Differential Topology Final Exam With Solutions

Similar documents
Differential Topology Solution Set #2

MATH 23a, FALL 2002 THEORETICAL LINEAR ALGEBRA AND MULTIVARIABLE CALCULUS Solutions to Final Exam (in-class portion) January 22, 2003

Practice Exam. 2x 1 + 4x 2 + 2x 3 = 4 x 1 + 2x 2 + 3x 3 = 1 2x 1 + 3x 2 + 4x 3 = 5

Problems in Linear Algebra and Representation Theory

j=1 x j p, if 1 p <, x i ξ : x i < ξ} 0 as p.

Review problems for MA 54, Fall 2004.

LINEAR ALGEBRA 1, 2012-I PARTIAL EXAM 3 SOLUTIONS TO PRACTICE PROBLEMS

Matrix Lie groups. and their Lie algebras. Mahmood Alaghmandan. A project in fulfillment of the requirement for the Lie algebra course

MA 1B ANALYTIC - HOMEWORK SET 7 SOLUTIONS

Review of linear algebra

Duke University, Department of Electrical and Computer Engineering Optimization for Scientists and Engineers c Alex Bronstein, 2014

Math 554 Qualifying Exam. You may use any theorems from the textbook. Any other claims must be proved in details.

LINEAR ALGEBRA QUESTION BANK

Mathematical Methods wk 2: Linear Operators

1. General Vector Spaces

Math 215B: Solutions 1

Exercise Sheet 1.

Linear Algebra- Final Exam Review

MATH 235. Final ANSWERS May 5, 2015

Differential Geometry qualifying exam 562 January 2019 Show all your work for full credit

PRACTICE FINAL EXAM. why. If they are dependent, exhibit a linear dependence relation among them.

OHSx XM511 Linear Algebra: Solutions to Online True/False Exercises

Notes 10: Consequences of Eli Cartan s theorem.

08a. Operators on Hilbert spaces. 1. Boundedness, continuity, operator norms

Math Camp Lecture 4: Linear Algebra. Xiao Yu Wang. Aug 2010 MIT. Xiao Yu Wang (MIT) Math Camp /10 1 / 88

Calculating determinants for larger matrices

Queens College, CUNY, Department of Computer Science Numerical Methods CSCI 361 / 761 Spring 2018 Instructor: Dr. Sateesh Mane.

1. Foundations of Numerics from Advanced Mathematics. Linear Algebra

2. Linear algebra. matrices and vectors. linear equations. range and nullspace of matrices. function of vectors, gradient and Hessian

Contents. Preface for the Instructor. Preface for the Student. xvii. Acknowledgments. 1 Vector Spaces 1 1.A R n and C n 2

Spring 2018 CIS 610. Advanced Geometric Methods in Computer Science Jean Gallier Homework 3

LECTURE 25-26: CARTAN S THEOREM OF MAXIMAL TORI. 1. Maximal Tori

THE INVERSE FUNCTION THEOREM FOR LIPSCHITZ MAPS

Math 147, Homework 1 Solutions Due: April 10, 2012

The goal of this chapter is to study linear systems of ordinary differential equations: dt,..., dx ) T

Math 302 Outcome Statements Winter 2013

LECTURE 10: THE ATIYAH-GUILLEMIN-STERNBERG CONVEXITY THEOREM

Notation. For any Lie group G, we set G 0 to be the connected component of the identity.

18.06 Problem Set 8 - Solutions Due Wednesday, 14 November 2007 at 4 pm in

Linear algebra I Homework #1 due Thursday, Oct Show that the diagonals of a square are orthogonal to one another.

Therefore, A and B have the same characteristic polynomial and hence, the same eigenvalues.

MANIFOLD STRUCTURES IN ALGEBRA

Math Linear Algebra Final Exam Review Sheet

Glossary of Linear Algebra Terms. Prepared by Vince Zaccone For Campus Learning Assistance Services at UCSB

Linear Algebra Final Exam Solutions, December 13, 2008

Lecture 1 and 2: Random Spanning Trees

Math Topology II: Smooth Manifolds. Spring Homework 2 Solution Submit solutions to the following problems:

Final A. Problem Points Score Total 100. Math115A Nadja Hempel 03/23/2017

Review of Linear Algebra

Chapter 5 Eigenvalues and Eigenvectors

NOTES ON LINEAR ODES

Homework set 4 - Solutions

Spectral Theorem for Self-adjoint Linear Operators

Review of Linear Algebra Definitions, Change of Basis, Trace, Spectral Theorem

The value of a problem is not so much coming up with the answer as in the ideas and attempted ideas it forces on the would be solver I.N.

GRE Subject test preparation Spring 2016 Topic: Abstract Algebra, Linear Algebra, Number Theory.

LECTURE 16: LIE GROUPS AND THEIR LIE ALGEBRAS. 1. Lie groups

Lecture 10 - Eigenvalues problem

SOLUTIONS TO THE FINAL EXAM

Math 141 Final Exam December 18, 2014

MATRIX LIE GROUPS AND LIE GROUPS

TOEPLITZ OPERATORS. Toeplitz studied infinite matrices with NW-SE diagonals constant. f e C :

University of Colorado at Denver Mathematics Department Applied Linear Algebra Preliminary Exam With Solutions 16 January 2009, 10:00 am 2:00 pm

SYLLABUS. 1 Linear maps and matrices

Introduction to Topology

ALGEBRA QUALIFYING EXAM PROBLEMS LINEAR ALGEBRA

Let us recall in a nutshell the definition of some important algebraic structure, increasingly more refined than that of group.

Examples include: (a) the Lorenz system for climate and weather modeling (b) the Hodgkin-Huxley system for neuron modeling

NONCOMMUTATIVE POLYNOMIAL EQUATIONS. Edward S. Letzter. Introduction

Linear Systems. Class 27. c 2008 Ron Buckmire. TITLE Projection Matrices and Orthogonal Diagonalization CURRENT READING Poole 5.4

Semidefinite and Second Order Cone Programming Seminar Fall 2001 Lecture 2

Def. A topological space X is disconnected if it admits a non-trivial splitting: (We ll abbreviate disjoint union of two subsets A and B meaning A B =

MAT Linear Algebra Collection of sample exams

MATH 1120 (LINEAR ALGEBRA 1), FINAL EXAM FALL 2011 SOLUTIONS TO PRACTICE VERSION

2. Every linear system with the same number of equations as unknowns has a unique solution.

Elements of Linear Algebra, Topology, and Calculus

Math 110 Linear Algebra Midterm 2 Review October 28, 2017

MATH 431: FIRST MIDTERM. Thursday, October 3, 2013.

Examples True or false: 3. Let A be a 3 3 matrix. Then there is a pattern in A with precisely 4 inversions.

Vector Spaces, Affine Spaces, and Metric Spaces

IMPORTANT DEFINITIONS AND THEOREMS REFERENCE SHEET

Lecture 23: Trace and determinants! (1) (Final lecture)

5.) For each of the given sets of vectors, determine whether or not the set spans R 3. Give reasons for your answers.

LINEAR ALGEBRA BOOT CAMP WEEK 2: LINEAR OPERATORS

Master Algèbre géométrie et théorie des nombres Final exam of differential geometry Lecture notes allowed

THE EULER CHARACTERISTIC OF A LIE GROUP

LINEAR ALGEBRA BOOT CAMP WEEK 4: THE SPECTRAL THEOREM

Topology Homework Assignment 1 Solutions

Notes on nilpotent orbits Computational Theory of Real Reductive Groups Workshop. Eric Sommers

LIE ALGEBRAS AND LIE BRACKETS OF LIE GROUPS MATRIX GROUPS QIZHEN HE

1 Smooth manifolds and Lie groups

Yale University Department of Mathematics Math 350 Introduction to Abstract Algebra Fall Midterm Exam Review Solutions

Final Exam Practice Problems Answers Math 24 Winter 2012

INTRODUCTION TO LIE ALGEBRAS. LECTURE 2.

Homework sheet 4: EIGENVALUES AND EIGENVECTORS. DIAGONALIZATION (with solutions) Year ? Why or why not? 6 9

MAT 150A, Fall 2015 Practice problems for the final exam

Eigenvalues and Eigenvectors

MATH 315 Linear Algebra Homework #1 Assigned: August 20, 2018

Linear Algebra Review. Vectors

Final Review Sheet. B = (1, 1 + 3x, 1 + x 2 ) then 2 + 3x + 6x 2

Transcription:

Differential Topology Final Exam With Solutions Instructor: W. D. Gillam Date: Friday, May 20, 2016, 13:00 (1) Let X be a subset of R n, Y a subset of R m. Give the definitions of... (a) smooth function f : X R, (b) diffeomorphism g : X Y, and (c) X is an m-manifold. Solution: (a) f : X R is smooth iff, for each x X, there is an open neighborhood U of x in R n and a smooth function F : U R such that F U X = f U X. It does not make any sense to say all partial derivatives of f exist and are continuous since, for general X R n (even for X a linear subspace of R n ), one cannot even meaningfully define the partial derivatives of a function f : X R. (This point is discussed on your first HW.) (2) Throughout, manifold means (smooth) manifold, as defined in our class, as in 1(c). (a) Let X R 2 be the graph of the function f : R R defined by f(x) = 3 x. Is X a manifold? Why or why not? (b) Let X R 3 be the graph of the function f : R 2 R defined by f(x, y) = x 2 + y 2. Is X a manifold? Why or why not? Solution: (a) The function g : R X R 2 defined by g(y) = (y 3, y) is clearly a smooth bijection with smooth inverse (x, y) y, hence X is indeed a smooth manifold, diffeomorphic to R. Note, however, that the function f is not smooth, even though it is a homeomorphism. (b) Suppose X is a manifold. Since X is connected and is clearly a 2- dimensional manifold away from 0, it must be a 2-manifold, so T 0 X must be a two-dimensional linear subspace of T 0 R 3 = R 3. On the other hand, for any a, b R, the (linear!) map g : R 0 X R 3 defined by g(t) := (at, bt, a 2 + b 2 t) has derivative g (0) = (a, b, a 2 + b 2 ) R 3 at t = 0. (Notice that we make use of the fact that there is a meaningful theory of tangent spaces for manifolds with boundary.) Since g(0) = 0, this derivative must lie in T 0 X. But as we range over all a, b (or even if just look at (a, b) equal to (1, 0), (0, 1), and (1, 1)), the vectors g (0) span R 3, so T 0 X can t be 2-dimensional after all!

2 (3) (a) State the classification theorem for 1-manifolds and the important corollary of this theorem that was central to our development of intersection theory mod 2. (b) Give the statements of three (3) theorems we proved using the above classification theorem and / or our study of intersection theory (mod 2 or with integral coefficients). (c) Choose one of theorems you mentioned in (b) and give a brief (six sentences maximum) sketch of the way we proved that theorem using the methods of differential topology. Try to give precise definitions of any concepts central to the proof. Also try to give precise statements of any intermediate results (Lemmas/Steps). Solution: (a) The classification theorem I had in mind says that every compact, connected, non-empty 1-manifold (possibly with boundary) is diffeomorphic to S 1 or [0, 1]. The important consequence is that every compact 1-manifold has an even number of boundary points. (b) The theorems I had in mind are: the Brouwer Fixed Point Theorem, the Jordan-Brouwer Separation Theorem, the Fundamental Theorem of Algebra (or at least half of it), the Borsuk-Ulam Theorem, the Hairy Ball Theorem (and / or other results about vector fields on spheres or real projective spaces), and the Lusternik-Schnirelmann Theorem. (4) Let E = R n2 be the vector space of n n matrices, GL n (R) E the open subspace consisting of invertible matrices. (a) Recall from whatever linear algebra class you took that, for any A E, you can define exp(a) GL n (R) E via the usual power series definition of exponentiation. The usual arguments one makes to justify term-by-term differentiation of convergent power series show that exp : E GL n (R) is a smooth map of manifolds. Show that there is a neighborhood V of I GL n (R) and a neighborhood U of 0 in exp 1 (V ) such that exp U : U V is a diffeomorphism. (It is not so easy to describe U and V explicitly. The map exp : E GL n (R) is neither injective (when n > 1) nor surjective (even onto the connected component of I when n > 1).) (b) Calculate the derivative of det : GL n (R) R at I GL n (R). Solution: (a) The derivative of exp at 0 is given by exp(0 + ɛb) exp(0) d exp I (B) = lim ɛb + (1/2)ɛ 2 B 2 + = lim = B.

3 In other words, d exp I is the identity map from T 0 E = E to T I GL n (R) = E, so the result follows from the Implicit Function Theorem. (b) For B T I GL n (R), we have Recall that d det(b) = lim I det(a) = σ ɛ 0 det(i + ɛb) det I. ɛ A 1,σ(1) A n,σ(n) sign(σ), where the sum is over permutations σ of {1,..., n}. When A = I + ɛb, all terms in this sum corresponding to nontrivial σ are divisible by ɛ 2 and hence will not contribute to the limit defining d det I (B). The summand corresponding to the trivial permutation σ = Id is (1 + ɛb 1,1 ) (1 + ɛb n,n ) = 1 + ɛ i B i,i +..., where the... are terms divisible by ɛ 2. We find d det (B) = I i B i,i = Tr(B). (5) Let E be a Euclidean space in the abstract sense i.e. a (finite dimensional, real) vector space E equipped with a (positive definite) inner product,. Let X E be a manifold, a E. Define f : X R by f(x) := x a 2, where e := e, e is the norm associated to the inner product on V. (a) For x X, describe the derivative df x : T x X R. (b) Use the result of (a) to show that x X is a critical point of f iff x a (T x X). (It can be shown that f is Morse for almost every a E.) (c) We now specialize to the case where E is the space of n n matrices equipped with the positive definite inner product A, B := Tr(A T B) and X E is the orthogonal group O(n) = {A E : A T A = I}. Here A T is the transpose of A and Tr denotes the trace (=sum of the diagonal entries). Recall (from your homework and / or Pages 22-23 in the textbook) that Show that T A O(n) = {B E : BA T = AB T }. (T A O(n)) = {C E : CA T = AC T }. Hint: Denote the RHS by V. First show that V (T A O(n)) by using standard properties of the trace (conjugation invariance, transpose

4 invariance, and the fact that Tr(UV ) = Tr(V U)). Next notice that for any D E we can certainly write D = 1 2 (D ADT A) + 1 2 (D + ADT A). Verify that the first summand is in T A O(n) and the second is in V, then conclude that the previously-established containment V (T A O(n)) must in fact be an equality. (d) Let λ 1,..., λ n be non-zero real numbers such that the n real numbers λ 2 1,..., λ 2 n are distinct. Set a := Diag(λ 1,..., λ n ) E. Show that if A, a 1 Aa O(n), then A is diagonal. Hint: The arguments you will use are standard tricks for studying commuting diagonalizable matrices. First show that if A, a 1 Aa O(n), then we have Aa 2 = a 2 A. Next notice that the way the λ i are chosen ensures that the standard basis vector e i is a basis for the λ 2 i eigenspace of a 2. Now show that Ae i is in this eigenspace. (e) Let a E be as in (d). Show that the critical points of the function f : O(n) R f(a) := A a 2 = Tr((A a) T (A a)) are precisely the 2 n diagonal matrices in O(n). (It can be shown that this f is Morse.) Solution: (a) We calculate df x (v) = lim ɛ 0 f(x + ɛv) f(x) ɛ x + ɛv a, x + ɛv a x a, x a = lim ɛ v, x a + ɛ x a, v + ɛ 2 v, v = lim = 2 x a, v. (b) To say that x X is a critical point of f is to say that df x (v) = 2 x a, v = 0 for every v T x X, which is to say that x a (T x X).

5 (c) Suppose B T A O(n) and C V. Then we compute B, C = Tr(B T C) = Tr(AB T CA 1 ) (Tr(SUS 1 ) = Tr(U)) = Tr(AB T CA T ) (A O(n)) = Tr( BA T CA T ) (B T A O(n)) = Tr(BA T CA T ) (Tr is linear) = Tr(BA T AC T ) (C V ) = Tr(BC T ) (A O(n)) = Tr(CB T ) (Tr(U) = Tr(U T )) = Tr(B T C) (Tr(UV ) = Tr(V U)) = B, C, hence B, C = 0. This proves V (T A O(n)). Next, for arbitrary D E, we compute and (D AD T A)A T = DA T AD T = A(D T A T DA T ) = A(D AD T A) T (D + AD T A)A T = DA T + AD T which shows that, in the decomposition = A(D T + A T DA T ) = A(D + AD T A) T, D = 1 2 (D ADT A) + 1 2 (D + ADT A), the first summand is in T A O(n) and the second is in V. In particular, this shows that T A O(n) and V together span E, so the containment V (T A O(n)) must be an equality on dimension grounds. (d) If a 1 Aa O(n), then, using the fact that a is symmetric, we compute I = (a 1 Aa)(a 1 Aa) T = a 1 Aaa T A T (a 1 ) T = a 1 Aa 2 A T a 1. Multiplying on the left and right by a shows a 2 = Aa 2 A T, then multiplying on the right by A and using A O(n) we find a 2 A = Aa 2. Since e i is an eigenvector for a 2 with eigenvalue λ 2 i, we have Aa 2 e i = λ 2 i Ae i. On the other hand, we have Aa 2 = a 2 A, so we find a 2 Ae i = λ 2 i Ae i, which shows that Ae i is in the λ 2 i eigenspace of a 2. But a 2 = Diag(λ 2 1,..., λ 2 n) is diagonal with distinct diagonal entries, so e i is a basis for the λ 2 i eigenspace of a 2, so we must have Ae i = µ i e i for some µ i R. This is true for each i, so we find that

6 A = Diag(µ 1,..., µ n ). Since A O(n), the µ i must be ±1 (the columns of A must have norm 1). (e) By (b), A O(n) is a critical point of f iff A a (T A O(n)). By the description of (T A O(n)) in (c), this is equivalent to which is equivalent to which is equivalent to (A a)a T = A(A a) T, aa T = Aa, A T = a 1 Aa. Obviously this holds when A is diagonal (for then A and a are both diagonal, so they commute and A = A T ) and (d) implies that it holds only when A is diagonal.