The University of Sydney School of Mathematics and Statistics Solutions to Tutorial 5 MATH3961: Metric Spaces Semester 1 Lecturer: Laurentiu Paunescu 1. Show that the mapping f : R → R+ defined by f (x) = ex is a homeomorphism from R onto R+ . Here R+ = {x ∈ R | x > 0} . Solution: It is known that f : R → R+ defined by f (x) = ex is continuous. d x e = ex > 0 for all x ∈ R , it follows that f is strictly increasing and so Since dx f is one-to-one. The image of R under f is clearly R+ , i.e. f is onto. Thus the inverse mapping f −1 : R+ → R exists, and in fact it is given by f −1 (y) = ln y , (y ∈ R+ ) . It is also known that the ln function is continuous. Hence the given function f is a homeomorphism from R onto R+ . [A stronger result is true. If I ⊆ R and J ⊆ R are intervals and f : I → J is a function which is strictly increasing (x < y ⇒ f (x) < f (y)) and onto then (a) f is invertible and f −1 : J → I is strictly increasing and onto; (b) f is continuous; (c) hence f −1 is also continuous, and so f is a homeomorphism.] 2. Show that the mapping f : R2 → R2 defined by f (x1 , x2 ) = (x1 2 , x2 2 ) gives a homeomorphism from the space A = {(x1 , x2 ) ∈ R2 | x1 ≥ 0, x2 ≥ 0} onto A ; but that it is not a homeomorphism from R2 onto A . Solution: The mapping f : R2 → R2 defined by f (x1 , x2 ) = (x1 2 , x2 2 ) is continuous on R2 , since each of the component functions f1 (x1 , x2 ) = x1 2 and f2 (x1 , x2 ) = x2 2 is continuous. Hence the restriction of f to A gives a continuous mapping, denoted again by f, from A to R2 . Now f is one-to-one on A : Suppose that (x1 , x2 ), (y1 , y2 ) ∈ A are such that f (x1 , x2 ) = f (y1 , y2 ) . Then (x1 2 , x2 2 ) = (y1 2 , y2 2 ) , so that x1 2 = y1 2 , x2 2 = y2 2 c 2014 The University of Sydney Copyright 1 which implies that x1 = y1 , x2 = y2 or (x1 , x2 ) = (y1 , y2 ) . The image of A under f is A : For if (a1 , a2 ) ∈ A , then a1 ≥ 0, a2 ≥ 0 so that √ √ a1 ≥ 0, a2 ≥ 0 and √ √ f ( a1 , a2 ) = (a1 , a2 ) . Thus f : A → A is one-to-one and onto and so the inverse f −1 of f exists. In fact, the inverse mapping f −1 : A → A is given by √ √ f −1 (x1 , x2 ) = ( x1 , x2 ), (x1 , x2 ) ∈ A . Since the component functions of f −1 are both continuous, it follows that f −1 is a continuous mapping. Hence the given f gives a homeomorphism from A onto A . Finally, since f is not one-to-one on R2 , for example, f (−1, 0) = f (1, 0) = (1, 0) , we conclude that f is not a homeomorphism from R2 to A . 3. Let A = (x1 , x2 , x3 ) ∈ R3 | x1 2 + x2 2 + x3 2 = 1, (x1 , x2 , x3 ) 6= (0, 0, 1) . Show that A is homeomorphic to R2 . Solution: Let S 2 be the unit sphere S 2 = (x1 , x2 , x3 ) ∈ R3 | x1 2 + x2 2 + x3 2 = 1 , and let N = (0, 0, 1) . Then A = S 2 \ {(0, 0, 1)} . x3 N = (0, 0, 1) R2 = x1 x2 -plane P O x2 P′ x1 The radial projection, with centre N , of A onto the x1 x2 -plane, corresponds to 2 f : A → R2 which is given by x x2 1 , , f (x1 , x2 , x3 ) = 1 − x3 1 − x3 (x3 6= 1) . Since the component functions are continuous, f is continuous. Since each ray, not in the tangent plane x3 = 1, meets each of A and x1 x2 -plane in exactly one point, and these rays contain all the points of A and the x1 x2 -plane, we see that f is a one-to-one mapping from A onto R2 . Thus the inverse mapping f −1 : R2 → A of f exists and it is given by (1) f −1 (y1 , y2 ) = 2y1 2y2 y1 2 + y2 2 − 1 , , . y1 2 + y2 2 + 1 y1 2 + y2 2 + 1 y1 2 + y2 2 + 1 In fact, suppose that f −1 (y1 , y2 ) = (x1 , x2 , x3 ) ∈ A . Then x1 2 + x2 2 + x3 2 = 1 and we have (2) y1 = x2 x1 , y2 = . 1 − x3 1 − x3 N = (0, 0, 1) 1 − x3 Q (x1 , x2 , x3 ) ′ O Q (y1 , y2 , 0) 1 As shown in the figure, we see that 1 − x3 N Q′ = . 1 NQ Thus x1 2 + x2 2 + (x3 − 1)2 y1 2 + y2 2 + 1 x1 2 + x2 2 + x3 2 − 2x3 + 1 = y1 2 + y2 2 + 1 2(1 − x3 ) . = 2 y1 + y2 2 + 1 (1 − x3 )2 = 3 Therefore 1 − x3 = y1 2 2 , + y2 2 + 1 and so by (2), we obtain x1 = y1 2 Also, x3 = 1 − 2y1 2y2 , x2 = 2 . 2 + y2 + 1 y1 + y2 2 + 1 2 y1 2 + y2 2 − 1 = . y1 2 + y2 2 + 1 y1 2 + y2 2 + 1 Hence f −1 (y1 , y2 ) is given by (1). Finally, since all the components of f −1 are continuous, f −1 is continuous. Consequently A is homeomorphic to R2 . 4. Let (X, d) be any metric space. Let d1 and d2 be the metrics on X defined by d1 (x, y) = min 1, d(x, y) , and d2 (x, y) = d(x, y) . 1 + d(x, y) Prove that d, d1 and d2 are all equivalent on X. Solution: Clearly d1 (x, y) ≤ d(x, y) for all x, y ∈ X, so that the identity mapping of (X, d) onto (X, d1 ) is continuous. Now, we will prove that the identity mapping of (X, d1 ) onto (X, d) is continuous. For any ε > 0, we choose δ = min{1, ε}. Then we see that d1 (x, y) < δ =⇒ d(x, y) < ε. Thus the identity mapping of (X, d1 ) onto (X, d) is continuous and so the identity mapping of (X, d1 ) onto (X, d) is a homeomorphism. Hence d and d1 are equivalent. Next, since d2 (x, y) ≤ d(x, y), for all x, y ∈ R, we see that the identity mapping of (X, d) onto (X, d2 ) is continuous. Now, we will prove that the identity mapping of (X, d2 ) onto (X, d) is continuous. To show this, observe that if ε > 0 and d(x, y) ε d2 (x, y) = < , 1 + d(x, y) 1+ε then d(x, y) < ε. Thus given ε > 0, if we set δ = ε/(1 + ε), then we have d2 (x, y) < δ =⇒ d(x, y) < ε. Thus the identity mapping of (X, d2 ) onto (X, d) is continuous and so the identity mapping of (X, d2 ) onto (X, d) is a homeomorphism. Hence d and d2 are equivalent. Consequently, d, d1 and d2 are all equivalent on X. 4 5. Let X = [1, ∞) with d the Euclidean metric and let f : X → X be given by 25 f (x) = (x + 1/x). 26 Show that 25 d(x, y), d f (x), f (y) ≤ 26 [and so f is a contraction mapping] and by solving the equation algebraically, show that 5 is the unique fixed point of f . Solution: For any x, y ∈ X, we have 25 25 d f (x), f (y) = |f (x) − f (y)| = (x + 1/x) − (y + 1/y) 26 26 1 25 25 d(x, y), 1 − |x − y| < = 26 xy 26 since x ≥ 1 and y ≥ 1. Hence f is a contraction mapping. If f (x) = x, then 25 (x + 1/x) = x 26 which implies that x2 = 25 and so x = 5 is the unique fixed point of f in X. 6. Let X = [0, 1] with d the Euclidean metric and let f : X → X be given by 1 f (x) = (x3 + x2 + 1). 7 Show that 5 d f (x), f (y) ≤ d(x, y). 7 (2) (3) Calculate f (0), f (0), f (0), . . . , and hence find, to three decimal places, the fixed point of f . Solution: For any x, y ∈ X, we see that 1 3 1 3 2 2 d f (x), f (y) = |f (x) − f (y)| = (x + x + 1) − (y + y + 1) 7 7 1 3 1 |(x − y 3 ) + (x2 − y 2 )| = |(x − y)(x2 + xy + y 2 + x + y)| 7 7 5 5 ≤ |x − y| = d(x, y), 7 7 = since 0 ≤ x ≤ 1 and 0 ≤ y ≤ 1. Using the calculator, one can compute that f (0) = 0.142857, f (2) (0) = f (0.142857) = 0.1457767, f (3) (0) = f (0.1457767) = 0.145897, · · · . Hence the fixed point of f is x = 0.146, to three decimal places. 5 7. Let f : [a, b] → [a, b] be differentiable over [a, b]. Show that f is a contraction mapping if and only if there exists a number K < 1 such that for all x ∈ (a, b), |f ′ (x)| ≤ K. Solution: Suppose first that f is a contraction mapping, i.e. that there exists K < 1 such that |f (x) − f (y)| ≤ K|x − y|, for all x, y ∈ [a, b]. Then, in particular, for and x and x + δx in [a, b], we have |f (x + δx) − f (x)| ≤ K|(x + δx) − x| = K|δx|. Hence for δx 6= 0, f (x + δx) − f (x) ≤ K, δx and the limit of the left-hand expression as δx → 0 is also less than or equal to K. But that limit is precisely |f ′ (x)|. Hence there exists K < 1 such that for all x ∈ (a, b), |f ′ (x)| ≤ K. Conversely, asssume that for all x ∈ (a, b), |f ′ (x)| ≤ K < 1. Now for any x 6= y in [a, b], there is c between x and y such that f (x) − f (y) = f ′ (c). x−y But |f ′ (c)| ≤ K so that Hence, for all x, y ∈ [a, b], f (x) − f (y) ′ x − y = |f (c)| ≤ K. |f (x) − f (y)| ≤ K|x − y|, and so f is a contraction mapping. 8. Let X = C[0, 1] be the metric space with d given by d(f, g) = sup |f (x) − g(x)|. x∈[0, 1] Define F : X → X by {F (f )}(x) = Z x f (t) dt 0 Show that 6 (f ∈ X). 1. {F (f )}(x) − {F (g)}(x) ≤ x d(f, g), 2. {F (2) (f )}(x) − {F (2) (g)}(x) ≤ x2 d(f, g), 2 for all f, g ∈ X, and deduce that F (2) is a contraction mapping. Show, however, that F is not a contraction mapping. Solution: 1. We have x {F (f )}(x) − {F (g)}(x) = Z ≤ Z (g)}(x) = Z ≤ Z f (t) − g(t) dt 0 x 0 d(f, g) dt ≤ x d(f, g). 2. We have {F (2) (f )}(x) − {F (2) x 0 {F (f )}(t) − {F (g)}(t) dt x t d(f, g) dt = 0 x2 d(f, g). 2 It follows that d F (f ), F (g) = sup |{F (f )}(x) − {F (g)}(x)| x∈[0, 1] ≤ sup x d(f, g) ≤ d(f, g), x∈[0, 1] and d F (2) (f ), F (2) (g) = sup |{F (2) (f )}(x) − {F (2) (g)}(x)| x∈[0, 1] ≤ sup x∈[0, 1] x2 1 d(f, g) ≤ d(f, g). 2 2 Hence F (2) is a contraction mapping. But F is not, since by taking f = 1 and g = 0, we have d F (f ), F (g) supx∈[0,1] x = 1 = d(f, g). = 9. Let A = S 1 be the unit circle in R2 and B = [0, 2π) ⊆ R. Prove that the mapping f : B → A defined by f (x) = (cos x, sin x) 7 is a continuous bijection, but that f −1 is not continuous. Solution: Since each of the components of f is continuous, it follows that f is continuous. The mapping f is clearly a bijection and so f −1 : A → B exists. To show that f −1 is not continuous, we need to produce an open set U in B for which the inverse image (f −1 )−1 (U ) = f (U ) is not open in A. Take U = [0, π/2). Then clearly U is open in B. But f (U ) = {(cos x, sin x) | 0 ≤ x < π/2} is not open in A. In fact, if we take the point y = (1, 0) ∈ f (U ), then we see that for any open ball BA (y; ε) in A, BA (y; ε) ∩ (A \ f (U )) 6= ∅, and so f (U ) is not open in A. Hence f −1 is not continuous. [Recall that BA (y; ε) = B(y; ε) ∩ A.] 10. Let X = Rn and let d1 and d2 be the metrics on X given by d1 (x, y) = n nX k=1 |xk − yk |p o1/p , where p ≥ 1, and d2 (x, y) = max(|x1 − y1 |, |x2 − y2 |, . . . , |xn − yn |) . Prove that d1 and d2 are equivalent over X. Solution: Clearly, for k = 1, 2, . . . , n d1 (x, y) ≥ |xk − yk | so that d1 (x, y) ≥ d2 (x, y) , for all x, y ∈ X. Next, we see that d1 (x, y) ≤ n nX = n nX k=1 (max |xk − yk |) K=1 k d2 (x, y) p o1/p p 1/p = n d2 (x, y) , and so for all x, y ∈ X. d1 (x, y) ≤ n1/p d2 (x, y) , Hence d1 and d2 are equivalent over X. 8 p o1/p 11. Let A = (x1 , x2 , x3 ) ∈ R3 | x1 2 + x2 2 + x3 2 = 1, x3 < 0 . Show that A is homeomorphic to R2 . Solution: Let P be the plane x3 = −1. Then it is easy to see that A touches the plane P at (0, 0, −1) . x3 x2 x1 P (0, 0, −1) P′ P : x3 = −1 Now we project A from O onto the plane P . Similar to Question 4, one finds that the homeomorphism f : A → P is given by x x2 1 f (x1 , x2 , x3 ) = − , − , −1 , (x1 , x2 , x3 ) ∈ A ; x3 x3 and its inverse f −1 : P → A is given by y2 −1 y1 −1 , p , p . f (y1 , y2 , −1) = p y1 2 + y2 2 + 1 y1 2 + y2 2 + 1 y1 2 + y2 2 + 1 (Please check the answer!) Hence A is homeomorphic to P . But P is clearly homeomorphic to R2 . Consequently A is homeomorphic to R2 . 12. Let A = (x1 , x2 ) ∈ R2 | x1 2 + x2 2 < 1 . Show that A is homeomorphic to R2 . Solution: By Question 5, we know that the southern hemisphere A1 (without the equator) is homeomorphic to R2 . Now, the mapping f : A1 → A given by f (x1 , x2 , x3 ) = (x1 , x2 ) , (x1 , x2 , x3 ) ∈ A1 , is clearly continuous, one-to-one and onto, and its inverse f −1 : A → A1 is given by p f −1 (y1 , y2 ) = y1 , y2 , − 1 − y1 2 − y2 2 . Again, we see that f −1 is continuous. Hence A is homeomorphic to A1 . Since A1 is homeomorphic to R2 , it follows that A is homeomorphic to R2 . 9