Math 535a Homework 5

Similar documents
Topological K-theory, Lecture 2

LECTURE 1: SOME GENERALITIES; 1 DIMENSIONAL EXAMPLES

Math 396. Linear algebra operations on vector bundles

(1) Let π Ui : U i R k U i be the natural projection. Then π π 1 (U i ) = π i τ i. In other words, we have the following commutative diagram: U i R k

Mini-Course on Moduli Spaces

M4P52 Manifolds, 2016 Problem Sheet 1

Chapter 2. Bundles. 2.1 Vector bundles: definitions and examples

LECTURE 3 MATH 261A. Office hours are now settled to be after class on Thursdays from 12 : 30 2 in Evans 815, or still by appointment.

LECTURE 28: VECTOR BUNDLES AND FIBER BUNDLES

FOUNDATIONS OF ALGEBRAIC GEOMETRY CLASS 2

Categories and functors

h M (T ). The natural isomorphism η : M h M determines an element U = η 1

COMMUTATIVE ALGEBRA LECTURE 1: SOME CATEGORY THEORY

Exercises on chapter 0

MONDAY - TALK 4 ALGEBRAIC STRUCTURE ON COHOMOLOGY

Categories and Quantum Informatics: Hilbert spaces

SECTION 2: THE COMPACT-OPEN TOPOLOGY AND LOOP SPACES

Equivalence of the Combinatorial Definition (Lecture 11)

FOUNDATIONS OF ALGEBRAIC GEOMETRY CLASS 24

EXAMPLES AND EXERCISES IN BASIC CATEGORY THEORY

Math 121 Homework 4: Notes on Selected Problems

3. Categories and Functors We recall the definition of a category: Definition 3.1. A category C is the data of two collections. The first collection

Math 215B: Solutions 3

Cohomology and Vector Bundles

The Algebra of Tensors; Tensors on a Vector Space Definition. Suppose V 1,,V k and W are vector spaces. A map. F : V 1 V k

EXERCISES IN POISSON GEOMETRY

PICARD GROUPS OF MODULI PROBLEMS II

1. Algebraic vector bundles. Affine Varieties

Differential geometry for physicists - Lecture 2

An introduction to cobordism

LECTURE 2. (TEXED): IN CLASS: PROBABLY LECTURE 3. MANIFOLDS 1. FALL TANGENT VECTORS.

Elementary (ha-ha) Aspects of Topos Theory

MILNOR SEMINAR: DIFFERENTIAL FORMS AND CHERN CLASSES

MATH 101A: ALGEBRA I PART C: TENSOR PRODUCT AND MULTILINEAR ALGEBRA. This is the title page for the notes on tensor products and multilinear algebra.

BEN KNUDSEN. Conf k (f) Conf k (Y )

Exercises of the Algebraic Geometry course held by Prof. Ugo Bruzzo. Alex Massarenti

Lecture 9: Sheaves. February 11, 2018

1 Differentiable manifolds and smooth maps

MATH 4030 Differential Geometry Lecture Notes Part 4 last revised on December 4, Elementary tensor calculus

Vector Bundles on Algebraic Varieties

Chern classes à la Grothendieck

Topological K-theory, Lecture 3

Assume the left square is a pushout. Then the right square is a pushout if and only if the big rectangle is.

Algebraic Geometry

and the diagram commutes, which means that every fibre E p is mapped to {p} V.

Patrick Iglesias-Zemmour

Topological K-theory

Direct Limits. Mathematics 683, Fall 2013

Differential Geometry MTG 6257 Spring 2018 Problem Set 4 Due-date: Wednesday, 4/25/18

is the desired collar neighbourhood. Corollary Suppose M1 n, M 2 and f : N1 n 1 N2 n 1 is a diffeomorphism between some connected components

Exercises in Geometry II University of Bonn, Summer semester 2015 Professor: Prof. Christian Blohmann Assistant: Saskia Voss Sheet 1

ALGEBRAIC GEOMETRY COURSE NOTES, LECTURE 9: SCHEMES AND THEIR MODULES.

Vector bundles in Algebraic Geometry Enrique Arrondo. 1. The notion of vector bundle

THE MODULI OF SUBALGEBRAS OF THE FULL MATRIX RING OF DEGREE 3

Characteristic Classes, Chern Classes and Applications to Intersection Theory

Algebraic Geometry Spring 2009

Adjunctions! Everywhere!

ERRATA FOR INTRODUCTION TO SYMPLECTIC TOPOLOGY

Geometry 9: Serre-Swan theorem

Integration and Manifolds

SUMMER COURSE IN MOTIVIC HOMOTOPY THEORY

Representations and Linear Actions

Holomorphic line bundles

HOMEWORK FOR SPRING 2014 ALGEBRAIC TOPOLOGY

NOTES ON DIFFERENTIAL FORMS. PART 3: TENSORS

Cutting and pasting. 2 in R. 3 which are not even topologically

NOTES ON FIBER BUNDLES

4. Symmetric and Alternating Products We want to introduce some variations on the theme of tensor products.

We can choose generators of this k-algebra: s i H 0 (X, L r i. H 0 (X, L mr )

Vector spaces, duals and endomorphisms

CATEGORICAL GROTHENDIECK RINGS AND PICARD GROUPS. Contents. 1. The ring K(R) and the group Pic(R)

De Rham Cohomology. Smooth singular cochains. (Hatcher, 2.1)

Systems of linear equations. We start with some linear algebra. Let K be a field. We consider a system of linear homogeneous equations over K,

1 Replete topoi. X = Shv proét (X) X is locally weakly contractible (next lecture) X is replete. D(X ) is left complete. K D(X ) we have R lim

1. Classifying Spaces. Classifying Spaces

A GLIMPSE OF ALGEBRAIC K-THEORY: Eric M. Friedlander

Equivalence Relations

Category theory and set theory: algebraic set theory as an example of their interaction

Math 231b Lecture 16. G. Quick

NOTES ON ATIYAH S TQFT S

Lectures - XXIII and XXIV Coproducts and Pushouts

which is a group homomorphism, such that if W V U, then

Symplectic and Poisson Manifolds

Lecture 17: Invertible Topological Quantum Field Theories

Derived Algebraic Geometry I: Stable -Categories

ON COSTELLO S CONSTRUCTION OF THE WITTEN GENUS: L SPACES AND DG-MANIFOLDS

THE H-PRINCIPLE, LECTURE 14: HAEFLIGER S THEOREM CLASSIFYING FOLIATIONS ON OPEN MANIFOLDS

Three Descriptions of the Cohomology of Bun G (X) (Lecture 4)

CATEGORY THEORY. Cats have been around for 70 years. Eilenberg + Mac Lane =. Cats are about building bridges between different parts of maths.

ELEMENTARY SUBALGEBRAS OF RESTRICTED LIE ALGEBRAS

Math 396. Quotient spaces

MATH 101B: ALGEBRA II PART A: HOMOLOGICAL ALGEBRA 23

PART II.1. IND-COHERENT SHEAVES ON SCHEMES

1 Recall. Algebraic Groups Seminar Andrei Frimu Talk 4: Cartier Duality

sset(x, Y ) n = sset(x [n], Y ).

ABSTRACT DIFFERENTIAL GEOMETRY VIA SHEAF THEORY

An Introduction to the Stolz-Teichner Program

Fibre Bundles and Chern-Weil Theory. Johan Dupont

Math 52H: Multilinear algebra, differential forms and Stokes theorem. Yakov Eliashberg

1 Fields and vector spaces

Transcription:

Math 535a Homework 5 Due Monday, March 20, 2017 by 5 pm Please remember to write down your name on your assignment. 1. Let (E, π E ) and (F, π F ) be (smooth) vector bundles over a common base M. A vector bundle morphism over M from E to F is a C map f : E F, compatible with projection, meaning that π F f = π E, so that on each fiber f p : E p F p is a linear map of vector spaces. An isomorphism of vector bundles E = F over potentially different manifolds M and N is a diffeomorphism f : M N and a vector bundle morphism covering f which is an isomorphism on each fiber. 1 We say E and F are isomorphic over M if the isomorphism of vector bundles covers the identity map; e.g, if the vector bundle morphism f : E F satisfies π F f = π E, so f maps E p to F p. 2 Let (E, π : E M) be a rank k vector bundle. We say a collection of sections s 1,..., s k Γ(E) is linearly independent if (s 1 )(p),..., (s k )(p) are linearly independent in E p for each p M (in particular, no s i (p) should be zero). Prove that there is an isomorphism of vector bundles over M, E = R k if and only if E has a basis of linearly independent sections s 1,..., s k. (Recall: R k denotes the trivial rank k bundle over M, defined as R k := M R k, with projection map π : M R k M sending (p, v) to p). 2. (double weight) Vector bundles via gluing data. Let M be a smooth manifold and U = {U α α I an open cover of M. A set of transition (or gluing) data (of rank k) for the cover U is a collection of C maps satisfying Φ αβ : U α U β GL(k, R) for all α, β I i. For any α, β I, Φ αβ Φ βα = I. 3 ii. (cocycle condition) for any α, β, γ I, Φ γα = Φ γβ Φ βα as functions on U α U β U γ. (a) Show, as asserted in class, that, given a collection of transition data T := {Φ αβ α,β I for the cover U, there is a C rank k vector bundle E U,T over M, formed by gluing 1 If f : M N is a C map, and E, π E : E M, F : π F : F N are vector bundles, a (general) bundle morphism covering f is a C map f : E F satisfying π F f = f π E, so that the induced map on fibers E p F f(p) is a linear map. 2 An implicit point is that isomorphism of bundles over M is an equivalence relation. In particular, you should check that if there is an isomorphism of vector bundles f : E F, then f is a diffeomorphism with f 1 : F E also an isomorphism of vector bundles. You do not need to prove this fact as part of your HW, but if would like to use this fact on the HW elsewhere, you should prove it as a Lemma. 3 Here and in ii., multiplication means pointwise multiplication of matrices, i.e., for f, g : V GL(k, R), (f g)(p) = f(p) g(p). Also, I means the constant function with value the identity matrix; V GL(k, R), p I. 1

together trivial vector bundles U α R k via the transition data maps: E U,T := (U α R k )/{(U α, p, v) (U β, p, Φ αβ (p)(v)) for all α, β, p U α U β. U α U with projection map to M defined as follows: first, define M U = U α /{(U α, p) (U β, p) for all α, β, p U α U β ; U α U check that M U is indeed a smooth manifold, and show that there is a canonical diffeomorphism i U : M U M sending [(U α, p)] to p. Then one defines π : E U,T M as the composition of the map π T : E U,T M U with i U, where π T : [(U α, p, v)] [(U α, p)] (again, you ll need to check this is C ). 4 Note: part of this problem involves verifying that (a) of M U as above is indeed a smooth manifold canonically diffeomorphic to M, and (b) E U,T has the structure of a smooth manifold with π a smooth map, and finally that (c) (E U,T, π) satisfy the axioms of a vector bundle. (b) Show that any vector bundle E on M is isomorphic, as vector bundles over M (in the sense of problem 1), to E U,T for some cover U and some transition data T. (c) Let U = {U i i I be a cover and T = {Φ αβ a set of transition data of rank k. A (Čech) 0-co-chain on U is a collection of functions F := {f α : U α GL(k, R). Given such a Čech 0-co-chain, define new transition data for the cover U by FT := {f 1 α Φ αβ f β. Show that E U,T and E U,FT are isomorphic as vector bundles over M. (d) Given an open cover of M, U = {U α α I, along with transition data T for U, a refinement of the cover U is a new open cover U = {V κ κ J along with a map of index sets f : J I such that for each κ J, V κ U f(κ). (Example: If M = R n, one can take U = {R n and U = {B ɛ (p) p J=R n. Then U, equipped with the canonical map on index sets f : J {, is a refinement of U). Given a refinement of U, call it U, the induced transition data on the refinement is T U := {Φ f(κ)f(τ) Vκ V τ κ,τ J. Show that there is an isomorphism of vector bundles over M, E U,T U = EU,T. 4 Note: In this entire problem, for clarity we re using the notation V V V := {(V, p) V V, p V. So elements of disjoint unions should be thought of as tuples (V, p) where V is one of the sets we re taking the disjoint union of and p is an element of V. 2

(e) Let V ect k (M) denote the set of smooth rank k vector bundles on M up to isomorphism. That is, V ect k (M) := {(E, π) a vector bundle over M/, where E F if E and F are isomorphic over M, in the terminology of problem 1. Show that there is a bijection of sets V ect k (M) = {(U, T) U is any open cover of M and T is any transition data of rank k for U/, where (U 1, T 1 ) (U 2, T 2 ) if, after passing to a common refinement U of U 1 and U 2, there is a Čech 0-co-chain F on U such that T 1 U = F(T 2 U ). Hint: there is a natural map {(U, T) as above V ect k (M) sending (U, T) to (E U,T, π). Show that this descends to a well-defined map {(U, T) as above/ V ect k (M) which is in fact a bijection. 3. (double weight) The category theory of vector bundle operations: In class, we have indicated or sketched how many vector space operations, such as direct sum, tensor product, self tensor/exterior product, quotient, etc. induce operations on vector bundles. In this homework assignment we will give a streamlined proof of all such constructions, using some category theory. First, if C 1,..., C k are categories, the product category C 1 C k has as objects elements of the Cartesian product (ob C 1 ) (ob C k ), and has morphisms, for a pair of tuples (X 1,..., X k ), (Y 1,..., Y k ), hom C1 C k ((X 1,..., X k ), (Y 1,..., Y k )) := hom C1 (X 1, Y 1 ) hom Ck (X k, Y k ), with composition given by component-wise composition (see exercise (a) below). We use the shorthand C k := C C A k-ary functor from C to D is a functor F : C k D. Now, let V := V ect iso R be the category whose objects are finite-dimensional R-linear vector spaces and whose morphisms are the R-linear isomorphisms/invertible linear transformations, with composition given by composition of linear maps (you should convince yourself this is a category). Note that in this category hom V (V, V ) = GL(V ). (a) Check that the product category C 1 C k is indeed a category if C 1,..., C k are. (b) Show that in V, hom V (V, W ) has the structure of a C manifold for any pair of finite-dimensional vector spaces V, W. Moreover, check that the composition map hom V (W, Z) hom V (V, W ) hom V (V, Z) is a C map of manifolds. (we have already seen this when V = W = Z) (in categorical language, we might say that V is 3

enriched in the category of smooth manifolds. ) (c) We say a k-ary functor T : V V is smooth if, for any pair of objects (V 1,..., V k ), (W 1,..., W k ) in V k, the the induced map on morphism spaces hom V (V 1, W 1 ) hom V (V k, W k ) = hom V k((v 1,..., V k ), (V 1,..., V k )) is a C map of manifolds. hom V (T (V 1,..., V k ), T (W 1,..., W k )) Show that the following give smooth k-ary functors on V. Note: you will in particular need to define a map on morphism spaces associated to the map on objects we define below. 2-ary functor T, defined on objects by T : (V, W ) V W. 2-ary functor T, defined on objects by T : (V, W ) V W. 1-ary functor T k, defined on objects by T k : V V k for some k. 1-ary functor T k, defined on objects by T k : V k V for some k. (d) Given a tuple of (smooth) vector bundles E 1,..., E k over a manifold M, and a smooth k-ary functor T as above, directly construct a smooth vector bundle T (E 1,..., E k ) with T (E 1,..., E k ) p = T ((E 1 ) p,..., (E k ) p ). Hint: This problem is very similar to the problem of constructing T M. First, as a set T (E 1,..., E k ) = {(p, v) p M, v T ((E 1 ) p,..., (E k ) p ). Next, you will need to equip this set with a topology (use a cover on M consisting of open sets for which the input vector bundles all simultaneously admit trivializations), and manifold structure, such that the projection map to M is C. Finally, you should argue that the resulting pair (T (E 1,..., E k ), π) is indeed a smooth vector bundle. (e) Give a construction of the vector bundle T (E 1,..., E k ) in terms of transition data (as in the previous problem), using as input the transition data associated to E 1,..., E k for some cover, and explain why the result is isomorphic (as vector bundles) to the construction above. Note: Now, if E, F are vector bundles, we have now defined vector bundles E F := T (E, F), E F := T (E, F), and so on for E k, and k E. Remark: A very similar argument leads to the construction of a vector bundle hom(e, F), whose fiber at a point p M is hom R (E p, F p ). (a special case of which gives a construction of E := hom(e, R)). However, the argument you ll give in this part doesn t verbatim apply because the functor sending (V, W ) hom R (V, W ) is contravariant in in the first entry V, (meaning isomorphisms V 1 V 2 induce, via pre-composition, maps in the opposite direction hom R (V 2, W ) hom R (V 1, W )). 4

Hence T hom := hom R (, ) should be thought of as a functor V op V V. Fortunately, there is still a (basically identical) notion for smoothness of functors T : V op op V V. The arguments of this section directly adapt over s times t times to give a construction of a vector bundle T (E 1,..., E s ; F 1,..., F t ) for such T. 4. (a) Let E = [0, 1] R/(0, t) (1, t) be the Möbius line bundle defined in class, with π : E S 1 = [0, 1]/0 1 sending (x, t) x. Verify that E is indeed a line bundle, and prove that E is not isomorphic to the trivial line bundle. (b) Let L = {(x, v) RP n R n+1 v x, 5 π : L RP n, (x, v) x denote the line bunde introduced in class; we call this bundle the tautological line bundle on RP n. Verify that L is indeed a line bundle. 5. Linear algebra of tensor products (a) Write in detail the construction of the canonical map V W α hom(v, W ), and give a careful proof that it is an isomorphism if V and W are finite dimensional (in class, we constructed the map and sketchily argued why it was surjective). (b) Let ev : V V R be the linear map induced by the bilinear map ev : V V R, (φ, v) φ(v). Given a linear operator T hom(v, V ) on a finite-dimensional vector space, define the trace of T as tr(t ) := ev(α 1 (T )), where α is the map defined in the previous section. Show that this definition agrees with the usual notion of trace, that is if v := {v 1,..., v k is any basis of V and A is the matrix of T with respect to v, then tr(t ) = tr(a) = i a ii. 6. Exterior algebra 1. (a) A formula for the determinant of 3 3 matrices. Recall from class that the determinant det(t ) of T hom R (V, V ) is defined as the scalar in R such that where v 1,..., v n is any basis for V. T (v 1 ) T (v n ) = det(t ) v 1 v n, Suppose that dim V = 3, and v = (v 1, v 2, v 3 ) is a basis for V. Let T : V V be the linear operator defined by T (v 1 ) = av 1 + dv 2 + gv 3 T (v 2 ) = bv 1 + ev 2 + hv 3 T (v 3 ) = cv 1 + fv 2 + iv 3. 5 By v x, we mean that, if x = [w], then v Span(w). 5

In other words, suppose the matrix of T with respect to v is M(T, v) = a b c d e f g h i Derive, using the definition we gave in class with exterior products, a formula for det(t ) in terms of a, b, c, d, e, f, h, and i. (Remark: this formula should have six terms. It is easy to find the formula online, but you should be able to derive it in terms of the definition of determinants given in lecture). 7. Exterior algebra 2. For the below problems, let V be a finite-dimensional vector space over R. (a) Let A k (V ) := AltMultiLin R (V, R) be the vector space of alternating multilinear maps from k copies of V to R (what is its vector space structure). Also let L k (V ) := MultiLin R (V, R) is the vector space of multilinear maps. Prove that there are canonical isomorphisms A k (V ) = k V = ( k V ). Similarly, prove that there is a canonical isomorphism L k (V ) = (V ) k = (V k ), and that under these inclusions, the natural map A k (V ) L k (V ) is sent to the (dual of) the projection map (V ) k k V. (here we are implicitly using the fact that (V k ) = (V ) k and similarly for the wedge product). (b) An element ω A 2 (V ) = 2 V is called non-degenerate, or a linear symplectic form, if ω(v, ) 0 hom R (V, R) for any non-zero v V. If V is finite-dimensional and V admits a linear symplectic form, prove that n = dim V is necessarily even, say n = 2m. (c) Prove that ω 2 V is non-degenerate if and only if ω m 0 in n V (where n = 2m). 8. Give a careful construction of the exterior differentiation operator d : Ω k (M) Ω k+1 (M) using local coordinates; show that this definition is independent of local coordinates and is well-defined. 9. Let M be a manifold. Prove that d satisfies the formula d(α β) = d(α) β+( 1) k α d(β), where α Ω k (M), β Ω l (M). 10. Prove that d commutes with pullback; that is, d f = f d for any smooth f : M N. 6