MATH 104, SUMMER 2006, HOMEWORK 2 SOLUTION

BENJAMIN JOHNSON Due July 5

Assignment: Section 4: 4.1(d)(e)(h)(t)(u)(v); 4.7, 4.12 B1: Prove that an ordered eld F is Archimedean if and only if satises (a F)(n N)(n > a) Section 5: 5.1, 5.2, 5.6 Section 6: 6.1, 6.3, 6.4 Section 7: 7.3(f)(g)(j)(t), 7.4 Section 4 4.1 For each set bounded above, list 3 upper bounds for the set. Otherwise, write NBA. (d) {, e}: , 4, 5 : n N }: 1, 2, 3 (e) { 1 n (h) n=1 [2n, 2n + 1]: NBA (t) { x R : x3 < 8}: 2, 3, 4 (u) { x2 : x R}: NBA (v) {cos( n3 ) : n N}: 1, 2, 3 4.7 Let S and T be nonempty bounded subsets of R. (a) Prove that if S T then inf T inf S sup S sup T . Proof. Note that S , T bounded and non-empty implies inf T , inf S , sup S , sup T all exist. Let x S . We prove the three inequalities separately. Since S T , we have x T , and since inf T is a lower bound for T , we have inf T x. Since x was arbitrary in S , the above argument shows inf T is a lower bound for S . Since inf S is the greatest lower bound for S , we have inf T inf S . Since inf S is a lower bound for S , we have inf S x. Since sup S is an upper bound for S we have x sup S . So we have inf S x sup S . By transitivity inf S sup S . Since S T , we have x T , and since sup T is an upper bound for T , we have x sup T . Since x was arbitrary in S , the above argument shows sup T is an upper bound for S . Since sup S is the least upper bound for S , we have sup S sup T . (b) Prove that sup(S T ) = max{sup S , sup T }. Proof. Let x S T . Either x S or x T (or both). If x S , then x sup S max{sup S , sup T }. Similarly if x T , then x sup T max{sup S , sup T }. We see that in any case we have x max{sup S , sup T }. This shows max{sup S , sup T } is an upper bound for S T. Let B be any upper bound for S T . Then B is an upper bound for S , and since sup S is the least upper bound for S , we have sup S B. Similarly, B is an upper bound for T , and since
Date: July 5, 2006.
1

BENJAMIN JOHNSON

sup T is the least upper bound for T , we have sup T B. Thus we have max{sup S , sup T } B. This shows max{sup S , sup T } B for all upper bounds B for S T . Weve shown max{sup S , sup T } satises the denition for supremum of the set S T . Since suprema are unique, sup(S T ) = max{sup S , sup T }. 4.12 Let I be the set of real numbers that are not rational. Prove that if a < b, then there exists x I such that a < x < b. Proof. First we show that if c is rational, then c + 2 is irrational. Proof by contradiction. Suppose c is rational. If c + 2 is also rational, then since rational numbers are closed under subtraction (c + 2) c = 2 is rational. But weve shown 2 is irrational, a contradiction. Continuing with the main assertion, let a, b R with a < b. By Denseness of Q in R, there is a rational number c with a 2 < c < b 2. Then we have a < c + 2 < b. And c + 2 is irrational by what we proved above. B1 Prove that an ordered eld F is Archimedean if and only if satises (a F)(n N)(n > a). Proof. To say that F is Archimedean (according to the books denition) means that a, b F : a, b > 0)(n N )(na > b). Call this (1), and call the other property listed above (2). (1) (2) Assume (1). Let a F. If a 0, then take n = 1, and we have n > a. If a > 0, then applying (1) substituting a for b and 1 for a yields (n N )(n 1 > a). This proves (a F)(n N)(n > a), i.e. (2). (2) (1) Assume (2). Let a, b F Assume a, b > 0. Then b F. Using (2), substituting b for a, we have a a b (n N)(n > b ). Note that since > 0 such an n must actually be in N . Also, for a , b > 0, n > b a a a is equivalent to n a > b. Thus we have proven (a, b F : a, b > 0)(n N )(na > b), i.e. (1). Section 5 5.1 Write the following sets in interval notation: (a) { x R : x < 0} = (, 0) (b) { x R : x3 8} = (, 2] (c) { x2 : x R} = [0, ) (d) { x R : x2 < 8} = ( 8, 8) 5.2 Give the inmum and supremum of each set listed in Exercise 5.1. (a) inf = , sup = 0 (b) inf = , sup = 2 (c) inf = 0, sup = (d) inf = 8, sup = 8 5.6 Let S and T be nonempty subsets of R such that S T . Prove that inf T inf S sup S sup T . Proof. Regardless of whether the values are allowed, inf(S ) is the unique extended number satisfying (1) inf(S ) x for every x S , and (2) for every lower bound b of S (where b could also be ), we have b inf(A). Similarly, sup(S ) is the unique extended real number satisfying (1) x sup(S ) for every x S , and (2) for every upper bound B of S , (where B could also be +), we have inf(S ) B.

MATH 104, SUMMER 2006, HOMEWORK 2 SOLUTION

With this in mind, the proofs for each of the three inequalities from Exercise 4.7(a) (where we did not allow ) go through without modication. (Copied below). Let x S Since S T , we have x T , and since inf T is a lower bound for T , we have inf T x. Since x was arbitrary in S , the above argument shows inf T is a lower bound for S . Since inf S is the greatest lower bound for S , we have inf T inf S . Since inf S is a lower bound for S , we have inf S x. Since sup S is an upper bound for S we have x sup S . So we have inf S x sup S . By transitivity inf S sup S . Since S T , we have x T , and since sup T is an upper bound for T , we have x sup T . Since x was arbitrary in S , the above argument shows sup T is an upper bound for S . Since sup S is the least upper bound for S , we have sup S sup T .

Section 6 6.1 Consider s, t Q. Show that (a) s t if and only if s t Proof. Assume s t. Let x s . Then x < s. So x < t. So x t . This shows s t . Assume s t . Proceed by contradiction. Assume t < s. Then there is some rational number c with t < c < s. So c s but c t This contradicts s t (b) s = t if and only if s = t Proof. Assume s = t. Then {r Q : r < s} = {r Q : r < t}. So s = t . Assume s = t . Proceed by contradiction. If s t, then one of s < t or t < s holds. Without loss of generality, assume that s < t holds. Then there is a rational number c with s < c < t. We have c t but c s . This contradicts s = t . Conclude s = t. (c) ( s + t) = s + t Proof. Let x ( s + t) . Then x Q and x < s + t. By algebra, x s < t. By density, there is a rational number c with x s < c < t. By algebra x c < s. Because x and c are rational, so is x c. We have c t , x c s , and x = c + ( x c). So x s + t . This proves ( s + t) s + t Let x s + t . Then there are rational numbers a s and b t with x = a + b. Since a, b are rational so is x. Since a < s and b < t, we have x < s + t. So x ( s + t) This proves s + t ( s + t) . Hence ( s + t) = s + t . 6.3 (a) Show that + 0 = for all Dedekind cuts . Proof. Let be any Dedekind cut. Let x + 0 . Then there exist r and s 0 with x = r + s. Since s < 0, we have x < r. since is closed under <, we have x . This argument shows + 0 . Let x . Since has no least rational number, there must be some rational number c with x < c. Then x c is rational and less than 0, so x c 0 . We have x = c + ( x c), so x + 0 . This argument shows + 0 . Hence + 0 =

BENJAMIN JOHNSON

(b) How would you dene ? = {r Q : r } 6.4 Let and be Dedekind cuts and dene the product: = {r1 r2 : r1 and r2 }. (a) Calculate some products of Dedekind cuts using the Dedekind cuts 0 , 1 , and (1) . 0 0 = { x Q : x > 0} 0 1 = Q 0 (1) = { x Q : x > 0} 1 1 = Q 1 (1) = Q (1) (1) = { x Q : x > 1}

(b) Discuss why the denition of product is totally unsatisfactory for dening multiplication in R. Aside from the less consequential shortcomings of not handling 0 and 1 correctly with respect to several of the ordered eld axioms, the product function dened above does not even output a Dedekind cut when given Dedekind cuts as inputs. Always giving a Dedekind cut as an output would be the rst prerequisite for any binary function on the set of all Dedekind cuts. A product function of this type would normally be expected to be a binary function. Section 7 7.3 For each sequence below, determine whether it converges, and if it converges, give its limit. 1 (f) sn = 2 n : The limit is 1. (g) yn = n!: Does not converge. n3 +8n (j) 7 : The limit is 7 . 2 2n3 31 6n+4 (t) 9n2 +7 : The limit is 0. 7.4 Give examples of (a) a sequence xn of irrational numbers having a limit that is a rational number. xn =
2 n

denes a sequence of irrational numbers with rational limit 0.

(b) a sequence rn of rational numbers having a limit that is an irrational number.