Math 414 Professor Lieberman February 18, 2003 HOMEWORK #5 SOLUTIONS Section 2.2 8. (a) We will show that A < B + ε for every ε > 0. So let ε > 0 be given. Then there are numbers n1 and n2 such that |an − A| < ε/2 if n ≥ n1 and |bn − B| < ε/2 if n ≥ n2 . We now choose n = max{n1 , n2 }. Then ε ε A − B = (A − an ) + (bn − B) + (an − bn ) < + + 0 = ε, 2 2 so A < B + ε as claimed. (b) There are lots of examples. I like an = 0 and b + n = 1/n but there are many others. 11. (d) The limit exists and it’s zero. To see this, let m ∈ N be the number from Theorem 1.6.8 such that m > |r|, and let ε > 0 be given. Then Theorem 1.6.8 gives a number n∗ such that 1 mm ∗ n > max , m+1 . ε m! If n ≥ n∗ , then n n r − 0 = |r| ≤ m m . . . m m . . . m . n! n! n n−1 mm−1 1 But m/n ≤ m/n∗ so m/s ≤ 1 if s = m, . . . , 1, so n m r − 0 ≤ m m < ε. n! n∗ m! √ √ (i) Let bn = n n and cn = n 2n. Then bn ≤ an ≤ cn and we know that (bn ) converges to 1 by Exercise 15 of Section 2.1 while (cn ) converges to 1 by combining Exercises 14 and 15 of Section 2.1 with Theorem 2.2.1(b). It then follows from the Squeeze Theorem that (an ) converges to 1. 13. It is not true √ that the sequence√(bn ) must converge. For example, we could have an = 1/n and bn = n. Then an bn = 1/ n, so an bn → 0, but (bn ) diverges. Section 2.3 4. (a) Take bn = an . (b) Take bn = (−1)n an . 8. First, we rewrite an = np−q bn with bn = sp + sp−1 n−1 + · · · + a0 n−p . tq + tq−1 n−1 + · · · + t0 n−q From the limit theorems along with the equation sp tq 6= 0 (which is the same as sp 6= 0 and tq 6= 0), we have that (bn ) converges to sp /tq , which is not zero. It then follows from 1 2 Theorem 2.1.12, that bn is bounded, so Theorem 2.3.3(b) tells us that (an ) diverges to infinity. Section 2.4 1. One possibility is ( n an = n/2 if n is even if n is odd. 6. (a) We can write 1/2 1 + , 2 2n + 1 so (an ) is decreasing, and it is bounded below by 1/2. Therefore it converges by Exercise 5 of Section 2.4 (which is the “decreasing” version of Theorem 2.4.2). (e) Now we have to be a little trickier. Let’s look at an+1 /an : an = an+1 3n+1 1 + 32n 3 + 32n+1 = = an 1 + 32n+2 3n 1 + 32n+2 2n+1 2n+1 3 +3 2 · 32n+1 32n+2 ≤ = ≤ ≤ 1, 1 + 32n+2 1 + 32n+2 1 + 32n+2 so the sequence is decreasing. Since the terms are positive, it’s also bounded below, so it converges by Exercise 5. (f) Now we have an+1 (n + 1)! 1 · 3 · · · · · (2n − 1) n+1 = = ≤ 1, an n! 1 · 3 · · · · · (2n + 1) 2n + 1 so the sequence is decreasing. Again the terms are positive, so it’s bounded below and hence convergent. 7. (a) Now 1 > 0, (n + 1)2 so the sequence is increasing. Exercise 2(t) from Section 1.3 says that an ≤ 2 − 1/n, so the sequence is also bounded. Therefore it converges by Theorem 2.4.2. In addition an ≥ 1, so Theorem 2.2.1(f) says that 1 ≤ A ≤ limn→∞ (2 − 1/n). This last limit is 2. In addition, an ≥ 5/4 if n ≥ 2, so Theorem 2.2.1(f) also says that A ≥ 5/4. an+1 − an = Section 2.5 3. (a) an = (−1)n n/(n + 1) has exactly two accumulation points −1 and 1. (The problem didn’t ask you to prove that these are all the accumulation points, but here’s a proof that they are accumulation points anyway. First, given ε > 0, Theorem 1.6.8 says that there is an integer n∗ such that n∗ > 1/ε. Then 2n∗ 1 1 ∗ |a2n − 1| = ∗ − 1 = ∗ ≤ ∗ < ε, 2n + 1 2n + 1 n a2n∗ ∈ S, and a2n∗ 6= 1. Similarly |a2n∗ +1 − (−1)| < ε, a2n∗ +1 ∈ S, and a2n∗ +1 6= −1. Hence −1 and 1 are accumulation points. Showing that there aren’t any more accumulation points 3 is similar but a lot of work!) (b) Same as (a). (c) S = (0, 1) ∪ {2}. (d) S = {−1} ∪ (0, 1) ∪ {2}. 4. (a) If s0 is an accumulation point for S, then, for each positive integer n, there is a point an ∈ S with an 6= s0 and |an − s0 | < 1/n (this is just the point corresponding to ε = 1/n). To see that (an ) converges to s0 , let ε > 0 be given. Then Theorem 1.6.8 gives a natural number n∗ such that n∗ > 1/ε. If n ≥ n∗ , then 1 1 |an − s0 | < ≤ ∗ < ε, n n so an → s0 . Conversely, if there is a sequence (an ) in S with an 6= s0 for all n and an → s0 , then, given ε > 0, there is a natural number n∗ such that |an − s0 | < ε if n ≥ n∗ . It follows that an∗ ∈ S, an∗ 6= s0 and |an∗ − s0 | < ε. Hence s0 is an accumulation point of S. (b) It’s necessary because the set S = {2} has no accumulation points but the sequence an = 2 for all n is a sequence of points in S such that an → s0 .