SOLUTION OF HW4 MINGFENG ZHAO September 30, 2012 1. [2 Points, Exercises 3, on Page 85] Use Lemma 2.3.2 to prove that if f 0 (x0 ) > 0, there is a δ > 0 such that f (x) < f (x0 ) if x0 − δ < x < x0 and f (x) > f (x0 ) if x0 < x < x0 + δ. Proof. By the Lemma 2.3.2, on Page 76, we know that f (x) = f (x0 ) + [f 0 (x0 ) + E(x)](x − x0 ), and lim E(x) = E(x0 ) = 0. x→x0 Since f 0 (x0 ) > 0, then there exists some δ > 0 such that for all 0 < |x − x0 | < δ, we have |E(x)| < f 0 (x0 ). In particular, we know that f 0 (x0 ) + E(x) > 0, ∀0 < |x − x0 | < δ. So for all x0 − δ < x < x0 , we have x − x0 < 0 and f (x) − f (x0 ) = [f 0 (x0 ) + E(x)](x − x0 ) < 0. For all x0 < x < x0 + δ, we have x − x0 > 0 and f (x) − f (x0 ) = [f 0 (x0 ) + E(x)](x − x0 ) > 0. In summary, we get there is a δ > 0 such that f (x) < f (x0 ) if x0 − δ < x < x0 and f (x) > f (x0 ) 1 if x0 < x < x0 + δ. 2 MINGFENG ZHAO 2. [2 Points, Exercises 7, on Page 85] Suppose that c0 (0) = a and s0 (0) = b where a2 + b2 6= 0, and c(x + y) = c(x)c(y) − s(x)s(y) s(x + y) = s(x)c(y) + c(x)s(y) for all x and y. a. Show that c and s are differentiable on (−∞, ∞), and find c0 and s0 in terms of c, s, a and b. b. Find c and s explicitly. Proof. a. For any x ∈ R and fixed, for any y ∈ R\{0}, then c(x + y) = c(x)c(y) − s(x)s(y) c(x) = c(x)c(0) − s(x)s(0) s(x + y) = s(x)c(y) + c(x)s(y) s(x) = s(x)c(0) + c(x)s(0). Then, we have c(x + y) − c(x) y = [c(x)c(y) − s(x)s(y)] − [c(x)c(0) − s(x)s(0)] y = c(x) · s(x + y) − s(x) y = s(y) − s(0) c(y) − c(0) − s(x) · y y [s(x)c(y) + c(x)s(y)] − [s(x)c(0) + c(x)s(0)] y = s(x) · c(y) − c(0) s(y) − s(0) + c(x) · . y y Since c0 (0) = a and s0 (0) = b, that is, c(y) − c(0) = a, y→0 y lim and s(y) − s(0) = b. y→0 y lim SOLUTION OF HW4 3 Hence, we know that lim c(x + y) − c(x) y = c(x) · a − s(x) · b = ac(x) − bs(x) lim s(x + y) − s(x) y = s(x) · a + c(x) · b = as(x) + bc(x). y→0 y→0 Hence c(x) and s(x) are differentiable on (−∞, ∞), c0 (x) = ac(x) − bs(x) and s0 (x) = as(x) + bc(x) for all x ∈ R. b. Since c0 (0) = a and s0 (0) = b, then a = ac(0) − bs(0), and b = as(0) + bc(0). That is, −b c(0) a = · b s(0) a a b . Since a2 + b2 6= 0, then c(0) = s(0) 1 2 a + b2 a · −b b a · b a 1 a2 + b2 = = a2 + b2 · −ab + ab 1 . 0 Hence, c(x) and s(x) satisfy the following ODE system: 0 c(x) a = s(x) b −b c(x) · a s(x) and c(0) 1 = . s(0) 0 4 MINGFENG ZHAO a Let’s compute the eigenvalues of b −b , that is, a λ−a det −b b λ−a = (λ − a)2 + b2 = 0. Hence, we get λ = a ± ib. Let’s solve the following λ−a −b y1 ±ib b = · −b ±ib y2 λ−a b y1 0 . = · 0 y2 Case I: λ = a + ib with b 6= 0, then y1 i = y2 · y2 1 Case II: λ = a − ib with b 6= 0, then −i y1 = y2 · 1 y2 So when b 6= 0, we know that there exists some A, B such that c(x) i −i = Ae(a+ib)x · + Be(a−ib)x · s(x) 1 1 and Hence, we get i −i 1 A· +B· = 1 1 0 That is, we get i(A − B) A+B 1 = . 0 . c(0) 1 = . s(0) 0 SOLUTION OF HW4 5 Which implies that i A=− , 2 and B = i . 2 Hence, we get 12 e(a+ib)x + 12 e(a−ib)x c(x) = s(x) − 2i e(a+ib)x + 2i e(a−ib)x eax cos bx = eax sin bx . That is, c(x) = eax cos bx and s(x) = eax sin bx. If b = 0, since a2 + b2 6= 0, then a 6= 0. Hence we get c0 (x) = ac(x) and s0 (x) = as(x), which implies that c(x) = c(0)eax = eax and s(x) = s(0)eax = 0. In summary, we know that c(x) = eax cos bx and s(x) = eax sin bx. 3. [2 Points, Exercises 13, on Page 86] What is wrong with the “proof” of the chain rule suggested after Example 2.3.3? Corrected it. Proof. If g(x) ≡ g(x0 ) in a neighborhood of x0 , then f (g(x)) − f (g(x0 )) g(x) − g(x0 ) make no sense. Correction: Assume that g(x) 6= g(x0 ) in a neighborhood of x0 , then the proof works. In fact, since f is differentiable at g(x0 ), then f (t) − f (g(x0 )) = [f 0 (g(x0 )) + E(t)][t − g(x0 )], 6 MINGFENG ZHAO where lim E(t) = E(g(x0 )) = 0. t→g(x0 ) So we know that lim t→g(x0 ) f (t) − f (g(x0 )) = f 0 (g(x0 )). t − g(x0 ) Replacing t be g(x) since g(x) 6= g(x0 ) for all x 6= x0 in a neighborhood of g(x0 ), then we get lim x→x0 f (g(x)) − f (g(x0 )) = f 0 (g(x0 )). g(x) − g(x0 ) 4. [2 Points, Exercises 14, on Page 86] Suppose that f is continuous and strictly increasing on [a, b]. Let f be differentiable at a point x0 in (a, b), with f 0 (x0 ) 6= 0. If g is the inverse of f , show that g 0 (f (x0 )) = 1 f 0 (x 0) . Proof. Since f is strictly increasing on [a, b], then f (a) < f (b). Since f is continuous on [a, b], then f ([a, b]) = [f (a), f (b)] and g : [f (a), f (b)] → [a, b]. Claim I: g is continuous on [f (a), f (b)]. For any yn , y0 ∈ [f (a), f (b)] such that yn → y0 as n → ∞. Let xn = g(yn ) and x0 = g(y0 ). We want to show that xn → x0 as n → ∞, otherwise, there exists some 0 > 0 such that there exists a subsequence of xn , without loss of generality, assume xn itself such that |xn − x0 | ≥ 0 , ∀n ≥ 1. In the sequence xn , we can choose a subsequence without loss of generality, assume xn itself such that x n − x 0 ≥ 0 , ∀n ≥ 1, or xn − x0 ≤ −0 , Without loss of generality, we assume x n − x 0 ≥ 0 , ∀n ≥ 1. ∀n ≥ 1. SOLUTION OF HW4 7 Hence we get x n ≥ x 0 + 0 , ∀n ≥ 1. Since f is continuous and strictly increasing, then yn = f (xn ) ≥ f (x0 + 0 ) > f (x0 ), ∀n ≥ 1. Which contradicts with yn → y0 as n → ∞. Hence we must have xn → x0 as n → ∞. Therefore, g is continuous. Claim II: x → x0 if and only if f (x) → x0 . If x → x0 , since f is continuous, then f (x) → f (x0 ). On the other hand, if f (x) → f (x0 ), applying g on both sides, since g is also continuous, then x = g(f (x)) → x0 = g(f (x0 )). Since f 0 (x0 ) 6= 0, then f 0 (x0 ) > 0. By the result of the Problem 1, we know that there exists some δ > 0 such that f (x) 6= f (x0 ) for all 0 < |x − x0 | < δ. Hence, we know that g(f (x)) − g(f (x0 )) f (x) − f (x0 ) = = → x − x0 f (x) − f (x0 ) 1 f (x)−f (x0 ) x−x0 1 f 0 (x0 ) as x → x0 ⇐⇒ as f (x) → f (x0 ) Since f is increasing and g is continuous. Hence, we know that g 0 (f (x0 )) exists and g 0 (f (x0 )) = 1 . f 0 (x0 ) 5. [2 Points, Exercises 16, on Page 86] Show that f (a+) and f (b−) exists if f 0 is bounded on (a, b). Proof. Since f 0 is bounded, then there exists some M > 0 such that |f 0 (x)| ≤ M, Claim I: f is uniformly continuous on (a, b). for all x ∈ (a, b). 8 MINGFENG ZHAO For any x, y ∈ (a, b), by the mean value theorem, there exist some z ∈ [x, y] or z ∈ [y, x] such that |f (x) − f (y)| = |f 0 (z)(x − y)| ≤ M |x − y|. Which implies that f is uniformly continuous on (a, b). In particular, for any > 0, there exists some δ > 0 such that for all |x − y| < 2δ, then we have |f (x) − f (y)| < . Now for any |x − a|, |y − a| < δ, then |x − y| < 2δ, then |f (x) − f (y)| < . That is, f (x) is Cauchy as x → a, Hence f (a+) exists. Now for any |x − b|, |y − b| < δ, then |x − y| < 2δ, then |f (x) − f (y)| < . That is, f (x) is Cauchy as x → b, Hence f (b−) exists. 0 6. [0 Points, Exercises 15, on Page 86] a. Show that f+ (a) = f 0 (a+) if both quantities exist. 0 b. Example 2.3.4 shows that f+ (a) may exist even if f 0 (a+) does not. Give an example where 0 f 0 (a+) exists but f+ (a) does not. c. Complete the following statement so it becomes a theorem, and prove the theorem: “ If f 0 (a+) exists and f is 0 at a, then f+ (a) = f 0 (a+).” 0 Proof. a. Recall the definition of f+ (a) and f 0 (a+): 0 f+ (a) = lim x→a+ f (x) − f (a) , x−a f 0 (a+) = lim f 0 (x). x→a+ SOLUTION OF HW4 9 In a neighborhood of x, say there exists some δ > 0, we know that f is continuous on [a, a + δ] and f is differentiable on (a, a + δ]. By the mean value theorem, for any n ≥ 1, there exists some yn ∈ (a, a + n1 ) such that f (a + 1 1 ) − f (a) = f 0 (yn ) . n n Hence, we get f 0 (yn ) = f (a + n1 ) − f (a) 1 n . By taking n → ∞, then f 0 (a+) = f+ (a). b. Let a = 0 and f (x) = x2 + 1, x 6= 0 0, x = 0. 0 Then f is not continuous at x = 0, hence f+ (0) does not exist. But f 0 (x) = 2x for all x > 0, and f 0 (0+) = 0. 0 c. Claim I: If f 0 (a+) exists and f is continuous at a, then f+ (a) = f 0 (a+). In a neighborhood of x, say there exists some δ > 0, we know that f is continuous on [a, a + δ] and f is differentiable on (a, a + δ]. By the mean value theorem, for any x, there exists some yx ∈ (a, x) such that f (x) − f (a) = f 0 (yx )(x − a). Hence, we get f 0 (yx ) = f (x) − f (a) . x−a By taking x → a, then yx → a. Since f 0 (a+) exists, then f (x) − f (a) 0 = f+ (a). x→a+ x−a f 0 (a+) = lim f 0 (yx ) = lim x→a 0 Hence f 0 (a+) = f+ (a). 10 MINGFENG ZHAO Department of Mathematics, University of Connecticut, 196 Auditorium Road, Unit 3009, Storrs, CT 06269-3009 E-mail address: mingfeng.zhao@uconn.edu