Scribd
Upload
93 views6 pages

HW140B 2 Solutions

This document provides solutions to 6 problems from a calculus homework assignment. The problems cover topics like limits, derivatives, integrals, and Hölder's inequality. For each problem, the solution provides the relevant work or steps to solve the given problem, with explanations of the mathematical reasoning. The longest solution is for Problem 3, which takes about 4 paragraphs to fully explain the proof.

Uploaded by

Idris Addou
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
93 views6 pages

HW140B 2 Solutions

This document provides solutions to 6 problems from a calculus homework assignment. The problems cover topics like limits, derivatives, integrals, and Hölder's inequality. For each problem, the solution provides the relevant work or steps to solve the given problem, with explanations of the mathematical reasoning. The longest solution is for Problem 3, which takes about 4 paragraphs to fully explain the proof.

Uploaded by

Idris Addou
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

MATH 140B - HW 2 SOLUTIONS

Problem 1 (WR Ch 5 #11). Suppose f is dened in a neighborhood of x, and suppose


f

(x) exists. Show that
lim
h0
f (x +h) + f (x h) 2f (x)
h
2
= f

(x).
Show by an example that the limit may exist even if f

(x) does not.
Solution. Since f

(x) = lim
h0
f

(x+h)f

(x)
h
, and since h 0 iff h 0, if we replace h by h in
this expression, we have
f

(x) = lim
(h)0
f

(x +(h)) f

(x)
(h)
= lim
h0
f

(x) f

(x h)
h
.
Therefore,
f

(x) = lim
h0
f

(x +h) f

(x)
h
= lim
h0
lim
h
1
0
f (x+h)f (x+hh
1
)
h
1
lim
h
2
0
f (x)f (xh
2
)
h
2
h
,
and letting h
1
= h = h
2
, i.e. taking them all to zero at the same rate (which we can do by
Theorem 4.2), we have
f

(x) = lim
h0
f (x+h)f (x+hh)
h

f (x)f (xh)
h
h
= lim
h0
f (x +h) + f (x h) 2f (x)
h
2
.
For the second portion, we would like to nd an f so that f

(x) = |x|. One such choice of f
could be f (x) =
_
x
0
|t | dt =
1
2
x|x|. Now let x =0. By continuity of f at x =0 we have
lim
h0
_
f (x +h) + f (x h) 2f (x)
_
= f (0) + f (0) 2f (0) =0.
Also, clearly lim
h0
h
2
=0. So by LHpitals rule we have
lim
h0
f (x +h) + f (x h) 2f (x)
h
2
LH
= lim
h0
f

(x +h) f

(x h)
2h
= lim
h0
|0+h| |0h|
2h
=0.
That means the limit exists at x =0, but f

(x) =|x| is not differentiable at 0, so f

(0) does not
exist.
Problem 2 (WR Ch 5 #12). If f (x) = |x|
3
, compute f

(x), f

(x) for all real x, and show
that f
(3)
(0) does not exist.
1
Solution. For x =0, |x| is a differentiable function with derivative
sgn(x) =
_
1 if x >0
1 if x <0
.
Thus by the chain rule in the rst line and by the product rule in the second line,
f

(x) =3|x|
2
sgn(x) =3x|x|.
f

(x) =3|x| +3x sgn(x) =3|x| +3|x| =6|x|.
Checking the cases for x =0 by hand, we have
f

(0) = lim
h0
f (x +h) f (x)
h
= lim
h0
|h|
3
0
h
= lim
h0
h|h| =0.
f

(0) = lim
h0
f

(x +h) f

(x)
h
= lim
h0
3h|h| 0
h
= lim
h0
3|h| =0.
f

(0) = lim
h0
f

(x +h) f

(x)
h
= lim
h0
6|h| 0
h
=6 lim
h0
|h|
h
=DNE
Problem 3 (WR Ch 6 #2). Suppose f 0, f is continuous on [a, b], and
_
b
a
f (x) dx =0.
Prove that f (x) =0 for all x [a, b].
Solution. Assume by way of contradiction that there is some y [a, b] such that f (y) >0, and
let =
f (y)
2
. Since f is continuous, there exists a >0 such that
0 <|y x| < = | f (y) f (x)| < =
f (y)
2
,
for x [a, b]. This last inequality gives us
f (y) f (x) | f (y) f (x)| <
f (y)
2
= f (x) >
f (y)
2
>0.
Let I = (y , y +) [a, b]. What we have shown so far is that if x I , then f (x) >
f (y)
2
> 0.
Now, given some partition P of [a, b], we make a renement P

by adding if necessary (and


if possible) a point in (y , y) [a, b] and a point in (y, y +) [a, b] so that y (x
k1
, x
k
) I
with x
k1
, x
k
P

for some 1 k n. Then


0 =
_
b
a
f (x) dx =sup
P
L(P, f ) L(P

, f ) =
n

i =1
_
inf
x
i 1
xx
i
f (x)
_
x
i

f (y)
2
x
k
>0,
a contradiction.
2
Problem4 (WR Ch 6 #4). If f (x) =0 for all irrational x, f (x) =1 for all rational x, prove
that f R on [a, b] for any a <b.
Solution. For any partition P ={a =x
0
, x
1
, . . . , x
n1
, x
n
=b}, we have
L(P, f ) =
n

i =1
_
inf
x
i 1
xx
i
f (x)
_
x
i
=
n

i =1
(0)x
i
=0 by the density of Q
c
,
U(P, f ) =
n

i =1
_
sup
x
i 1
xx
i
f (x)
_
x
i
=
n

i =1
(1)x
i
=(b a) by the density of Q.
_
b
a
f =sup
P
L(P, f ) =0 =(b a) =inf
P
U(P, f ) =
_
b
a
f ,
so f R on [a, b].
Problem5 (WR Ch 6 #5). Suppose f is a bounded real function on [a, b], and f
2
Ron
[a, b]. Does it follow that f R? Does the answer change if we assume that f
3
R?
Solution. In the rst case, we have the following counterexample. Let
f (x) =
_
1 if x (Q
c
[a, b])
1 if x (Q[a, b])
.
Then by the previous proof with 1 in place of 0, f R on [a, b]. But f
2
1 R on [a, b]. So
it does not necessarily follow that if f
2
R on [a, b] then f R.
In the second case, it does necessarily follow that if f
3
R on [a, b] then f R by the
following proof. The reason this works for f
3
and not for f
2
is that the inverse of the cube
function on R is well-dened and is (x) =
3

x. The square function does not have a well-


dened inverse on all of R (since if y =x
2
then x =

y).
By Theorem 6.11, since is continuous on all of R, then
( f
3
(x)) =
3
_
f
3
(x) = f (x) is in R on [a, b].
Problem6 (WR Ch 6 #10). Let p and q be positive real numbers such that
1
p
+
1
q
=1.
3
Prove the following statements.
(a) If u 0 and v 0, then
uv
u
p
p
+
v
q
q
.
Equality holds if and only if u
p
=v
q
.
(b) If f R(), g R(), f 0, g 0, and
_
b
a
f
p
d=1 =
_
b
a
g
q
d,
then
_
b
a
f g d=1.
(c) If f and g are complex functions in R(), then

_
b
a
f g d

__
b
a
| f |
p
d
_
1/p
__
b
a
|g|
q
d
_
1/q
.
This is Hlders inequality. When p = q = 2 it is usually called the Schwarz
inequality.
(d) Showthat Hlders inequality is also true for the improper integrals described in
Exercises 6.7 and 6.8.
Solution.
Claim. f (x) =e
x
is a convex function.
Let x <t < y. By the Mean Value Theorem, there exists some a (x, t ) such that
f (t ) f (x) =(t x) f

(a) which means f

(a) =
f (t ) f (x)
t x
.
Once again, by the Mean Value Theorem, there exists some b (t , y) such that
f (y) f (t ) =(y t ) f

(b) which means f

(b) =
f (y) f (t )
y t
.
Notice that f

(x) =e
x
>0 for all x R. This means that f

(x) is strictly increasing. Therefore,
since a <b, we have f

(a) f

(b), so
f (t ) f (x)
t x
= f

(a) f

(b) =
f (y) f (t )
y t
.
Now for any (0, 1) we have x <(x +(1)y) < y, so letting t =(x +(1)y) the above
4
inequality becomes
f (t ) f (x)
(x +(1)y) x

f (y) f (t )
y (x +(1)y)
f (t ) f (x)
(1)

(y x)

f (y) f (t )

(y x)
f (t ) f (x) (1) f (y) (1) f (t )
f (t ) f (x) +(1) f (y)
f (x +(1)y) f (x) +(1) f (y),
so f (x) =e
x
is convex.
(a) From here, we let =
1
p
, so that (1) =
1
q
. The desired result is trivial if u =0 or v =0, so
assume they are both strictly positive. Letting x =logu
p
and y =logv
q
, the above inequality
becomes
e
1
p
logu
p
+
1
q
logv
q

1
p
e
logu
p
+
1
q
e
logv
q
e
logu+logv

u
p
p
+
v
q
q
uv
u
p
p
+
v
q
q
.
(b) By part (a), for every x [a, b] we have
f (x) g(x)
( f (x))
p
p
+
(g(x))
q
q
.
Therefore, taking integrals, we have
_
b
a
f g d
_
b
a
f
p
d
p
+
_
b
a
g
q
d
q
=
1
p
+
1
q
=1.
(c) If
_
b
a
| f | d=0 or
_
b
a
|g| d=0 the inequality is trivial. Otherwise, let A =
_
_
b
a
| f |
p
d
_
1/p
>
0 and let B =
_
_
b
a
|g|
q
d
_
1/q
>0, and let
F(x) =
| f (x)|
A
and G(x) =
|g(x)|
B
.
These functions satisfy the hypotheses of part (b), so
_
b
a
FGd1
_
b
a
| f |
A
|g|
B
d1

_
b
a
f g d

_
b
a
| f | |g| d AB =
__
b
a
| f |
p
d
_
1/p
__
b
a
|g|
q
d
_
1/q
.
5
(c) Since x |x| and x x
1/p
and x x
1/q
are continuous functions (for x >0), we have

_
1
0
f g d

lim
c0
_
1
c
f g d

= lim
c0

_
1
c
f g d

lim
c0
_
__
1
c
| f |
p
d
_
1/p
__
1
c
|g|
q
d
_
1/q
_
=
_
lim
c0
__
1
c
| f |
p
d
_
1/p
_

_
lim
c0
__
1
c
|g|
q
d
_
1/q
_
=
_
lim
c0
_
1
c
| f |
p
d
_
1/p

_
lim
c0
_
1
c
|g|
q
d
_
1/q
=
__
1
0
| f |
p
d
_
1/p
__
1
0
|g|
q
d
_
1/q
,
assuming the integrals are all nonzero and nite. If they are not, the inequality is trivial. The
proof follows similarly for
_

a
.
6

You might also like