1 September 12, 2014

1 September 12, 2014

1 September 12, 2014 Definition 1.1. f : U ⊂ Rn ! Rm is smooth if it has continuous partial derivatives of all orders. Need U to be open in order to define derivative. In order to define for arbitrary subsets, need local extension: Definition 1.2. f : X ⊂ Rn ! Rm is smooth if at each point x 2 X, there is an open set x 2 U ⊂ Rn and a smooth map F : U ! Rm such that F jU\X = f, i.e. f has a local extension in a neighborhood of each point. Recall open subsets of X in subspace topology are precisely sets of the form U \ X, so smoothness is a \local" property { f : X ! Rn is smooth if it smooth in a neighborhood of each point of x (in the subspace topology). Definition 1.3. f : X ⊂ Rn ! Y ⊂ Rm is a diffeomorphism if f and f −1 are smooth (and f is a bijection). examples: circles = knots 6= triangle, etc. Definition 1.4. A subset X ⊂ RN is a(smooth) k-dimensional manifold if it islocally diffeomorphic to Rk, that is every x 2 X has a neighborhood (open set) x 2 V ⊂ X that is diffeomorphic to some open set U ⊂ Rk. A diffeomorphism φ : U ! V is a parametrization of (the neighborhood) V . The inverse diffeomorphism φ−1 : V ! U is called a coordinate system on V . −1 −1 When φ is written in coordinates φ = (x1; : : : ; xk), the smooth func- tions x1; : : : ; xk are called coordinate functions. We say that dim X = k is the dimension of X. We can think of U and V as identified by phi { given coordinates (x1; : : : ; xk) on U, this gives us a point on V by using the parametrization φ, and given a point v 2 V , we can obtain the coordinates \of v" by taking (x1(v); : : : ; xk(v)). Example: we can show that S1 = f(x; y) 2 R2 : x2 + y2 = 1g is a 1-dimensional manifold. 1 We can parametrizep the set of (x; y) 2 S on the upper semicircle, i.e. 2 y > 0. φ1(x) = (x; 1 − x ) takes the open interval (−1; 1) to the upper −1 semicircle with a smooth map. The inverse φ1 (x; y) = x is smooth. For the 1 p 2 1 lower semicircle, we take φ2(x) = (x; − 1 − x ). This parametrizes all of S p 2 except for two points (±1; 0). To parametrize, we can take φ3 = ( 1 − y ; y) p 2 1 and φ4 = (− 1 − y ; y). These four parametrizations cover S and, for each point (at least) one of these gives a local parametrization. Proposition 1.5. Given X ⊂ RN and Y ⊂ RM manifolds, X × Y is a submanifold of RN+M . Also, dim X × Y = dim X + dim Y . Proof. Let k = dim X and l = dim Y . Suppose (x; y) 2 X × Y . Then, there exists and open set W ⊂ Rk and a local parametrization φ : W ! φ(W ) ⊂ X, as well as an open subset U ⊂ Rl along with a local parametrization : U ! Y . Define φ × : W × U ⊂ Rk+l ! X × Y by φ × (w; u) = (φ(w); (u)). One can check that φ × is a local parametrization of X × Y in a neighborhood of (x; y). Definition 1.6. If X and Z are both manifolds in RN and Z ⊂ X, then Z is a submanifold of X. 2 September 15, 2014 Suppose f is a smooth map from (an open set of) Rn to Rm and x is in the domain. Then, for any vector h 2 Rn, the derivative of f in the direction of h at x is f(x + th) − f(x) dfx(h) = lim : t!0 t n m We can define dfx : R R by h 7! dfx(h). This map is linear and has matrix form 2 @f1 ::: @f1 3 @x1 @xn 6 . 7 4 . 5 @fm ::: @fm @x1 @xn Proposition 2.1 (Chain Rule). d(g ◦ f)x = dgf(x)dfx: A linear map can be thought of as its best linear approximation. We can use derivatives to identify the best linear approximation a manifold X at x. Let φ : U ! X be a local parametrization around x, and assume φ(0) = x for convenience. Then, the best linear approximation is u 7! φ(0) + dφ0(u) = x + dφ0(u): 2 Definition 2.2. The tangent space of X at x, Tx(X) is the image of the map k N dφ0 : R ! R . N Tx(X) is a vector subspace of R , and the translation x + Tx(X) is the \best linear approximation" of X at x. For this to be a real definition, need to make sure it is well-defined. Let : V ! X be another parametrization. Then, by shrinking U and V , can assume that φ(U) = (V ). The map h : −1 ◦φ : U ! V is a diffeomorphism (composition). Then, φ = ◦ h, so by chain rule dφ0 = d 0 ◦ dh0. Hence, the image of dφ0 is contained in the image of d 0. The reverse also holds similarly, so they must be identical. Proposition 2.3. dim Tx(X) = dim X. Proof. Let φ : U ! V be smooth, and let Φ0 : W ! U be a smooth map −1 0 0 that extends φ . Then Φ ◦ φ = id, so dΦx ◦ dφ = id. This implies that dφ must be an isomorphism, i.e. dim = k. 3 September 17, 2014 1 For an examplep of a tangent space, we can take the parametrization of S , φ : x 7! (x; 1 − x2). Then 1 dφ = p x : 1−x2 1 pFor each point x 2 R, this gives a vector that is tangent to S at (x; 1 − x2). We can do a similar example for S2, with a parametrization φ :(x; y) 7! (x; y; p1 − x2 − y2). Then, 2 1 0 3 dφ = 6 0 1 7 : 4p x p y 5 1−x2−y2 1−x2−y2 The column space of the matrix gives a 2-dimensional subspace of R3 which is tangent to S2 at φ(x; y). Now want to take derivative of f : X ! Y . 3 We take parametrizations φ : U ⊂ Rk ! RN and : V ⊂ Rl ! RM . WLOG take φ(0) = x; (0) = y. We have commutative diagram: X ! (f) Y U ! (h = −1 ◦ f ◦ φ) V −1 So we can define dfx = d 0 ◦ dh0 ◦ dφ0 . We claim this is independent of parametrization. Indeed, if Ψ and Φ are two different parametrizations of X and Y , and H = Ψ−1 ◦ f ◦ Φ, then h = −1 ◦ Ψ ◦ Ψ−1 ◦ f ◦ Φ ◦ Φ−1 ◦ φ = −1 ◦ Ψ ◦ H ◦ Φ−1 ◦ φ. Then, −1 −1 −1 −1 −1 d 0 ◦ dh0 ◦ dφ0 = d ◦ d ◦ dΨ ◦ dH ◦ dΦ ◦ φ ◦ dφ = dΦ ◦ dH ◦ dΦ : A similar argument shows: Proposition 3.1 (Chain Rule). d(g ◦ f)x = dgf(x) ◦ dfx. Proof. Let φ : U ! X, : V ! Y and ν : W ! Z. Then h = −1 ◦ f ◦ φ : U ! V and j = ν−1 ◦ g ◦ : V ! W , and j ◦ h : U ! W . −1 By definition, d(g ◦ f)x = dν0 ◦ d(j ◦ h)0 ◦ dφ0 . N By the chain rule in R , d(j ◦ h) = (dj)0 ◦ (dh)0, and also (dj)0 ◦ (dh)0 = −1 (dj)0 ◦ d 0 ◦ d ◦ (dh)0. Substituting yields the desired equality. 4 September 19, 2014 We want to use the tangent space to study smooth manifolds. Definition 4.1. A function f : X ! Y is a local diffeomorphism if for every x 2 X, there exists a neighborhood x 2 U that maps diffeomorphically to a neighborhood f(U) of y = f(x). In order to be a local diffeomorphism, note that dfx : Tx(X) ! Ty(Y ) must be an isomorphism. This follows from a chain rule argument if f : U ! V is a diffeomorphism. There is a version of the inverse function theorem for smooth manifolds: Theorem 4.2 (Inverse Function Theorem). Suppose that f : X ! Y is a smooth map whose derivative dfx at the point x is an isomorphism (of vector spaces). Then f is a local diffeomorphism at x. As an example, f : R ! S1 : t 7! (cos t; sin t) is a local diffeomorphism. 4 5 September 22, 2014 A local diffeomorphism says that locally, X and Y \look the same". We can make this more precise by noting that we can find local coordinates around x and y so that f(x1; : : : ; xk) = (x1; : : : ; xk). In particular, if φ : U ! X is a parametrization around x 2 X, and dfx is an isomorphism, then the composition : f ◦ φ : U ! Y is a parametrization around y = f(x) 2 Y (here it may be necessary to shrink U to a smaller set so that f is a local diffeomorphism). Definition 5.1. Two maps f : X ! Y and f 0 : X0 ! Y 0 are equivalent if there exist diffeomorphisms α : X0 ! X and β : Y 0 ! Y so that f ◦α = β ◦f 0 (i.e. diagram commutes). f is locally equivalent to f 0 at x if there is a local diffeomorphism at x and a local diffeomorphism at y = f(x) that makes the diagram commute. The inverse function theorem can be interpreted to say that if dfx is an isomorphism, then f is locally isomorphic to the identity. If dim X < dim Y , then IFT cannot apply, since the matrix dfx is not square. The closest thing we can hope for is dfx is injective. Definition 5.2. f : X ! Y is an immersion at x if dfx : Tx(X) ! Ty(Y ) is injective.

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    49 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us