Chapter 7

Sequences of Functions
Sequences of functions play in important role approximation theory. They can be
used to show a solution of a dierential equation exists. We recall in Chapter three
we dene a sequence to be a function whose domain is the natural numbers. Thus,
if f
n
(x) : D R for each n N, then {f
n
} is sequence of functions. We need a
notion of convergence.
Denition 7.1. We say the sequence of functions {f
n
} dened on a set D converges
pointwise, if for each x D, the sequence of real numbers {f
n
(x)} converges. We
set f(x) = lim
n
f
n
(x).
We would like to know what f inherits from the f
n
. In general, the answer is
not much.
Example 7.2. Consider f
n
(x) = x
n
on [0, 1]. Note that f
n
C

(0, 1) for all

n N. However, the limit
f(x) = lim
n
f
n
(x) =
_
0 x [0, 1)
1 x = 1
is not even continuous on [0, 1].
Example 7.3. Consider f
n
(x) =
sin nx

n
on R. Once again f
n
C

(R for all
n N, and the pointwise limit is f(x) = 0 is also innite dierentiable. However,
f

n
(x) =

ncos nx has no limit and certainly does not converge to f

(x) = 0. Thus,
we see
( lim
n
f
n
(x))

= lim
n
f

n
(x)
Example 7.4. Back in Chapter one in Example 1.5 we considered the sequence of
functions given by f
n
(x) = nxe
nx
2
on [0, 1]. One checks the limit is f(x) = 0. Once
again both the sequence and the limit have innite many derivatives. However,
lim
n
_
1
0
f
n
(x)dx = lim
n

1
2
e
nx
2

1
0
=
1
2
,
91
92 7. Sequences of Functions
while
_
1
0
lim
n
f
n
(x)dx =
_
1
0
lim
n
0 dx = 0.
Thus
lim
n
_
1
0
f
n
(x)dx =
_
1
0
lim
n
f
n
(x)dx.
So what does f inherit from f
n
? We need a stronger form of convergence.
Denition 7.5. A sequence of functions {f
n
} on D is uniformly convergent on D
if, for all > 0, there exists a N N such that |f
n
(x) f(x)| < for all x D and
n N.
Lemma 7.6. A sequence of functions {f
n
} on D does not converge uniformly to
its pointwise limit f if there exists a
0
> 0 and a subsequence {f
n
k
} of {f
n
} and a
sequence {x
k
} D such that |f
n
k
(x
k
) f(x
k
)| for all k N.
Proof. In logic notation uniform convergence is
( > 0)(N N)(x D)(n N)(|f(x) f
n
(x)| < ).
The negation of this is
(
0
> 0)(N N)(x D)(n N)(|f(x) f
n
(x)|
0
).
That is
(
0
> 0)(k N)(x
k
D)(n
k
k)(|f(x
k
) f
n
k
(x
k
)|
0
)
as advertised.
Example 7.7. Show that f
n
(x) = x
n
on [0, 1] does not converge uniformly on
[0, 1].
Proof. We construct the sequence {x
k
} to converge to the trouble spot. Here, the
trouble is at x = 1 (a graph of several f
n
will help to see this). We pick
0
= 1/2,
n
k
= k and x
k
= (1/2)
k
. Then
|f(x
k
) f
n
k
(x
k
)| = |f
n
k
(x
k
)| =
1
2

0
,
and {f
n
} does not converge uniformly on [0, 1].
Example 7.8. Show that f
n
(x) = x
n
on [0, 1/2] does converge uniformly on
[0, 1/2].
Proof. We just need to show that the N in the denition of convergence does not
depend on x. The pointwise limit f = 0. We calculate
|f
n
(x) f(x)| = |f
n
(x) 0|
_
1
2
_
n
.
Thus, given any > 0 we choose N N so that (1/2)
n
< for all n N (how do
we do this?). Then |f
n
(x) f(x)| < for all x [0, 1/2] and n N.
Example 7.9. Show that the sequence of functions given by f
n
(x) = nxe
nx
2
on
[0, 1] does not converge uniformly on [0, 1].
7.1. Preservation Theorems 93
Proof. A graph of the f
n
is given in Example 1.5. The peak of the curves (the
maximum of f
n
occurs at x = 1/

2n, and f
n
(1/

2n) =
_
n
2
e
1/2
. This suggest
choosing
0
equal to any positive number, n
k
= k, x
k
= 1/

2k. Then
|f
n
k
(x
k
) f(x
k
)| = |f
n
k
(x
k
)| =
_
k
2
e
1/2
which tends to innity.
We might expect uniform convergence to imply f inherits the properties of f
n
.
This is not quite true.
Example 7.10. Consider again f
n
(x) =
sin nx

n
on R. The pointwise limit is f(x) =
0. Moreover, the convergence is uniform since
|f
n
(x) f(x)| = |f
n
(x) 0|
1

n
which can be made small independently of x. However, recall that f

n
does not
converge to f

.
Example 7.11. The sequence of functions f
n
(x) = x/n on R does not converge
uniformly on R. Indeed, set
0
= 1, n
k
= k, and x
k
= k. Then |f
n
k
(x
k
) f(x
k
)| =
|1 0| . However, f inherits the continuity of the f
n
and f

n
converges to f

.
7.1. Preservation Theorems
In this section we nd conditions on the sequence {f
n
} so that its properties are
retained by the pointwise limit f.
Theorem 7.12. Let D be any nonempty subset of R and suppose f
n
C(D) for
each n N. If {f
n
} converges uniformly to f on D, then f is continuous on D.
Proof. Since f
n
converges uniformly on D, for any > 0, there exists N() such
that
(7.1) |f
n
(x) f(x)| <
for all x D and n N. Since each f
n
is continuous on D, for any > 0, x
0
D,
and n N, there exists (, x
0
, n) such that
(7.2) |f
n
(x) f
n
(x
0
)| <
for all x D and |x x
0
| < .
We need to show f is continuous on D. Let > 0 and x
0
D. Equation (7.1)
shows the existence of a N
0
N so that
|f
n
(x) f(x)| <

3
for all n N
0
and all x D. Equation (7.2) provides a (, x
0
, N
0
) such that
|f
N0
(x) f
N0
(x
0
)| <

3
94 7. Sequences of Functions
for all x D, |x x
0
| < . Thus
|f(x) f(x
0
)| = |f(x) f
N0
(x) + f
N0
(x) f
N0
(x
0
) + f
N0
(x
0
) f(x
0
)|
|f(x) f
N0
(x)| +|f
N0
(x) f
N0
(x
0
)| +|f
N0
(x
0
) f(x
0
)|
<

3
+

3
+

3
=
for all x D with |x x
0
| < , and hence f C(D).
Example 7.13. Show (again) that f
n
(x) = x
n
on [0, 1] does not converge uniformly
on [0, 1].
Proof. We note (again) that the f
n
are continuous. Since the pointwise limit f is
not continuous, the convergence cannot be uniform.
Example 7.14. Consider
f
n
(x) =
_
1
n
x Q
0 x Q
c
.
The pointwise limit is f = 0, and it is easily seen that f
n
f uniformly on [0, 1].
While the f
n
are discontinuous the limit f is smooth. Thus uniform convergence
preserves good properties, not bad, [6].
Theorem 7.15. Suppose f
n
R[a, b] for all n N. Furthermore, suppose {f
n
}
converges uniformly to f on [a, b]. Then f R[a, b] and
(7.3) lim
n
_
b
a
f
n
(x) dx =
_
b
a
lim
n
f
n
(x) dx.
Proof. We rst show f is integrable. Let > 0 be given. Choose N N so that
|f
n
(x) f(x)| <

3(b a)
for all n N and x [a, b]. It follows that U(P, f f
n
) < /3 and L(P, f f
n
) >
/3 for any partition P of [a, b]. Since F
N
is Riemann integrable, there exists a
partition of [a, b] so that
U(P, f
N
) L(P, f
N
) < /3.
For any two bounded functions h, g and any partition, note that
U(P, h + g) U(P, h) + U(P, g), L(P, h + g) L(P, h) + L(P, g).
Thus
U(P, f) L(P, f) U(P, f f
N
) +
_
U(P, f
N
) L(P, f
N
)
_
L(P, f f
N
)
< 3

3
.
Now that we know f is Riemann integrable, a simple calculation shows

_
b
a
f
n
(x) dx
_
b
a
f(x) dx

_
b
a
|f
n
(x) f(x)| dx <
_
b
a

3(b a)
<
for all n N and (7.3) follows.
7.1. Preservation Theorems 95
If we assume that f is Riemann integrable, we may drop the uniform conver-
gence.
Theorem 7.16. (Bounded-Convergence Theorem) Let f
n
R[a, b] for all n N
and suppose the pointwise limit f is also Riemann integrable on [a, b]. If a M exists
so that |f
n
(x)| M for all x [a, b] and n N, then
lim
n
_
b
a
f
n
(x) dx =
_
b
a
lim
n
f
n
(x) dx.
Proof. For a proof at the level of this book see [5].
Example 7.17. Note that the bounded convergence theorem applies to the func-
tion f
n
(x) = x
n
on [0, 1] while Theorem 7.15 does not.
Theorem 7.18. Suppose
i) f
n
C
1
(I)
ii) lim
n
f
n
(x
0
) converges for some x
0
I.
iii) f

n
converges uniformly on I
Then
(7.4) lim
n
f

n
(x) =
_
lim
n
f
n
(x)
_

Proof. Suppose f

n
converge uniformly on I to g. By the Fundamental Theorem
of Calculus
f
n
(x) = f
n
(x
0
) +
_
x
x0
f

n
(x) dx.
Theorem 7.15 applies to f

n
, and we therefore conclude f
n
(x) converges pointwise
to, say, f. In fact, the previous equation shows
f(x) = f(x
0
) +
_
x
x0
g(x) dx.
The Fundamental Theorem of Calculus applies again, and we see that f

= g. This
is (7.4).
96 7. Sequences of Functions
7.2. Old Final
Work Three problems from 1-4 and Two problems from 5-7. (Total of ve prob-
lems worked).
1. Suppose f is continuous on [a, b]. Prove that f is Riemann integrable.
2. Suppose f is continuous on [a, b] and
_
c
a
f = 0 for all c [a, b]. Show that f = 0
on [a, b].
3. Consider the function
h(x) =
_
1 x = 1/n, n N,
0 x = 1/n,
on the interval [0, 1].
(a) Prove h R[0, 1]. State any theorems you use.
(b) Set F(x) =
_
x
0
h(t)dt. Find a formula for F. Justify your claims.
4. Suppose f is a bounded function on [a, b] and that f R[c, d] for all [c, d] (a, b).
Show f R[a, b].
5. Suppose that f : R R is uniformly continuous on R and for n N, let
f
n
(x) = f(x + 1/n) for x R. Show that {f
n
} converges uniformly to f on R.
6. Suppose 0 < a < 2. Show that lim
n
_
2
a
e
nx
2
dx = 0. What happens if a = 0?
As usual state any theorems you use.
7. Suppose f
n
converges to f uniformly on a set D and each f
n
is continuous on
D. Prove f is continuous. Show that the statement is false if either the uniform
convergence or the continuity of the f
n
is dropped.
7.3. Homework 97
7.3. Homework
Exercise 7.1. This is the same as Problem 6.25. Work it again using on of the
theorems in this chapter. Suppose f is continuous on [0, 1]. Dene g
n
(x) = f(x
n
)
for n N. Prove that lim
n
_
1
0
g
n
(x) dx = f(0).
Exercise 7.2. Give an example of a sequence of continuous functions which con-
verge to a continuous function but where the convergence is not uniform.
Exercise 7.3. Give an example of a sequence of functions discontinuous everywhere
which converge uniformly to a continuous function.
Exercise 7.4. Let f
n
(x) = 1/(nx+1) and g
n
(x) = x/(nx+1) for x (0, 1). Show
that {f
n
} converges uniformly on (0, 1), but {g
n
} does not converge uniformly on
(0, 1).
Exercise 7.5. Show that f
n
(x) = sin
n
x does not converge uniformly on [0, /2].
Exercise 7.6. Let {f
n
} be a sequence of continuous functions on a nonempty set
E such that f
n
(x) f(x) uniformly on E as n , and let {x
n
} be a sequence
in E converging to an x E. Does it follow that f
n
(x
n
) converges to f(x)?
Exercise 7.7. Consider the following functions dened for x 0
(a)
x
n
n
.
(b)
x
n
n + x
n
.
(c)
x
n
1 + x
n
.
(d)
x
n
e
x/n
.
Discuss the convergence and the uniform convergence of these sequences and the
continuity of the limit functions. In the case on nonuniform convergence consider
the sequences on an appropriate interval in E.
Exercise 7.8. Prove the following result. If {f
n
} is uniformly convergent on a set
D and each f
n
is bounded, then f
n
is uniformly bounded. That is, there exists a
M such that |f
n
(x)| M for all x D and n N. Use this result to give another
proof that the sequence in Example 7.9 does not converge uniformly.
Exercise 7.9. Let f
n
(x) = x + 1/n on R. Show that f
n
converges uniformly on
R, but f
2
does not.
Exercise 7.10. Let f
n
(x) =
_
x
2
+ 1/n
2
on R. Show that f
n
converges uniformly
to |x| on R.
Exercise 7.11. Suppose {f
n
} and {g
n
} converge uniformly on a set D.
(a) Show that {f
n
g
n
} converges uniformly on D.
(b) If each f
n
and g
n
are bounded, show that {f
n
g
n
} converges uniformly on D.
98 7. Sequences of Functions
Exercise 7.12. If 0 < a < 2, show that lim
n
_
2
a
e
nx
2
dx = 0. What happens if
a = 0?
Exercise 7.13. Show that f
n
=
_
x
2
+ 1/n
2
converges uniformly to |x| on R.
[Hint: Rationalize to show |f
n
(x)

x
2
| 1/n for each n.
Exercise 7.14. If f
n
converges uniformly to f on a set E and each f
n
is continu-
ous, prove f is continuous. Show that the statement if false if either the uniform
convergence or the continuity of the f
n
are dropped.
Exercise 7.15. Suppose that f : R R is uniformly continuous on R and for
n N, let f
n
(x) = f(x +1/n) for x R. Show that {f
n
} converges uniformly to f
on R.
Exercise 7.16. Prove the Monotone Convergence Theorem. Let {f
n
} be a se-
quence of monotone increasing Riemann integrable functions. That is, f
1
(x)
f
2
(x) . . . f
n
(x) . . . for all x [a, b]. If the pointwise limit f is Riemann
integrable, then lim
n
_
b
a
f
n
(x) dx =
_
b
a
f(x) dx.
Exercise 7.17. (Dinis Theorem) Suppose f
n
converges to f pointwise on [a, b]
with both f
n
and f continuous on [a, b]. If the sequence is monotone decreasing,
f
n
(x) f
n+1
(x) for all n N and x [a, b], then f
n
f uniformly on [a, b].
Bibliography
[1] T.M Apostol, Mathematical Analysis, Addison-Wesley, 1964.
[2] Robert G. Bartle, The Elements of Real Analysis, John Wiley & Sons, 1976.
[3] Frank Dangello & Michael Seyfried, Introductory Real Analysis, Houghton
Miin Company, 2000.
[4] Ralph P. Boas Jr., A Primer of Real Functions, 1981, QA331.5 B63x.
[5] A Truly Elementary Approach to the Bounded Convergence Theorem, Jonathan W.
Lewin, The American Mathematical Monthly Vol. 93, No. 5, 1986, pp. 395-397,
http://www.jstor.org/pss/2323608
[6] Gelbaum & Omstead, Counter Examples in Analysis, Holden-Day, San Francisco,
1964.
99