Homework 5 Key 1. (3 points) Here’s the scratch work portion of the proof: We want |x2 + x − 6| < ϵ or equivalently (use x2 − x + 6 = (x + 3)(x − 2)) |x − 2| < ϵ/|x + 3|. Let’s restrict δ ≤ 1 so that |x − 2| < 1, or 1 < x < 3. This means that 4 < |x + 3| < 6, so from above we want |x − 2| < ϵ/6. We can then take δ = ϵ/6. The formal proof would be this: Given ϵ > 0, let δ = max(1, ϵ/6). If |x − 2| < δ we have |x − 2| < 1 so that 1 < x < 3 and hence 4 < |x + 3| < 6. Thus |x2 + x − 6| = |x + 3||x − 2| (ϵ) < 6 6 = ϵ so that limx→2 x2 + x = 6. 2. Many ways this can be done. For simplicity, let f (x) = sin(1/x). Let δ > 0 and consider the interval I = |x − 0| < δ, that is, −δ < x < δ. Note that if xk = 1/(2πk) for k ∈ Z then f (xk ) = 1, while if zk = 1/(2πk + π) we have f (zk ) = −1. And in each case, as k → ∞ we find xk , zk → 0. In short, we can find sequence xk , zk → 0 such that f (xk ) → 1 and f (zk ) → −1, so f (x) can have no limit as x → 0. 3. If f is bounded then for some constant M we have |f (x)| ≤ M for all x ∈ X. Since g limits to 0 at p, for any ϵ > 0 there is a δ > 0 so that |g(x)| < ϵ/M for |x − p| < δ. If |x − p| < δ we then have |f (x)g(x) − 0| = |f (x)||g(x)| < M (ϵ/M ) = ϵ which shows that limx→p f (x)g(x) = 0. 1 4. A one-liner: Since the set V = {y ∈ R; y ≤ 0} is closed and f is continuous, the inverse image f −1 (V ) = {x ∈ X : f (x) ∈ V } = {x ∈ X : f (x) ≤ 0} is closed, by Theorem 4.8. 5. (a) Let x be a fixed real number. We can choose a sequence rn of rational numbers that converge to x. Since f is continuous, f (rn ) converges to f (x). But f (rn ) must converge to zero (since f (rn ) = 0 for all n). Thus f (x) = 0. (b) Consider the function f −g, for which (f −g)(r) = 0 for all rational numbers. Part (a) implies (f − g)(x) = 0 for all real x, that is, f (x) = g(x) for all real x. 6. (a) Let x0 be any point in X. For ϵ > 0 let δ = ϵ. Then for dX (x0 , x) < δ we have dX (f (x0 ), f (x)) ≤ cdX (x0 , x) < cδ = cϵ < ϵ. So f is continuous. (b) Note that for adjacent terms in the sequence we have dX (xk , xk+1 ) ≤ cdX (xk−1 , xk ). A simple induction yields dX (xk , xk+1 ) ≤ ck dX (x0 , x1 ) = ck dX (x0 , f (x0 )). WLOG assume m < n. By the triangle inequality n−1 ∑ dX (xm , xn ) ≤ dX (xk , xk+1 ) k=m ≤ dX (x0 , f (x0 )) n−1 ∑ ck k=m cm − cn = dX (x0 , f (x0 )) 1−c 1 − cn−m ≤ cm dX (x0 , f (x0 )) 1−c d (x , f (x )) X 0 0 ≤ cm . 1−c Because |c| < 1 we have cm → 0 as m → ∞. For any ϵ > 0 we can thus clearly choose some N so that for m ≥ N (and note then 2 n ≥ N ) we have from above dX (xm , xn ) < ϵ. The sequence is thus Cauchy. The sequence has a limit because X is complete and the sequence in Cauchy. (c) We have x∗ = lim xk = lim xk+1 = lim f (xk ) = f (x∗ ). k k k The last equality requires continuity of f . (d) Suppose another point x = y ∗ satisfies y ∗ = f (y ∗ ). Then dX (x∗ , y ∗ ) = dX (f (x∗ ), f (y ∗ )) ≤ cdX (x∗ , y ∗ ). If dX (x∗ , y ∗ ) > 0 we can divide both sides above by dX (x∗ , y ∗ ) and obtain 1 ≤ c, a contradiction. Thus dX (x∗ , y ∗ ) = 0, so that x∗ = y ∗ . 7. (a) Continuous functions on R map bounded open sets to bounded open sets: FALSE. Let f be any constant function. (b) Continuous functions on R map bounded open sets to open sets. FALSE: Same counterexample. (c) Continuous functions on R map bounded closed sets to bounded closed sets. TRUE: Bounded closed sets in R are compact, the image of a compact set under a continuous function is compact. (d) Continuous functions on R map bounded closed sets to closed sets. TRUE: Just a special case of the last question! 8. Note, for example, that p(−2) < 0 while p(0) = 8. By the Intermediate Value Theorem, there is at least one x ∈ (−2, 0) such that p(x) = 0. 3