Math 414 Professor Lieberman February 24, 2003 EXAM #1 SOLUTIONS 1. the statement P (n) is “n5 − n is divisible by 5.” To check P (1), we compute 15 − 1 = 0, which is divisible by 5 because 0 = 5 · 0. Now, we suppose that P (k) is true for some k ∈ N. Then (k + 1)5 − (k + 1) = k 5 − k + 5(k 4 + 2k 3 + k 2 ) by algebra, so (k + 1)5 − (k + 1) is also divisible by 5. Therefore P (k + 1) is also true. Now that we have shown that P (1) is true and that P (k) implies P (k + 1), mathematical induction implies that P (n) is true for any n ∈ N. 2. Given ε > 0, we want to find a natural number n∗ so that √ n c − 1 < ε √ √ if n ≥ n∗ . Since c > 1, we always have n c − 1 > −ε, so we only have to arrange n c − 1 < ε, which is the same as (1 + ε)n > c. By the Archimedean axiom, there is an integer n∗ such that n∗ ε > c − 1, and then n ≥ n∗ implies that 1 + nε > c. From Bernoulli’s inequality, we see that (1 + ε)n ≥ 1 + nε > c. 3. A homework problem says that, if f and g are injections, then g ◦ f is also an injection, and, if f and g are onto, then g ◦ f is onto. Therefore g ◦ f is a bijection. To see that (g ◦ f )−1 = f −1 ◦ g −1 , we note that if c ∈ C with b = g −1 (c) and a = f −1 (b), then f −1 ◦ g −1 (c) = f −1 (b) = a. But also g ◦ f (a) = c, so (g ◦ f )−1 (c) = a and therefore (g ◦ f )−1 = f −1 ◦ g −1 . 4. There are several different ways to do this. Method 1: Use the limit theorems: 2 3 2n + 3 2 + n3 lim 3 = lim n = 0. n→∞ n + 1 n→∞ 1 + 13 n Method 2: Use the definition, along with knowing that the limit is zero. Given ε > 0, take n∗ (by the Archimedean principle) so that n∗ > (3/ε)1/2 . If n ≥ n∗ , then 2n + 3 2n + 3 3n + 3 3 3 3 n3 + 1 − 0 = n3 + 1 ≤ n3 + 1 = n2 + n + 1 ≤ n2 ≤ (n∗ )2 < ε. Method 3: Note first that (2n + 3)/(n3 + 1) ≥ 0 and that 2(n + 1) + 3 2n + 5 = 3 . (n + 1)3 + 1 n + 3n2 + 3n + 2 By cross-multiplying, we have (2n + 5)(n3 + 1) = 2n4 + 5n3 + 2n + 5 < 2n4 + 9n3 + 15n2 + 13n + 2 = (2n + 3)(n3 + 3n2 + 3n + 2), so 2(n + 1) + 3 2n + 3 < 3 . (n + 1)3 + 1 n +1 1 2 Hence we have a decreasing sequence, which is bounded below and therefore it converges. Method 4: We have 2n + 3 3 0≤ 3 ≤ 2. n +1 n (See the algebra in Method 2 to verify this last inequality.) But limn→∞ 0 = 0 and limn→∞ 3/(n2 ) = 0, so 2n + 3 =0 lim n→∞ n3 + 1 by the Squeeze Theorem. √ 5. Method 1: First, a1 = 2 ≤ 2 and, if an ≤ 2, then √ √ an+1 = 2an ≤ 2 · 2 = 2, so by induction an ≤ 2 for all n, so the sequence is bounded above. Also an ≥ 0 by the formula, so √ √ an+1 = 2an ≥ an an = an , so the √ sequence is also increasing. It therefore √ has a limit A. To evaluate A, we note that A = √2A, so A = 0 or A = 2. Since a1 = 2, and (an ) is an increasing sequence, we have A ≥ 2, so A = 2. Method 2: The basic idea is to see that an = 2b(n) with b(n) 2−n . We proceed by √= 1 −1/2 induction, or look at the pattern from the first few a’s: a1 = 2 = 2 and b(1) = 2−1 = 1/2. If we know that ak = 2b(k) for some k ∈ N, then √ ak+1 = 2ak = 21/2 2b(k)/2 = 2(b(k)+1)/2 . But (b(k) + 1)/2 = (1 − 2−k + 1)/2 = 1 − 2−(k+1) , so the formula for an is true by induction. (Any verification of this formula by examining the first few a’s was worth few points.) Now we note that b(n) → 1 as n → ∞, so lim an = 2limn→∞ b(n) = 2. n→∞ 6. The set of accumulation points is [−2, 4]. It’s easy to show that points outside of this interval are not accumulation points because if s0 is outside of this interval, then the interval (s0 − ε, s0 + ε) has no points in S if ε = min{|s0 − (−2)|, |s0 − 4|}. For any point s0 in the open interval (−2, 4), there is a number δ(s0 ) such that x ∈ S for all x with 0 < |x − s0 | < δ. (To be technically correct, you have to check the cases s0 = 1 and s0 6= 1 separately since δ = 3 if s0 = 1 and δ = min{|s0 − (−2)|, |s0 − 1|, |s0 − 4|} if s0 6= 1.) So the number x = s0 +min{δ, ε/2} is in S with 0 < |x−s0 | < ε, which means s0 is an accumulation point. For s0 = −2 and s0 = 4, a similar argument works with x = −2 + min{3, ε/2} if s0 = −2 and x = 4 − min{3, ε/2} if s0 = 4. As mentioned in class, your argument did not need anywhere near this much detail.