Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Math 114 PSet 1 PDF

Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

Math 114 Problem Set 1

Roman Berens
with collaboration from George Torres

September 9, 2014

1. Let I denote the set of all nonempty open intervals (a, b) R (we allow a =
and b = +). A choice function is a map c : I Q such that c(I) I for all I.
(a) Give an explicit example of choice function.
There are several different cases we must consider as we construct the function. First, let
a > 0 and b =
6 . Let A = {q N | 1/q < b a}. This set is nonempty by the Archimedean
property, and since it is a subset of N, has a least member q0 . Let B = {n N | n/q0 b}.
Again, this set is nonempty and has a least member p. Since 1/q < b, p > 1. Thus p 1 N
and (p 1)/q < b. In addition,
a = b (b a) <

p 1
p1
=
.
q q
q

Thus r = (p 1)/q is a rational number in (a, b). We define c((a, b)) = r in this case.
Next we let a < 0, a 6= , and b 6= . By the Archimedean property, there is a
natural number n such that n > a. By the first case, there is a rational number r that lies
in (n + a, n + b). Thus r n lies in (a, b). We define c((a, b)) = r n in this case.
We now let a = and b 6= . We define c((a, b)) = bbc, the largest integer less than b.
We now let b = and a 6= . We define c((a, b)) = dae the smallest integer greater than a.
Lastly, we let a = and b = . We define c((a, b)) = 0 in this case.
(b) Show that there exists a translation-invariant choice function, meaning one that
satisfies c(I + r) = c(I) + r for all r Q and for all I I except for I = (, ). (You
may use the Axiom of Choice in your construction.)
We define an equivalence relation on intervals such that for I, J I, I J if I = J + r for
some r Q. Using the Axiom of Choice, we can select a representative member I0 from each
equivalence class. Using the Axiom of Choice again, we can select a rational number q in I0 .
We then define c(I0 ) = q. Since for every element I in the equivalence class, I = I0 + r, we
define c(I) = c(I0 ) + r = q + r.
2. Let K = R(t) be the field of rational functions f (t) with real coefficients, ordered
so that f > 0 if and only if f (t) > 0 for all t sufficiently large.
(a) Which axiom(s) of the real number system does K not satisfy?
The field K does not satisfy the completeness
the set A ={f K | f 2 < t}.
axiom. Consider
We see that this set is bounded above by t, as t f 2 > 0 implies that t > f . However,
this set has no supremum in K, as we can always find a rational function that is arbitrarily

close to x (e.g., a finite number of terms of the Taylor series).

(b) Show that the element  K defined by (t) = 1/t satisfies 0 <  < 1/n for all
positive integers n.
We see that  < 1/n means that there exits an x such that 1/n1/t > 0 for all t x. Since we
can always choose x > n, it follows that  < 1/n. Similary, we see that  > 0 means that there
exits an x such that 1/t > 0 for all t x. We can always choose x > 0, so it follows that  > 0.
(c) Let P (x) = x3 K[x]. What is Q(x) = (P (x + ) P (x)) /?
We see that



3x
3x2 3x
1
1
3
3
Q(x) = t x +
+ 2 + 3 x = 3x2 +
+ 2.
t
t
t
t
t

Note that as shown in part (b), 0 < 1/t < 1/n, so we can take 1/t 0 to yield 3x2 , carrying
over our notion of differentiation into this field.
3. Prove that for any x R, lim sup cos(nx) = 1. (Here n ranges through the positive
n

integers.)
We will show that for any x R and any > 0, there exists positive integers k and
m such that
2k mx < .
We see that this is equivalent to showing that
2

k
<
.
m
m

Since 2k/m Q, x R and the rational numbers are dense in the real numbers, we can find
a p Q such that p = 2k/m and the above holds. In addition, we can multiply k and m by a
natural number without changing p to ensure that m n. Thus we see that
p

x
< .

By the continuity of cos(x), we see that there exists a  such that


cos(2k) cos(mx) = 1 cos(mx) < .
Thus for any x R, we can find a natural number m such that cos(mx) is arbitrarily close
to 1. It then follows that
sup (cos(mx)) = 1.
mn

We then see that



lim sup cos(nx) = lim
n


sup cos(mx) = lim 1 = 1.
mn

4. Let f : [0, 1] [0, 1] be a monotone increasing function. Prove that points where f is
discontinuous form at most a countable set.
Let x0 be a point in [0, 1] [0, 1]. We know that a bounded monotone sequence of real
numbers converges, so if {xn } is a sequence in [0, 1] (x0 , ) which converges to x0 , then

the sequence {f (xn )} converges. We denote this limit by f (x+


0 ) and define f (x0 ) similarly.
3

+
Then f is continuous at x0 if and only if f (x
0 ) = f (x0 ) = f (x0 ). If f is not continuous
at x0 , that it must have jump discontinuity at x0 . For each jump discontinuity, there is
+
a rational number in the range of f between f (x
0 ) and f (x0 ). Since f is monotonic, the
rational number will be different for every jump discontinuity. Thus there is an injection
from the set of points at which f is discontinuous to a subset of the rational numbers. Since
the rational numbers are countable, the set of points at which f is discontinuous is countable.

5. Define fn : [0, 2] [0, 1] by fn (x) = 1/q n if x = p/q is rational (expressed in lowest terms), and fn (x) = 0 otherwise.
(a) At what points x [0, 2] is fn continuous?
We know that fn is continuous at x if for each  > 0 there is a > 0 such that if |x0 x| < ,
then |f (x0 )f (x)| < . Since the rational numbers are dense in the real numbers, for any x Q,
we can find an irrational x0 [0, 2] such that |x0 x| < . However, |f (x0 ) f (x)| = 1/q n ,
so we cannot find a for every  such that the above holds. This means fn cannot be
continuous at any rational point. However, if x and x0 are irrational, |f (x0 ) f (x)| = 0, so
the condition holds for any . If x0 is rational, we see that |f (x0 ) f (x)| = 1/q n . Thus for a
given , we pick the smallest q N such that 1/q n < . We then let = 1/q. Since there
will be a rational number p/q such that p/q < x < (p1 )/q, and all rational numbers in this
range will have denominators greater than q, we see that this choice of satisfies the condition.
(b) Where is f1 differentiable?
Since differentiability implies continuity, we only consider the irrational points in [0, 2].
Let a be one such point. We apply the definition of the derivative:
f1 (x)
f1 (x) f1 (a)
= lim
xa x a
xa
xa

f 0 (a) = lim

We see that as x approaches a, f1 (x) either takes on the value 0 or 1/q. As x gets closer to a.
the value of q will increase. Thus we see that f1 (x) goes to zero faster than x a, so the
f 0 (a) = 0 for any irrational a. In fact, this result holds for any fn , as fn (a) either takes on
the value 0 or 1/q n .

(c) Show f30 ( 2) = 0. (Hint: consider (p/q)2 2.)


We seein part (b) that the value of the derivative of fn must be zero wherever it is defined.
Since 2 is irrational, f30 is defined there, and takes the value 0.
6. Let (X, d) be a connected metric space containing at least 2 points. Show that X
has at least the cardinality of the continuum, that is, show that |X| |R|.
Let a and b be the two points in X. Consider the set of distances, C = {d(a, x) | x X}.
We know that this contains at least two numbers, namely d(a, a) = 0 and d(a, b) 6= 0. We
consider the interval (0, d(a, b)) R. We assume that there is a real number c in this interval
such that there is no x X with d(a, x) = c. If this is true, we can decompose X as
X = {x X | d(a, x) < c} {x X | d(a, x) > c}.
However, both of these sets are open, so we have found a disjoint open cover of X. This contradicts the fact that X is connected. Thus there can be no such c, so for every c (0, d(a, b)),
4

there is at least one x with d(a, x) = c. Thus we have a injection from the an interval in R to
the set C defined above. Since we also have an injection from C to X, and since any interval
in R has the same cardinality as R, we see that |X| |R|
7. Let B R be a Borel set and let f : R R be a continuous map. Prove that
f 1 (B) is a Borel set.
We shall show that any Borel set can be built from open and closed sets, using countable intersections and unions. Since the Borel sets form a -algebra, they must contain the
open intervals. In addition, they must be closed under complements and countable unions
(which implies that they must be closed under countable intersections). This means that the
Borel sets must contain the closed sets, as well as any countable intersection of open sets
(the G sets) and any countable union of closed sets (the F sets). Lastly, the Borel sets are
defined as the minimal -algebra that contains the open sets. We see that being a -algebra
implies the membership of the sets described above. However, the condition of minimality
ensures that there are no other sets. If there was a set that could not be built from open and
closed sets using countable intersections and unions, its membership would not be required
by the definition of a -algebra, so it would not be included in the minimal -algebra.
As described above, a general Borel set U can be written as
[
[
\
\
U=
On
C n On C n
n

where the On are open sets and Cn are closed sets. We know that the inverse image of an
open or closed set under a continuous function must be open or closed, respectively. We also
know that the inverse image of a union or intersection of sets under a continuous function is
the union or intersection of the inverse images of the sets. It thus follows that
[
[
\
\
f 1 (U ) =
f 1 (On ) f 1 (Cn )
f 1 (On ) f 1 (Cn )
n

[
n

On0

[
n

Cn0

On00

Cn00

where the primes indicate a new collection of open and closed sets.

You might also like