Anda di halaman 1dari 11

Theorem 8.6.8. F+ is closed under addition and multiplication.

PROOF: this follows easily from the fact that I+ is closed under addition
and multiplication. We leave the details to be done as an excercise.

Problems

1. In F, let us define the relation R by saying x R y is and only if x = y =


0 or if x and y have the same sign. Is R an equivalence relation in F? If so,
what are its equivalence classes?
2. Is F+ closed under
(a) subtraction (b)division?

Explain how you know.


1. In F-- closed under
(a) addition (b)subtraction

(c) multiplication (d)division?

Explain how you know.


4. Prove Theorem 8.6.8.
5. Let Q = F+∪{0} and S = F--∪ {0}. Is the system (Q, +)
isomorphic tp he system (S, +)? Explain.
6. If x, y, and z are positif rational number, we say that x is y per
cent(or y%) of z if and only if xz = y100. In other words, x is y per
cent of z when x = z(y100). To find y per cent of z, we multiply z by
y100. For example, 5 is 1 per cent of 500 since 5 = 5001100. What
number is
(a) 3 per cent of 960
(b)7 pre cent of 14,000
(c) 6 per cent of 5400
(d) 12 per cent of 17,5000
7. (a) 3200 is 6 per cent of what number?
(b) 560 is 8 per cent of what number?
(c) 43,000 is 3 per cent of what number?
8. (a) 75 is what per cent of 75,000?
(b) 3600 is what per cent of 18,000?
(c) 596 is what per cent of 298?

8.7 Inequalities

One-third of a pie is larger than one-fourth of it. We would get less


money if we inherited one-fifth of our rich uncle’s estate than if we
inherited two-ninths of it. If seven-eights of the people in a town support a
certain candidate for mayor, he has more support than if only three-
quarters of them favor him.

Clearly, there are many, many practical situations where we need to


compare the relative sizes of two rational numbers.

As a direct analogue of the corresponding definition in I we have:

When x and y rational number we say that x ≥ y if there is an


nonnegative relation number z such that x = y + z.

This is the third time we have used “≥” to name a relation. We might
have labeled this new relation by means of a symbol such as “≥F”. But
there is no to do this. Sine the system integral rational numbers with
addition and multiplication is isomorphic to the system (I, +, ∙) it follows
that the relation ≥ as defined above in the set of integral rational numbers
is isomorphic to the relation ≥ we had in I. Thus, we are free to interpret
such a statement as 6 ≥-5 as either an assertion about integers or an
assertion about the rational numbers 61 and -51.

We define the relation>, ≤, and < in F in terms of ≥ as we have done


before. We also carry over to F our definition for “betwen”, “ maximum”,
and “minimum”.

In terms of the number line, x > y in F if and only if the rational point x
is at least as far to the right as the rational point y. Since F is closed under
subtraction we have the result similar to one in I that when x and y are
rational numbers.

x ≥ y if and only if x – y is non negative.

We may buil on this result to get a rather simple produrce for


comparing the sizes of two rational numbers. Suppose a, b, c, and d are
integers and b and d are positive. We know that ab≥cd if and only if ab- cd
is nonnegative. But ab- cd= adbd- cbbd= ad-cbbd. When b and d are
posotove integers, so is bd and, hence, this last fraction names a
nonnegative rational number if and only if ad-cb is non negative (in I). And
ad-cb is nonnegative if and only if ad≥cb. We have thus shown that when
a, b, c, and d are integers:

If b and d are posotive, then ab≥cd if and only if ad≥cb.

In a similar fashion we may prove that:

If b and d are posotive, then ab≥cd if and only if ad≥cb.


We now have a very convenient method for comparing two rational
numbers when we wrirte them as faction with positive denominators. The
procedure is summarized by the flow chart of figure 8.8.

As examples aof the use of the above we see that

13>14 since 4 ∙ 1 = 4 > 1 ∙ 3 = 3,

15<29 since 1 ∙ 9 = 9 < 2 ∙ 5 = 10,

Read “a”, “b”, “c”,


and “d”

Compute ad

Compute ad

Is ad = bc?
Ye
Compute ab=
s
cd

N
o
Is ad > bc?
Ye
Compute
s
ab>cd

N
o
Compute
ab<cd
Figure 8.8 flow chart for comparing aband cd when b and d are
positive.

78>34 since 7 ∙ 4 = 28 > 3 ∙ 4 = 12,

-75<35 since -7 ∙ 5 = -35 < 3 ∙ 5 = 15,

-59>-173 since -5 ∙3= -15> -17 ∙9= -153,

and
-10731000<- since -107,300< -107,000
107100

One immediate result of the above method is:

If a, b, and c are integer and c is positive, then ac≥bc if and only if a


≥ b.

Also

If a, b, and c are integer and c is positive, then ac>bc if and only if a>
b.

These results tie in nicely with our geometric notions about rational
numbers. When c is a positive integer, 1c is the length of any one of the
segments obtained when we break up the segment from 0 to 1 into c
pieces of equal length. And ac=a1c represents the point reached from the
origin by laying out a consecutive segments of length 1c in the proper
direction – to the right if a is positive and to the left a is negative. bc may
be interpreted similarly. We certainly should have the point ac at least as
far to the right as the point bc when and only when a ≥b.

Many of the properties that we discovered for ≥ in N carry over to F. In


fact, the assertions numbered from 2 to 12 on page 156 are all true if
wake take it that x, y, z, and w are rational numbers.

And some additional properties of ≥ in I carry over to F. In particular, we


have that when x, y, and z are rational numbers

(a) x >0 if and only if x is positive


(b) x <0 if and only if x is negative
(c) if z>0, thenx ≥y if and only if xz ≥yz
(d) if z<0, then x ≥y if and only if xz ≤yz
(e) if z ≥0, thenx-y ≤z if and only if y-z ≤x ≤y+z
(f) x+y ≤ x+ y.

Since all nonzero elements in F have reciprocals, results (c) and (d)
above are especially useful. If, for instance, we wish to find the maximum
and minimum for the set x-6 ≤7-3x ≤4, we may do so as follows.

x-6 ≤7-3x ≤4= x-6-7 ≤ -3x ≤4-7= x-13 ≤ -3x ≤ -3= x-13-13 ≥ -13-3x ≥ -13-
3= x133 ≥x ≥1= x1 ≤x ≤ 133.

We see immediately that the maximum of this set is 133 and the
minimum is 1. Note well the reseval of the inequalities at the third step
above.
As another example, we have x2x-1 ≤6= x1-6 ≤2x ≤1+6= x12-5 ≤ 122x ≤
127= x-52 ≤x ≤ 72 so that this set has a maximum of 72and a minimum of-
52.

Despite the fact that some of the basic facts about inequalities in I carry
over to F, the order situation in F is strikingly different from that in I. it
isfalse in F that when x<y we must have x+1 ≤y. For example, when x= 12
and y=1, we have x<y and yet x+1= 32 > 1.

It also false in F that there is no number between 0 and 1. For example,


we have 0< 12 < 1.

In fact, we havethe following theorem which points up a very important


difference between I and F.

If x and y are rational numbersand x<y, then there is a rational number


z such that x<z<y.

This is easy to prove because we can show that x +y2 is between x and
y. This may be done as follows. From x<y it follows that x+x<y+x also
x+y<y+y. Then 2x<x+y<2x. And multiplying both sides of these two
inequalities by the positive number 12 we obtain x< x+y2<y.

Whenever we have two distinct rational numbers x and y we must have


x<y or y<x. So it follows from what we just proved that given any two
distinct rational numbers there is a rational number that is betwen them.
In terms of this number line tells us that between any two rational points
there is another rational point.

A truly remarkable fact follows from these considerations. Between any


two distinct rational numbers there any infinitely many rational numbers.

For suppose that x and y are in F and x ≠ y. there is no loss in generality


if we suppose x<y. We know there is an x1 such that x < x1 < y. but then
there must be an x2 such that x< x2 < x1. Then there must be an x3 such
that x < x3 < x2. And so forth without end. The distinct numbers x1, x2, x3,
… are all between x and y.

In N and I we found that bounded sets were always finite. We see


thatthis is not true in F. we shall dig more deeply into this result and
related matters in the next section.

Definition 8.7.1. when x and y are in F, wesay that x ≥y if there is


an nonnegative rational number z such that x=y+z. We define >, ≤,
<, “between”, max(A), and min(A) as before.
Theorem 8.7.1. let Q be the collection of integral rational numbers. The
relation ≥ of definition 8.7.1 in Q is isomorphic to the relation ≥ in I.

Theorem 8.7.2. Suppose x ∈F and y ∈F. x ≥ y if and only if x-y is


nonnegative.

PROOF: see proof of Theorem 7.8.4.

Theorem 8.7.3 if a, b, c, and d are integers and b and d are positive,


then ab ≥ cd when and only when ad ≥bc.

Theorem 8.7.4. When a, b, and c are integers and c is positive, ac ≥ bc if


and only if a ≥b.

Theorem 8.7.5. Suppose x, y, z, and w are in F. Then

(a) x≥x
(b) x≥y and y≥z implies x≥z
(c) x>y and y>z implies x>z
(d) x≥y or y ≥x
(e) x≥y and y≥x implies x=y
(f) x>y or y>x or x=y and only one of these may hold
(g) x+1>x
(h) x≥y if and only if x+z≥y+z
(i) x>y if and only if x+z>y+z
(j) x≥y and z≥w implies z+z≥y+w
(k) x≥y and z>w implies x+z>y+w.

PROOF: the proofs are similar to those for similar statemens in I.


They will be left as exercise.

Theorem 8.7.6. Suppose x, y, and z are in F. Then

(a) x>0 if and only if x is positive


(b) x<o if and only if x is negative
(c) if z>0, then x≥y if and only if xz ≥yz
(d) If z<0, then x≥y if and only if xz≤yz
(e) If z≥0, then x-y≤z if and only if y-z≤x≤y+z
(f) x+y≤x+y.

Theorem 8.7.7. Between any two distinct rational numbers there is


another ratioanal number.

Theorem 8.7.8. Between any two distinct rational numbers there


are infinitely many distinct rational numbers.

PROBLEMS
1. Which of the following are true?
(a) 3719≥4919 (b) -3719≥-4919

(c) -54≥-23 (d) -593≥-753

(e) 517≥-734 (f) 5-7≥-23

(g) 10057≥9956 (h) -7581≥-2527

(i) -436≥42-5 (j) 10211000≥102100

(k) -25741000≥-19,65010,000.
1. Find the maximum and minimum (if any) for each of the following
sets:
(a) x2x≤1 (b) x-34≤x≤75

(c) x-2x≤1 (d) x-34≤x+1≤0

(e) xx≤117 (f) x-2≤3+4x≤7

(g) x-4≤5-6≤3.
(a) We have 520 items we wish to sell to gross at least
$1780. What selling prices are feasible ones for us to set?

(b)We purchase 150 shares of a certain stock at $57 per


share. The market price of the stock has fallen very low.
We wish to sell all the shares when the prices gets back
up to where we shall lose no more than a total of $1500
on the shares. Which market prices will allow us to do
this?

(c) In a certain town of 7500 people, it is known that at least


13 are below voting age. How many people in the town
could be of voting age?

(d) Would you rather get as a prize 1349 of $10,000 or 14 of


$9,750?

1. Suppose a, b, and c are positive integers. Prove that ab≥ac if and


only if c≥b.
2. Prove Theorem 8.7.5.
3. Prove Theorem 8.7.6.
4. What is the isomorphism that establishes Theorem 8.7.1?

7.7 Least upper bounds

Let A be the set of rational numbers between 0 and 1. That is, A=xx∈F
and 0<x<1. 12, 13, 14, 15, and 16 are all in A. in fact if n is any positive
integer than 1n is in A. 23, 34, 45, 56, and 67 are all in A. in fact, nn+1 is in
A whenever n is a positive number.

The rational points on the number line which are in A all fall between
the poin o and the point 1. A is certainly a bounded set and yet A is
infinite.

We shall now put this notion of a set being bounded a bit being
precisely and less geometrically than we have done before. We shall say
that a rational number a is an upper bound for a nonempty set S of
rationals if x≤a for all x in S. A set which has an upper bound is said to be
bounded above. We shall say that a number b is a lower bound for S if b≤x
for all x in S. A set which has a lower bound is said to be bounded below. If
a set is bounded above and also bounded below than we say that it is
simply bounded.

For the set A=x0<x<1, 1, 5, 7, 9, and 106 are all upper bounds. -1000, -4,
-2, and 0 are all lower bounds for A. A is a bounded set.

The set B={xx=1n where n∈I+} is bounded above by 2 and below bi -2.
Hence B is a bounded set.

The set C={xx=nn+1 where n∈I+} has 10 as an upper bound and -50 as
a lower bound. Hence, C is bounded.

Notice that when we once find an upper bound for a set, than any larger
number will also be an upper bound for the same set. Also if we find any
one lower bound for a set, any smaller number will also be a lower bound.

What is the maximum of the above set B=1, 12, 13, 14, …? Evidently, 1
is the largest element in B sothat max(B)=1. What is the minimum of B?
There is none! We see that 1n+1<1n for any positive integer n. and any
element in B must be in the form 1n for some positive integer n. So, given
any element in B, there is another element in B which is smaller. Hence, B
can have no minimum.

This situation is in marked contrast to that in N where every nonempty


set of natural numbers has a minimum. And it is also quite different from
what we know about I. any nonempty bounded set of integers has both a
maximum and a minimum. Yet B is a bounded set of rationals with no
minimum.

The bounded setC=12, 23, 34, 45, …, as defined above has a minimum of
12 and yet has no minimum.

The bounded set A=x0<x<1 has neither a maximum nor a minimum. For
let x be any element of A. Then 0<x and there is a rational number y such
that 0<y<x. Then y∈A. Since y∈A, x cannot be the minimum of A. Thus, A
does not have a minimum. A similar argument may be used to show that A
has no maximum.

Notice that the argument showing that A has no minimum established


that there is no such thing as the smallest positive rational number. Given
any positive rational number, we can find a smaller one.

in the other words, it does not make sense to speak of the next rational
point to the right of the origin on the number line. There is no such
animal. Given any rational point to the right of the origin, we can find
another point to the right of 0 that is closer to 0.

Putting this another way, we claim that there is no such thing as the
rational point which is closest to the origin, and yet to the right of the
origin. Given any rational point to the right of the origin, we cand find
another point to the right of 0 that is closer to 0.

Although the set A=x0<x<1 has no minimum, there is some special


connection between the number 1 and this set. 1 is an upper bound for A.
But, more than this , no number less than 1 can be an upper bound for A.
That is, A is the least upper bound for A. or in other words, 1=minxx is an
upper bound for A. This notion is useful more generally.

If S is a set of rational numbers we say that a number a is the least


upper bound for S if a is the minimum of the collection of all upper bound
forS.

We state more briefly that a is the least upper bound for S by writing
“a=lub(S)”.

For the set A above we have lub(A)=1.

For the set B above we have lub(B)=1. Notice that when a set S a
maximum element, then maxS= lub(S).

For the set C above we have lub(C)=1.

How do we tell if a number is a lub for a set? First of all, a set must be
bounded above if it is to have a lub. For if a set is not bounded above it
has no upper bounds at all, let alone a least upper bound.

Suppose a is an upper bound for a set S. how can we show that is it the
least upper bound? We can prove that is by establishing any one of the
following:

I. Whenever b is an upper bound for S we have a≤b


II. If b<a, then b is not an upper bound for S
III. If b<a, then there is some element x is S such that b<x≤a
IV. Given any positive rational number d, no matter how small,
there is a member x of S such that a-d<x≤a.

For example, we may use the form (IV) to show that lub(C)=1 for the set
C above. But, before we do, we must mention a basic property of the
rationals called the Archimedean property:

If x and y are positive rationals, then there is a positive integer n such


that nx>y.

Inother words, the positive integral multiples of a fixed positive rational


get arbitrarily large. This ties in with the geometric idea that if we have a
segment with nonzero rational length , we can build up a segment that is
as long as we please by laying enough copies of our given segment end to
end.

Accepting this fact, weshall show that the lub of C=12, 23, 34, 45, … is 1.
We first see that when n is any positive integer we have n<n+1 so that
nn+1<1. Hence 1 is an upper bound for C. let d be any positive rational. By
the archimedian property of F, there isa positive integer n such that nd>1.
Then d>1n. Hence (-1)d<(-1)1n and 1-d<1-1n=nn-1n=n-1n<1. If n=1 then
d>1 and 1-d<0 so that every member of C is between 1-d and 1. If n≠1,
then n-1n is a member of S which falls between 1-d and 1. We conclude by
(IV) that lubC=1.

If we look at things from the left instead of from the right on the number
line, we are led to make the following definitions for a set S of rational
numbers. A number b is said to be the greatest lower bound (glb) for S if b
is the maximum of collection of all lower bounds for S.

For the sets A, B, and C above we have glb(A)=0, glb(B)=0, and glb(C)=
12. Note that when S has a minimum we have minS= glb(S). suppose b is a
lower bound for S. then b=glb(S) if any of the following holds:

I. If c is a lower bound for S, then c≤b


II. If c>b, then c is not a lower bound for S
III. If c>b, then there is a member x of S such that b≤x<c
IV. Given any positive rational d, no matter how small, there is a
member x of S such that b≤x<b+d

It is quite easy to use the Archimedean property and condition (IV) to


prove that lub(B)=0 for the set B above.
A veru deep question can now be asked about F. is it the case every
nonempty set that has an upper bound in F also has a least upper bound
in F? it turns out the answer is negative. Of this, more later.

Definition 8.8.1. A number a is said to be an upper bound for


nonempty set S of rationals if x≤a for all x in S. b is said to be upper
bound for S if b≤x for all x in S.

Definition 8.8.2. If a nonempty set S has an upper bound it is said


to be bounded above. If S has s lower bound it is said to be bounded
below. If S is bounded above and bounded below, it is said to be
bounded.

Definition 8.8.3. lub(S) is the minimum of the collection of all upper


bounds for S. glb(S) is the maximum of the set of all lower bounds
for S.

Theorem (a) If a set S has maximum, then


8.8.1. lub(S)=max(S).

(b) If a set S has a minimum, then


glb(S)=min(S).

Theorem 8.8.2. (the Archimedean property) if x and y are positive


rationals then there is an integer n such that nx>y.

PROOF: Let x=ab and y=cd where a, b, c, and d are positive integers. If
n is a positive integer we have nx=nab=nad if and only if nad>bc.
But by Theorem 6.16.2 we know that whatever the values of a, b,
c, and d, there does exist an n which makes this last equality
hold.

Theorem 8.8.3. Suppose that a is an upperbound for nonempty set S.


then a=lub(S) if and only if one of the following holds:

I. Whenever b is an upper bound for S we have a≤b


II. If b<a, then b is not an upper bound for S
III. If b<a, then there is an x in S such that b<x≤a
IV. If d is any positive number, then there is an x in S such that a-
d<x≤a.

PROOF: Part (I) is simply a restatement ofdefinition 8.8.3. (II) is


simply a restatement of the meaning of (I). (III) is a restatement of
(II). (IV) is a

Anda mungkin juga menyukai