). Let
m=infA=infx∈Df(x).
In particular, m∈R. For every n∈N, there exists an∈A such that
m≤an<m+1/n.
For each n, since an∈A=f(D), there exists xn∈D such that an=f(xn) and, hence,
m≤f(xn)<m+1/n.
By the compactness of D, there exists an element ˉx∈D and a subsequence {xnk} that converges to ˉx∈D as k→∞. Because
m≤f(xnk)<m+1nk for every k
by the squeeze theorem (Theorem 2.1.6) we conclude limk→∞f(xnk)=m. On the other hand, by continuity we have limk→∞f(xnk)=f(ˉx). We conclude that f(ˉx)=m≤f(x) for every x∈D. Thus, f has an absolute minimum at ˉx. The proof is similar for the case of absolute maximum. ◻
The proof of Theorem 3.4.2 can be shortened by applying Theorem 2.6.4. However, we have provided a direct proof instead.
If f:[a,b]→R is continuous, then it has an absolute minimum and an absolute maximum on [a,b].
- Proof
-
Add proof here and it will automatically be hidden
Corollary 3.4.4 is sometimes referred to as the Extreme Value Theorem. It follows immediately from Theorem 3.4.2, and the fact that the interval [a,b] is compact (see Example 2.6.4).
The following result is a basic property of continuous functions that is used in a variety of situations.
Let f:D→R be continuous at c∈D. Suppose f(c)>0. Then there exists δ>0 such that
f(x)>0 for every x∈B(c;δ)∩D.
- Proof
-
Let ε=f(c)>0. By the continuity of f at c, ther exists δ>0 such that x∈D and |x−c|<δ, then
|f(x)−f(c)|<ε.
This implies, in particular, that f(x)>f(c)−ε=0 for every x∈B(c;δ)∩D. The proof is now complete. ◻
An analogous result holds if f(c)<0.
Let f:[a,b]→R be a continuous function. Suppose f(a)⋅f(b)<0 (this means either f(a)<0<f(b) or f(a)>0>f(b)). Then there exists c∈(a,b) such that f(c)=0.
- Proof
-
We prove only the case f(a)<0<f(b) (the case f(a)>0>f(b) is completely analogous). Define
A={x∈[a,b]:f(x)≤0}.
This set is nonempty since a∈A. This set is also bounded since A⊂[a,b]. Therefore, c=supA exists and a≤c≤b. We are going to prove that f(c)=0 by showing that f(c)<0 and f(c)>0 lead to contradictions.
Suppose f(c)<0. Then there exists δ>0 such that
f(x)<0 for all x∈B(c;δ)∩[a,b].
Because c<b (since f(b)>0), we can find s∈(c,b) such that f(s)<0 (indeed s=min{c+δ/2,(c+b)/2} will do). This is a contradiction because s∈A and s>c.
Suppose f(c)>0. Then there exists δ>0 such that
f(x)>0 for all x∈B(c;δ)∩[a,b].
Since a<c (because f(a)<0), there exists t∈(a,c) such that f(x)>0 for all x∈(t,c) (in fact, t=max{c−δ/2,(a+c)/2} will do). On the other hand, since t<c=supA, there exists t′∈A with t<t′≤c. But then t<t′ and f(t′)≤0. This is a contradiction. We conclude that f(c)=0. ◻
Let f:[a,b]→R be a continuous function. Suppose f(a)<γ<f(b). Then there exists a number c∈(a,b) such that f(c)=γ.
The same conclusion follows if f(a)>γ>f(b).
Figure 3.3: Illustration of the Intermediate Value Theorem.
- Proof
-
Define
φ(x)=f(x)−γ,x∈[a,b].
Then φ is continuous on [a,b]. Moreover,
φ(a)φ(b)=[f(a)−γ][f(b)−γ]<0.
By Theorem 3.4.7, there exists c∈(a,b) such that φ(c)=0. This is equivalent to f(c)=γ. The proof is now complete. ◻
Let f:[a,b]→R be a continuous function. Let
m=min{f(x):x∈[a,b]} and M=max{f(x):x∈[a,b]}.
Then for every γ∈[m,M], there exists c∈[a,b] such that f(c)=γ.
- Proof
-
Add proof here and it will automatically be hidden
We will use the Intermediate Value Theorem to prove that the equation ex=−x has at least one real solution. We will assume known that the exponential function is continuous on R and that ex<1 for x<0.
Solution
First define the function f:R→R by f(x)=ex+x. Notice that the given equation has a solution x if and only if f(x)=0. Now, the function f is continuous (as the sum of continuous functions). Moreover, note that f(−1)=e−1+(−1)<1−1=0 and f(0)=1>0. We can now apply the Intermediate Value Theorem to the function f on the interval [−1,0] with γ=0 to conclude that there is c∈[−1,0] such that f(c)=0. The point c is the desired solution to the original equation.
We show now that, given n∈N, every positive real number has a positive n-th root.
Solution
Let n∈N and let a∈R with a>0. First observe that (1+a)n≥1+na>a (see Exercise 1.3.7). Now consider the function f:[0,∞)→R given by f(x)=xn. Since f(0)=0 and f(1+a)>a, it follows from the Intermediate Value Theorem that there is x∈(0,1+a) such that f(x)=a. That is, xn=a, as desired. (We show later in Example 4.3.1 that such an x is unique.)
We present below a second proof of Theorem 3.4.8 that does not depend on Theorem 3.4.7, but, instead, relies on the Nested Intervals Theorem (Theorem 2.3.3).
We construct a sequence of nested intervals as follows. Set a1=a, b1=b, and let I1=[a,b]. Let c1=(a+b)/2. If f(c1)=γ, we are done. Otherwise, either
f(c1)>γ or f(c1)<γ.
In the first case, set a2=a1 and b1=c1. In the second case, set a2=c1 and b2=b1. Now set I2=[a2,b2]. Note that in either case,
f(a2)<γ<f(b2).
Set c2=(a2+b2)/2. If f(c2)=γ, again we are done. Otherwise, either
f(c2)>γ or f(c2)<γ.
In the first case, set a3=a2 and b3=c2. In the second case, set a3=c2 and b3=b2. Now set I3=[a3,b3]. Note that in either case,
f(a3)<γ<f(b3).
Proceeding in this way, either we find some cn0 such that f(cn0)=γ and, hence, the proof is complete, or we construct a sequence of closed bounded intervals {In} with In=[an,bn] such that for all n,
- In⊃In+1,
- bn−an=(b−a)/2n−1, and
- f(an)<γ<f(bn).
In this case, we proceed as follows. Condition (ii) implies that limn→∞(bn−an)=0. By the Nested Intervals Theorem (Theorem 2.3.3, part(b)), there exists c∈[a,b] such that ⋂∞n=1In={c}. Moreover, as we see from the proof of that theorem, an→c and bn→c as n→∞.
By the continuity of f, we get
limn→∞f(an)=f(c) and limn→∞f(bn)=f(c).
Since f(an)<γ<f(bn) for all n, condition (iii) above and Theorem 2.1.5 give
f(c)≤γ and f(c)≥γ.
It follows that f(c)=γ. Note that, since f(a)<γ<f(b), then c∈(a,b). The proof is now complete. ◻
Now we are going to discus the continuity of the inverse function. For a function f:D→E, where E is a subset of R, we can define the new function f:D→R by the same function notation. The function f:D→E is said to be continuous at a point ˉx∈D if the corresponding function f:D→R is continuous at ˉx.
Let f:[a,b]→R be strictly increasing and continuous on [a,b]. Let c=f(a) and d=f(b). Then f is one-to-one, f([a,b])=[c,d], and the inverse function f−1 defined on [c,d] by
f−1(f(x))=x where x∈[a,b],
is a continuous function from [c,d] onto [a,b].
- Proof
-
The first two assertions follow from the monotonicity of f and the Intermediate Value Theorem (see also Corollary 3.4.9). We will prove that f−1 is continuous on [c,d]. Fix any ˉy∈[c,d] and fix any sequence {yk} in [c,d] that converges to ˉy. Let ˉx∈[a,b] and xk∈[a,b] be such that
f(ˉx)=ˉy and f(xk)=yk for every k.
Then f−1(ˉy)=ˉx and f−1(yk)=xk for every k. Suppose by contradiction that {xk} does not converge to ˉx. Then there exists ε0>0 and a subsequence {xkℓ} of {xk} such that
|xkℓ−ˉx|≥ε0 for every ℓ.
Since the sequence {xkℓ} is bounded, it has a further subsequence that converges to x0∈[a,b]. To simplify the notation, we will again call the new subsequence {xkℓ}. Taking limits in (3.7), we get
|x0−ˉx|≥ε0>0.
On the other hand, by the continuity of f, {f(xkℓ)} converges to f(x0). Since f(xkℓ)=ykℓ→ˉy as ℓ→∞, it follows that f(x0)=ˉy=f(ˉx). This implies x0=ˉx, which contradicts (3.8). ◻
A similar result holds if the domain of f is the open interval (a,b) with some additional considerations. If f:(a,b)→R is increasing and bounded, followin the argument in Theorem 3.2.4 we can show that both limx→a+f(x)=c and limx→b−f(x)=d exist in R (see Exercise 3.2.10). Using the Intermediate Value Theorem we obtain that f((a,b))=(c,d). We can now proceed as in the previous theorem to show that f has a continuous inverse from (c,d) to (a,b).
If :(a,b)→R is increasing, continuous, bounded below, but not bounded above, then limx→a+f(x)=c∈R, but limx→b−f(x)=∞ (again see Exercise 3.2.10). In this case we can show using the Intermediate Value Theorem that f((a,b))=(c,∞) and we can proceed as above to prove that f has a continuous inverse from (c,∞) to (a,b).
The other possibilities lead to similar results.
A similar theorem can be proved for strictly decreasing functions.
Exercise 3.4.1
Let f:D→R be continuous at c∈D and let γ∈R. Suppose f(c)>γ. Prove that there exists δ>0 such that
f(x)>γ for every x∈B(c;δ)∩D.
- Answer
-
Add texts here. Do not delete this text first.
Exercise 3.4.2
Let f,g be continuous functions on [a,b]. Suppose f(a)<g(a) and f(b)>g(b). Prove that there exists x0∈(a,b) such that f(x0)=g(x0).
Exercise 3.4.3
Prove that the equation cosx=x has at least one solution in R. (Assume known that the function cosx is continuous.
Exercise 3.4.4
Prove that the equation x2−2=cos(x+1) has at least two real solutions. (Assume known that the function cosx is continuous.)
Exercise 3.4.5
Let f:[a,b]→[a,b] be a continuous function.
- Prove that the equation f(x)=x has a solution on [a,b].
- Suppose further that
|f(x)−f(y)|<|x−y| for all x,y∈[a,b],x≠y.
Prove that the equation f(x)=x has a unique solution on [a,b]
Exercise 3.4.6
Let f be a continuous function on [a,b] and x1,x2,…,xn∈[a,b]. Prove that there exists c \in [a,b] with
\[f(c)=\frac{f\left(x_{1}\right)+f\left(x_{2}\right)+\cdots f\left(x_{n}\right)}{n}.\
Exercise \PageIndex{7}
Suppose f is a continuous function on \mathbb{R} such that |f(x)| < |x| \text { for all } x \neq 0.
- Prove that f(0)=0.
- Given two positive numbers a and b with a < b, prove that there exists \ell \in[0,1) such that |f(x)| \leq \ell|x| \text { for all } x \in[a, b].
Exercise \PageIndex{8}
Let f, g:[0,1] \rightarrow[0,1] be continuous functions such that
f(g(x))=g(f(x)) \text { for all } x \in[0,1] . \nonumber
Suppose further that f is monotone. Prove that there exists x_{0} \in[0,1] such that
f\left(x_{0}\right)=g\left(x_{0}\right)=x_{0} . \nonumber