Monotone Functions: S. Kumaresan School of Math. and Stat. University of Hyderabad Hyderabad 500046

Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

Monotone Functions

S. Kumaresan
School of Math. and Stat.
University of Hyderabad
Hyderabad 500046
[email protected]

1. We say that a function f : J ⊂ R → R is strictly increasing if for all x, y ∈ J with x < y,


we have f (x) < f (y).
One defines strictly decreasing in a similar way. A monotone function is either strictly
increasing or strictly decreasing.
We shall formulate and prove the results for strictly increasing functions. Analogous
results for decreasing functions f can be arrived at in a similar way or by applying the
result for the increasing functions to −f .

2. Let J ⊂ R be an interval. Let f : J → R be continuous and 1-1. Let a, c, b ∈ J be such


that a < c < b. Then f (c) lies between f (a) and f (b), that is either f (a) < f (c) < f (b)
or f (a) > f (c) > f (b) holds.
Proof. Since f is one-one, we assume without loss of generality that f (a) < f (b). If the
result is false, either f (c) < f (a) or f (c) > f (b).
Let us look at the first case. Then the value y = f (a) lies between the values f (a)
and f (c) at the end points of [a, c]. Since f (c) < f (a) < f (b), y = f (a) also lies
between the values of f at the end points of [c, b]. Hence there exists x ∈ (c, b) such
that f (x) = y = f (a). Since x > a, this contradicts the fact that f is one-one.
In case, you did not like the way we used y, you may proceed as follows. Fix any y such
that f (c) < y < f (a). By intermediate value theorem applied to the pair (f, [a, c]), there
exists x1 ∈ (a, c) such that f (x1 ) = y. Since f (a) < f (b), we also have f (c) < y < f (b).
Hence there exists x2 ∈ (c, b) such that f (x2 ) = y. Clearly x1 6= x2 .
The second case when f (c) > f (b) is similarly dealt with.

3. Theorem. Let J ⊂ R be an interval. Let f : J → R be continuous and 1-1. Then f is


monotone.
Proof. Fix a, b ∈ J, say with a < b. We assume without loss of generality that
f (a) < f (b). We need to show that for all x, y ∈ J with x < y we have f (x) < f (y).
(i) If x < a, then x < a < b and hence f (x) < f (a) < f (b).
(ii) If a < x < b, then f (a) < f (x) < f (b).
(iii) If b < x, then f (a) < f (b) < f (x).

1
In particular, f (x) < f (a) if x < a and f (x) > f (a) if x > a. (1)

If x < a < y, then f (x) < f (a) < f (y) by (1).


If x < y < a, then f (x) < f (a) by (1) and f (x) < f (y) < f (a) by the last item.
If a < x < y, then f (a) < f (y) by (1) and f (a) < f (x) < f (y) by the last item.
Hence f is strictly increasing.

4. Proposition. Let J be an interval and f : J → R be monotone. Assume that f (J) = I


is an interval. Then f is continuous.
Proof. We deal with the case when f is strictly increasing. Let a ∈ J. Assume that a
is not an endpoint of J. We prove the continuity of f at a using the ε-δ definition.
Since a is not an endpoint of J, there exists x1 , x2 ∈ J such that x1 < a < x2 and hence
f (x1 ) < f (a) < f (x2 ). It follows that there exists η > 0 such that (f (a) − η, f (a) + η) ⊂
(f (x1 ), f (x2 )) ⊂ I.
Let ε > 0 be given. We may assume ε < η. Let s1 , s2 ∈ J be such that f (s1 ) = f (a) − ε
and f (s2 ) = f (a) + ε. Let δ := min{a − s1 , s2 − a}. If x ∈ (a − δ, a + δ) ⊂ (s1 , s2 ),
then, f (a) − ε = f (s1 ) ≤ f (x) < f (s2 ) = f (a) + ε, that is, if x ∈ (a − δ, a + δ), then
f (x) ∈ (f (a) − ε, f (a) + ε).
If a is an endpoint of J, an obvious modification of the proof works.

5. Consider the n-th root function f : [0, ∞) → [0, ∞) given by f (x) := x1/n . We can use
the last item to conclude that f is continuous, a fact seen by us earlier.

6. Let J ⊂ R be an interval and f : J → R be increasing. Assume that c ∈ J is not an


endpoint of J. Then
(i) limx→c− f = l.u.b. {f (x) : x ∈ J; x < c}.
(ii) limx→c+ f = g.l.b. {f (x) : x ∈ J; x > c}.

7. Let the hypothesis be as in the last item Then the following are equivalent:
(i) f is continuous at c.
(ii) limx→c− f = f (c) = limx→c+ f .
(iii) l.u.b. {f (x) : x ∈ J; x < c} = f (c) = g.l.b. {f (x) : x ∈ J; x > c}.
What is the formulation if c is an endpoint of J?

8. Let J ⊂ R be an interval and f : J → R be increasing. Assume that c ∈ J is not an


endpoint of J. The jump at c is defined as

jf (c) := lim f − lim f ≡ g.l.b. {f (x) : x ∈ J; x > c} − l.u.b. {f (x) : x ∈ J; x < c}.
x→c+ x→c−

How is the jump jf (c) defined if c is an endpoint?

9. Let J ⊂ R be an interval and f : J → R be increasing. Then f is continuous at c ∈ J


iff jf (c) = 0.

2
10. Theorem. Let J ⊂ R be an interval and f : J → R be increasing. Then the set D of
points of J at which f is discontinuous is countable.
Proof. Assume that f is increasing. Then c ∈ J belongs to D iff the interval Jc :=
(f (c− ), f (c+ )) is nonempty. For c1 , c2 ∈ D, the intervals Jc1 and Jc2 are disjoint.
(Why?) Thus the collection {Jc : c ∈ D} is a pairwise disjoint family of open intervals.
Such a collection is countable. For, choose rc ∈ Jc ∩ Q. Then the map c 7→ rc from D
to Q is one-one.

11. Let f : R → R be an additive homomorphism. If f is monotone, then f (x) = f (1)x for


all x ∈ R.

You might also like