Sunteți pe pagina 1din 8

Overview of Real Analysis

EECE 506

May 31, 2017

1 Axioms of the real number line


The set of real numbers is denoted R, and is defined by the following sixteen
axioms.

1.1 Arithmetic axioms


There is a binary operation + : R R R, typically written as x + y to
mean +(x, y), with the following properties:
1. a + (b + c) = (a + b) + c (associativity)

2. There is an element 0 R with a + 0 = 0 + a = a.

3. For every a R there is a b R with a + b = 0. This element is


typically denoted as a.

4. a + b = b + a (commutativity)
There is a binary operation : R R R, typically written as x y
or xy to mean (x, y), with the following properties:
5. a (b c) = (a b) c (associativity)

6. There is an element 1 R with a 1 = 1 a = a.

7. For every a R {0} there is a b R {0} with a b = 1. This


element is typically denoted as a1 or 1/a.

8. a b = b a (commutativity)

9. a (b + c) = (a b) + (a c) (distributive property)
These are the arithmetic axioms, or field axioms, of the real numbers.

1
1.2 Order axioms
The order axioms describe how real numbers are arranged relative to one
another, to form a real number line. They are as follows:
There is a binary relation <: R R {true, false} where we write
x < y to mean that < (x, y) = true, which has the following properties:

10. a a

11. If a < b then b a

12. If a < b and b < c then a < c (transitivity)

13. For any x, y R, either x = y, x < y, or y > x.

Mathematically we say that R is totally ordered by these four axioms. We


also have these axioms tying the arithmetic operations to the order relation
<:

14. If x < y then x + z < y + z

15. If x < y and z > 0, then xz < yz

1.3 The completeness axiom


The completeness axiom describes how the real numbers form a continuous
fabric of numbers without any gaps.

16. Any nonempty set S R that has an upper bound, has a least upper
bound.

That last axiom requires some definitions. For a set S R, we say


M R is an upper bound of S if M x for every x S. For example, 5 is
an upper bound of (0, 1].
A least upper bound of S is a number m that is not larger than any other
upper bound. That is to say, if M is an upper bound of S, then m M .
We also call this number the supremum of S, and denote it sup[S]
Some sets of real numbers have a least element. A set T has a least
element m if m T and m x for every x T . Sets with least elements
include [0, 1] and {1, 2, 3}. Sets without least elements include Z, the set of
integers, the interval (, 0], and the interval (0, ) there is no smallest
positive number.

2
The completeness axiom says that if a nonempty set S has an upper
bound, we can imagine a set T of all upper bounds of Sand that T has a
least element.
Note that this axiom only holds for a nonempty set. That is because
any number is an upper bound of the empty set. For example, -5 is an
upper bound of the empty set because by vacuous implication, if x ,
then x < 5. Since every real number is an upper bound of the empty set,
and there is no ssmallest real number, then the empty set cannot have a
least upper bound.

2 Facts
All of these facts follow from the axioms listed above, and we go over the
proofs in class.

Theorem 2.1. If a + a = a then a = 0.

Proof. 0 = a + (a) = (a + a) + (a) = a + (a + (a)) = a + 0 = a.

Note the use of axioms here: a has an additive inverse by axiom 3; we


can substitute (a + a) for (a) using the precondition of the theorem; then
we invoke associativity (axiom 1) and the identity element (axiom 2).

Theorem 2.2. a 0 = 0

Proof. a0 = a(0+0) = (a0)+(a0), and we have already established


that only the number 0 can have this property.

Theorem 2.3. If ab = 0 then a = 0 or b = 0.

Proof. If a = 0, when a = 0 or b = 0. If a 6= 0, then there is an a1 ,


and b = 1b = (a1 a)b = a1 (ab) = a1 0 = 0, because of the previous
theorem.

Theorem 2.4. a = (1) a.

Proof. 0 = a 0 = a (1 + 1) = (1) a + a 1 = (1) a + a.


Therefore a and (1) a are additive inverses.

Corollary 2.5. (1)(1) = 1

So far, all of these theorems only require the field axioms. The following
theorems also use the order axioms.

3
Theorem 2.6. The positive numbers are closed under addition and multi-
plication. That is, if x > 0 and y > 0, then x + y > 0 and xy > 0.

Proof. If x > 0 then x+y > 0+y = y, and transitivity gives us 0 < y < x+y.
If x > 0 and y > 0 then axiom 15 says that xy > 0y = 0.

Theorem 2.7. If x > 0 and y > 0, then (x < y) if and only if (x2 < y 2 ).

Proof. You can verify that (y x)(y +x) = y 2 x2 . If y > x then (y x) > 0
(axiom 14), and (y x)(y + x) > 0 owing to the closure of the positive
numbers under + and .
Contrariwise, if y 2 x2 > 0 then (y x)(y + x) > 0. Since (y + x) is
positive it is nonzero and has an inverse, and that inverse has to be positive
(why?). We then have (y x) = (y x)(y + x)(y + x)1 > 0(y + x)1 = 0.

None of these theorems involve the final axiom. Just what do we get from
that final, completeness axiom? Heres a big one, called the Archimedean
principle:

Theorem 2.8. For any x R, there is an integer n with n 6= x < n + 1.

Proof. Let A be the set of integers n x. Lets assume A is nonempty,


and has a least upper bound we will denote m = sup[A]. Since m is the
least upper bound, (m 1) is not an upper bound of A, so there is at least
one n x greater than (m 1). We can write (m 1) < n m, and add
one to get m < n + 1. The number (n + 1) cant be in A because m is an
upper bound for A; by the definition of A, x < n + 1; but since n A,
n x < n + 1.

What if A was empty? It doesnt seem like that is possible, but why
can we assume that {n Z|n x} is a nonempty set? We can because if
it was empty, then every integer is greater than x, and at least one integer
is less than (-x) (see exercise 2). We could then find an integer n with
n (x) < n + 1, but then (n + 1) < x (n)contradicting the
assumption that A was empty.
This principle is really useful, because it allows us to construct very
concrete values, using whole numbers, that are greater than or less than or
between numbers. We get the following corollaries from this theorem:

Corollary 2.9. There is no infinite real number.

Corollary 2.10. There is no infinitesimal real number.

4
Why not? Because for any real x there is an integer larger than x,
and for every positive real x there is an integer n with x1 < n + 1, so
0 < 1/(n + 1) < x. There is is always a bigger number, and always a tinier
one. There is no biggest or tiniest nonzero real number.

Corollary 2.11. If x < y there is a third number z with x < z < y.

Proof. Since (y x) > 0, (y x)1 > 0 (see exercise 4), and there is an
integer N > (y x)1 > 0. Thus we have 0 < (1/N ) < (y x), and
x + 0 < x + (1/N ) < x + (y x) so x < x + (1/N ) < y.

This is the general approach to using the Archimedean principle: by


some gymnastics we find a number N greater than some expression that
can be rearranged into an inequality.
Heres an example of this strategy:

Theorem 2.12. Every positive number N has a square root. That is, for
every N R, there is an m R with m m = N .

Proof. Let S = {n|n n N }. This is nonempty because 0 < N , and


0 0 = 0 < N . It has an upper bound because if 1 N , 1 1 N 1 = N
so 1 S; and if N 1, N N 1 N = N , so N S.
Now let m = sup[S]. We will show that m2 = N , because there are only
two other possibilitiesm2 < N , or N < m2 that we will show impossible
by the Archimedean principle.
First, what if m2 < N ? Then (N m2 ) > 0 and there is an integer
n > (2m + 1)/(N m2 ). We can rearrange this to get

N m2 > (2m + 1)/n


N > m2 + (2m)/n + 1/n
> m2 + (2m)/n + 1/n2 = (m + 1/n)2

So there is an n with m2 < (m + 1/n)2 < N . But that is not possible,


because it means (m + 1/n) S, while m is an upper bound of the set S.
Second, what if m2 > N ? Then (m2 N ) > 0 and there is an integer
n > (2m)/(m2 N ). We can rearrange this to get

N < m2 (2m)/n
< m2 (2m)/n + 1/n2 = (m 1/n)2

5
...so we have an n with N < (m 1/n)2 < m2 . But this would make
(m 1/n) an upper bound for S, contradicting the fact that m is the least
such upper bound.
The completeness axiom says that this number m exists, that there is a
real number that is the least upper bound of the set S, and its square cant
be greater than N or less than N; therefore there exists a real number whose
square must equal N.

This is an example of what we get from the completeness axioms: there


are no gaps in the real number line. There are no positive numbers that
dont have square roots, for example. That means that there is at least one
irrational real number; we can show that there is no rational number whose
square is 2, but we just proved that there must be a real number whose
square is 2. This necessarily follows from the 16 axioms.
Here are some other facts pertaining to the fabric of the reals.

Theorem 2.13. Between any two real numbers, there is a rational number.

Proof. Choose reals x < y. The Archimedean principle gives us an N with


1
0 < yx < N . We have that x < x + 1/N < y, but (x + 1/N ) is not
necessarily rational.
However, we can also find an n with n N x < n + 1, so

n + 1 N x + 1 < N x + N (y x) = N y
n+1
. This gives us N x < n + 1 < N y, and so x < N < y.

Theorem 2.14. Between any two rational numbers, there is an irrational


number.

Proof. Suppose we have rationals (a/b) < (c/d), where a, c Z and b, d N.


Then we have integers ad < cb. The square
root of 2 exists, is not rational,
and is between 1 and 2, so ad < ad + 1/ 2 < cb. Dividing gives us a/b <
ad+1/ 2
< c/d.
bd
That middle number cant be rational: if 2 cant be expressed as a
n+1/ 2
fraction, neither can its reciprocal, and neither can m for integers m
and n.

Theorem 2.15. Between any two real numbers, there is an irrational num-
ber.

6
Proof. Choose reals x < y. We already can show there exists a z with
x < z < y. The previous theorem gives us rational numbers R and S with
x < R < z < S < y. There is an irrational number between them.

Corollary 2.16. Any interval (a, b), no matter how small, contains at least
one rational number, and at least one irrational number.
We actually have a word for this. We say that a set D is dense in the
reals if a member of D can be found in any real interval.

3 Facts about integers


We mentioned integers and natural numbers, so we should probably provide
some details about them.
Definition 3.1. The set of natural numbers N R is the smallest set
containing 1, which is closed under addition.
By smallest set we mean the intersection of all sets having this prop-
erty. That is, if you find any set of numbers S containing 1 which is closed
under addition, then N S.
The set N is the set {1, 1 + 1, 1 + 1 + 1, }. The set Z also includes 0,
and n for every n N.
Definition 3.2. The set of integers Z R is the smallest set containing 1
and -1, which is closed under addition.
Here are some facts about integers.
Theorem 3.3. If a set of integers S has a least upper bound n = sup[S],
n S.
That is to say, if a set of integers S has a supremum n, what it really has
is a greatest element n. Integers are so well-behaved and so spaced apart
that we dont have anything like an open interval of integers.

Proof. Let S be a set of integers and let m = sup[S]. First lets show that
m has to be an integer: there is an integer n with n m < n + 1 by
the Archimedean principle. There is no integer between n and m because
n + 1 > m. Hence any integer m is also n. This makes n an upper
bound of S, but m is the least upper bound, so n = m.
Next, lets show that m, the integer, must be in S. If m
/ S then n < m
for every n S; this means that n m 1 for every n S, and m 1 is
an upper bound, contradicting the fact that m is the least such.

7
This leads to a fact we call the well-ordering principle:

Corollary 3.4. Any nonempty set of natural numbers has a smallest ele-
ment.

4 Exercises
Prove the following from axioms:

1. Assuming that R contains more than one element, 0 6= 1. That


is, the additive and multiplicative identities can not be the same
element.
2. If x < y then y < x.
3. 0 < 1.
4. If a > 0 then a1 > 0
5. If 0 < x < y then 0 < (1/y) < (1/x).
6. Any nonempty set of numbers with a lower bound has a greatest
lower bound. This is called an infimum, and the greatest lower
bound is denoted inf[S].

We have six order axioms, but we could actually do the same with less.
Suppose we have a set P R that satisfies the following two axioms:

1. P is closed under addition and multiplication


2. For every x R, either x = 0, x P , or x P .

Suppose we furthermore define x < y to mean that (y x) P .


Show that all six order axioms follow from these two facts.
We could therefore define the reals using only 14 axioms if we wanted
two: nine axioms defining + and , two axioms defining the set P of
positive numbers which implies the order axioms, and the complete-
ness axiom.

Identify two axioms in our list of sixteen axioms that can be crossed
out without having any effect on what we can prove.

Show that a set with a single element {a} satisfies all sixteen axioms
if we define the + and and < operator correctly.

S-ar putea să vă placă și