An Introduction to Sheaf Cohomology in Algebraic Geometry

Total Page:16

File Type:pdf, Size:1020Kb

An Introduction to Sheaf Cohomology in Algebraic Geometry An Introduction to Sheaf Cohomology in Algebraic Geometry Ryan Mike University of Colorado Boulder Department of Mathematics Defended April 7, 2020 Thesis Advisor: Dr. Sebastian Casalaina-Martin, Department of Mathematics Defense Committee: Dr. Sebastian Casalaina-Martin, Department of Mathematics Dr. Nathaniel Thiem, Department of Mathematics Dr. Rafael Frongillo, Department of Computer Science Contents 0 Introduction and Preliminaries 1 1 Homological Algebra in R-Mod 2 1.1 (Co)chain Complexes . .2 1.2 Injective and Projective Modules . .5 1.3 Resolutions . .7 1.4 Derived Functors . 10 2 Sheaves 12 2.1 Complex Manifolds . 13 2.2 Basic Definitions and Examples . 14 2.3 Homological Algebra in Sh(X).............................. 17 2.4 Cohomology of Sheaves . 19 3 Applications 20 3.1 Relation to Other Cohomology Theories . 20 3.2 The Exponential Sequence and the Complex Logarithm . 22 3.3 The Riemann-Roch Theorem for Line Bundles . 23 0 Introduction and Preliminaries Given a topological space X and an abelian group A, the singular cohomology groups Hn(X; A) provide important algebraic invariants of X which form obstructions to X being contractible. Sheaf cohomology may be thought of as a generalization of singular cohomology (and other cohomology theories) in the sense that it accommodates more general \coefficients" than simply an abelian group A. The abelian group A is replaced by a sheaf of abelian groups F which in turn gives rise to the sheaf cohomology groups Hn(X; F). Broadly speaking, these groups constitute obstructions to solving geometric problems globally, given that they can be solved locally. In this paper, we intend to give an introduction to sheaf cohomology via the derived functors approach of Grothendieck. This paper is expository in nature, and the interested reader can find more information about all of the topics we cover in the list of references we provide at the end. Section 1 covers all of the homological algebra needed to define sheaf cohomology. Assuming familiarity with R-modules, we begin with the basics of cochain complexes, exactness, and coho- mology. This brings us to a discussion of exact functors and injective resolutions which then lets us define derived functors. Section 2 is devoted to developing the theory of sheaves on a topological space. We begin with an informal review of complex manifolds and line bundles. This leads to some examples of sheaves later on, and gives an impression of the types of geometric objects which can be studied using sheaves and cohomology. We then provide basic definitions and examples of sheaves and address how the tools from homological algebra which we developed in section 1 can be transferred to the category of sheaves on a topological space. Finally, we define the sheaf cohomology functors as the derived functors of the global sections functor from the category of sheaves on a topological space to the category of abelian groups. In section 3, we attempt to give the reader a sense for how sheaf cohomology is used in algebraic geometry and more broadly. First, we examine how sheaf cohomology fits within the larger picture 1 of cohomology theories, and in particular, its relation to singular cohomology and de Rham coho- mology. We then introduce the exponential sequence and explore its cohomological consequences to find some interesting properties of the complex logarithm from complex analysis. Finally, we discuss Chern classes and the Euler-Poincar´echaracteristic of a sheaf which lets state the Riemann-Roch theorem for line bundles on curves. Before we begin, we will review a couple preliminaries to set-up the category of R-modules. For our purposes, R will be a commutative ring so we will not distinguish between left and right R-modules. Definition 0.1. Let R be a commutative ring with 1. An R-module is an abelian group M together with an action of R on M (that is, a map R × M ! M) denoted rm for all r 2 R and m 2 M, such that • r(m + n) = rm + rn for all r 2 R and m; n 2 M. • (r + s)m = rm + sm for all r; s 2 R and m 2 M, • r(sm) = (rs)m for all r; s 2 R and m 2 M, and • 1m = m for all m 2 M. Given two R-modules M and N, we say a function φ : M ! N is an R-module homomorphism if it is a homomorphism of the underlying abelian groups which additionally satisfies φ(rm) = rφ(m) for all r 2 R and m 2 M. Note that if R = Z, the ring of integers, then the definition of R-module collapses to the definition of abelian group. Thus Z-modules are abelian groups. There is a category whose objects are R-modules and whose morphisms are R-module homo- morphisms. This is the category of R-modules, denoted R-Mod. For convenience, we will often call a homomorphism of R-modules an R-module map, or just a map when the context is understood. 1 Homological Algebra in R-Mod The purpose of this section is to establish the algebraic preliminaries needed to develop sheaf cohomology. Homological algebra unfolds in the abstract setting of abelian categories. Loosely speaking, these are categories equipped with sensible notions of product, coproduct, zero object, kernel, cokernel, and image, in which the sets of morphisms are enriched with an abelian group structure. General abelian categories can look very foreign at first glance, making results more difficult to state and prove. For this reason, we will restrict attention to the category R-Mod, where we assume our reader will have more familiarity. This is the prototypical example of an abelian category, and a good place to develop intuition and understanding for homological algebra. We emphasize, however, that all of the key results we prove extend to the more general setting of abelian categories. Indeed, this will be an important point when we move to the category of sheaves on a topological space in the next section. For the remainder of this section, we will fix a commutative ring R with 1. 1.1 (Co)chain Complexes In many areas of mathematics, objects often appear not by themselves, but in sequences with other objects of the same type. Given a collection of R-modules M1;M2;:::;Mn and R-module maps 2 fi : Mi ! Mi+1 for 1 ≤ i < n, we can form the sequence f1 f2 f3 fn−1 M1 M2 M3 ··· Mn: One curiosity about such a sequence is how the image of the incoming map fi−1 interacts with the kernel of the outgoing map fi in the module Mi. We say the sequence is exact at Mi if these submodules are equal, that is, if ker fi = im fi−1. The sequence is called exact if it is exact at Mi for 1 < i < n. Exactness also tells us something about the composite maps fi ◦ fi−1. In particular, exactness at Mi implies fi ◦ fi−1 = 0. Our next definition singles out precisely this case. Definition 1.1. A chain complex C• of R-modules is a family fCngn2Z of R-modules together with R-module maps fdn : Cn ! Cn−1gn2Z called differentials satisfying the property that dn ◦dn+1 = 0 for all n 2 Z. Because dn ◦ dn+1 = 0 we have 0 ⊆ im dn+1 ⊆ ker dn ⊆ Cn th and we define the n homology module of Cn to be the quotient Hn(C•) := ker dn= im dn+1. It is common to depict a chain complex C• by the diagram d d d d ··· Cn+1 Cn Cn−1 ··· and omit indices on the differentials for convenience. A chain map f : C• ! D• is family of R-module maps ffn : Cn ! Dngn2Z which commute with the differentials, that is, so that following diagram commutes d d d d ··· Cn+1 Cn Cn−1 ··· fn+1 fn fn−1 d d d d ··· Dn+1 Dn Dn−1 ··· : Note that if C• is exact at Cn, then Hn(C•) = 0. In this sense, we can view the homology modules Hn(C•) as obstructions to C• being exact. Examples 1.1. • The sequence consisting of all zero modules and zero differentials ··· 0 0 0 ··· is a chain complex which we will call the zero complex and often simply denote by 0. • More generally, any arbitrary family of R-modules fCng with zero differentials is a chain complex. • The sequence ·2 ·2 ·2 ·2 ··· Z=4 Z=4 Z=4 ··· with all differentials as `times-two' maps x 7! 2x is a chain complex which is exact. Anyone familiar with some category theory will be well-acquainted with the `co-' prefix, indi- n n n n+1 cating arrow reversals. This gives a variant in notation: C = C−n and d = d−n : C ! C . 3 • n Definition 1.2. A cochain complex C of R-modules is a family fC gn2Z of R-modules together n n n+1 n n−1 th with R-module maps fd : C ! C gn2Z such that d ◦ d = 0 for all n 2 Z. The n cohomology module of Cn is the quotient Hn(C•) := ker dn= im dn−1. Maps of cochain complexes are defined in the same way as chain maps and are called cochain maps. Elements of Hn(C•) are called cohomology classes and for a representative c 2 ker dn, we use square brackets to denote its cohomology class [c] 2 Hn(C•). Note on conventions. When we develop sheaf cohomology later in this paper, we will be chiefly concerned with cochain complexes and, as the reader may have guessed, cohomology.
Recommended publications
  • Cellular Sheaf Cohomology in Polymake Arxiv:1612.09526V1
    Cellular sheaf cohomology in Polymake Lars Kastner, Kristin Shaw, Anna-Lena Winz July 9, 2018 Abstract This chapter provides a guide to our polymake extension cellularSheaves. We first define cellular sheaves on polyhedral complexes in Euclidean space, as well as cosheaves, and their (co)homologies. As motivation, we summarise some results from toric and tropical geometry linking cellular sheaf cohomologies to cohomologies of algebraic varieties. We then give an overview of the structure of the extension cellularSheaves for polymake. Finally, we illustrate the usage of the extension with examples from toric and tropical geometry. 1 Introduction n Given a polyhedral complex Π in R , a cellular sheaf (of vector spaces) on Π associates to every face of Π a vector space and to every face relation a map of the associated vector spaces (see Definition 2). Just as with usual sheaves, we can compute the cohomology of cellular sheaves (see Definition 5). The advantage is that cellular sheaf cohomology is the cohomology of a chain complex consisting of finite dimensional vector spaces (see Definition 3). Polyhedral complexes often arise in algebraic geometry. Moreover, in some cases, invariants of algebraic varieties can be recovered as cohomology of certain cellular sheaves on polyhedral complexes. Our two major classes of examples come from toric and tropical geometry and are presented in Section 2.2. We refer the reader to [9] for a guide to toric geometry and polytopes. For an introduction to tropical geometry see [4], [20]. The main motivation for this polymake extension was to implement tropical homology, as introduced by Itenberg, Katzarkov, Mikhalkin, and Zharkov in [15].
    [Show full text]
  • Derived Functors for Hom and Tensor Product: the Wrong Way to Do It
    Derived Functors for Hom and Tensor Product: The Wrong Way to do It UROP+ Final Paper, Summer 2018 Kevin Beuchot Mentor: Gurbir Dhillon Problem Proposed by: Gurbir Dhillon August 31, 2018 Abstract In this paper we study the properties of the wrong derived functors LHom and R ⊗. We will prove identities that relate these functors to the classical Ext and Tor. R With these results we will also prove that the functors LHom and ⊗ form an adjoint pair. Finally we will give some explicit examples of these functors using spectral sequences that relate them to Ext and Tor, and also show some vanishing theorems over some rings. 1 1 Introduction In this paper we will discuss derived functors. Derived functors have been used in homo- logical algebra as a tool to understand the lack of exactness of some important functors; two important examples are the derived functors of the functors Hom and Tensor Prod- uct (⊗). Their well known derived functors, whose cohomology groups are Ext and Tor, are their right and left derived functors respectively. In this paper we will work in the category R-mod of a commutative ring R (although most results are also true for non-commutative rings). In this category there are differ- ent ways to think of these derived functors. We will mainly focus in two interpretations. First, there is a way to concretely construct the groups that make a derived functor as a (co)homology. To do this we need to work in a category that has enough injectives or projectives, R-mod has both.
    [Show full text]
  • A NOTE on COMPLETE RESOLUTIONS 1. Introduction Let
    PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 138, Number 11, November 2010, Pages 3815–3820 S 0002-9939(2010)10422-7 Article electronically published on May 20, 2010 A NOTE ON COMPLETE RESOLUTIONS FOTINI DEMBEGIOTI AND OLYMPIA TALELLI (Communicated by Birge Huisgen-Zimmermann) Abstract. It is shown that the Eckmann-Shapiro Lemma holds for complete cohomology if and only if complete cohomology can be calculated using com- plete resolutions. It is also shown that for an LHF-group G the kernels in a complete resolution of a ZG-module coincide with Benson’s class of cofibrant modules. 1. Introduction Let G be a group and ZG its integral group ring. A ZG-module M is said to admit a complete resolution (F, P,n) of coincidence index n if there is an acyclic complex F = {(Fi,ϑi)| i ∈ Z} of projective modules and a projective resolution P = {(Pi,di)| i ∈ Z,i≥ 0} of M such that F and P coincide in dimensions greater than n;thatis, ϑn F : ···→Fn+1 → Fn −→ Fn−1 → ··· →F0 → F−1 → F−2 →··· dn P : ···→Pn+1 → Pn −→ Pn−1 → ··· →P0 → M → 0 A ZG-module M is said to admit a complete resolution in the strong sense if there is a complete resolution (F, P,n)withHomZG(F,Q) acyclic for every ZG-projective module Q. It was shown by Cornick and Kropholler in [7] that if M admits a complete resolution (F, P,n) in the strong sense, then ∗ ∗ F ExtZG(M,B) H (HomZG( ,B)) ∗ ∗ where ExtZG(M, ) is the P-completion of ExtZG(M, ), defined by Mislin for any group G [13] as k k−r r ExtZG(M,B) = lim S ExtZG(M,B) r>k where S−mT is the m-th left satellite of a functor T .
    [Show full text]
  • The Geometry of Syzygies
    The Geometry of Syzygies A second course in Commutative Algebra and Algebraic Geometry David Eisenbud University of California, Berkeley with the collaboration of Freddy Bonnin, Clement´ Caubel and Hel´ ene` Maugendre For a current version of this manuscript-in-progress, see www.msri.org/people/staff/de/ready.pdf Copyright David Eisenbud, 2002 ii Contents 0 Preface: Algebra and Geometry xi 0A What are syzygies? . xii 0B The Geometric Content of Syzygies . xiii 0C What does it mean to solve linear equations? . xiv 0D Experiment and Computation . xvi 0E What’s In This Book? . xvii 0F Prerequisites . xix 0G How did this book come about? . xix 0H Other Books . 1 0I Thanks . 1 0J Notation . 1 1 Free resolutions and Hilbert functions 3 1A Hilbert’s contributions . 3 1A.1 The generation of invariants . 3 1A.2 The study of syzygies . 5 1A.3 The Hilbert function becomes polynomial . 7 iii iv CONTENTS 1B Minimal free resolutions . 8 1B.1 Describing resolutions: Betti diagrams . 11 1B.2 Properties of the graded Betti numbers . 12 1B.3 The information in the Hilbert function . 13 1C Exercises . 14 2 First Examples of Free Resolutions 19 2A Monomial ideals and simplicial complexes . 19 2A.1 Syzygies of monomial ideals . 23 2A.2 Examples . 25 2A.3 Bounds on Betti numbers and proof of Hilbert’s Syzygy Theorem . 26 2B Geometry from syzygies: seven points in P3 .......... 29 2B.1 The Hilbert polynomial and function. 29 2B.2 . and other information in the resolution . 31 2C Exercises . 34 3 Points in P2 39 3A The ideal of a finite set of points .
    [Show full text]
  • Sheaves and Homotopy Theory
    SHEAVES AND HOMOTOPY THEORY DANIEL DUGGER The purpose of this note is to describe the homotopy-theoretic version of sheaf theory developed in the work of Thomason [14] and Jardine [7, 8, 9]; a few enhancements are provided here and there, but the bulk of the material should be credited to them. Their work is the foundation from which Morel and Voevodsky build their homotopy theory for schemes [12], and it is our hope that this exposition will be useful to those striving to understand that material. Our motivating examples will center on these applications to algebraic geometry. Some history: The machinery in question was invented by Thomason as the main tool in his proof of the Lichtenbaum-Quillen conjecture for Bott-periodic algebraic K-theory. He termed his constructions `hypercohomology spectra', and a detailed examination of their basic properties can be found in the first section of [14]. Jardine later showed how these ideas can be elegantly rephrased in terms of model categories (cf. [8], [9]). In this setting the hypercohomology construction is just a certain fibrant replacement functor. His papers convincingly demonstrate how many questions concerning algebraic K-theory or ´etale homotopy theory can be most naturally understood using the model category language. In this paper we set ourselves the specific task of developing some kind of homotopy theory for schemes. The hope is to demonstrate how Thomason's and Jardine's machinery can be built, step-by-step, so that it is precisely what is needed to solve the problems we encounter. The papers mentioned above all assume a familiarity with Grothendieck topologies and sheaf theory, and proceed to develop the homotopy-theoretic situation as a generalization of the classical case.
    [Show full text]
  • 81151635.Pdf
    View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Elsevier - Publisher Connector Topology and its Applications 158 (2011) 2103–2110 Contents lists available at ScienceDirect Topology and its Applications www.elsevier.com/locate/topol The higher derived functors of the primitive element functor of quasitoric manifolds ∗ David Allen a, , Jose La Luz b a Department of Mathematics, Iona College, New Rochelle, NY 10801, United States b Department of Mathematics, University of Puerto Rico in Bayamón, Industrial Minillas 170 Car 174, Bayamón 00959-1919, Puerto Rico article info abstract Article history: Let P be an n-dimensional, q 1 neighborly simple convex polytope and let M2n(λ) be the Received 15 June 2011 corresponding quasitoric manifold. The manifold depends on a particular map of lattices Accepted 20 June 2011 λ : Zm → Zn where m is the number of facets of P. In this note we use ESP-sequences in the sense of Larry Smith to show that the higher derived functors of the primitive element MSC: functor are independent of λ. Coupling this with results that appear in Bousfield (1970) primary 14M25 secondary 57N65 [3] we are able to enrich the library of nice homology coalgebras by showing that certain families of quasitoric manifolds are nice, at least rationally, from Bousfield’s perspective. © Keywords: 2011 Elsevier B.V. All rights reserved. Quasitoric manifolds Toric topology Higher homotopy groups Unstable homotopy theory Toric spaces Higher derived functors of the primitive element functor Nice homology coalgebras Torus actions Cosimplicial objects 1. Introduction Given an n-dimensional q 1 neighborly simple convex polytope P , there is a family of quasitoric manifolds M that sit over P .
    [Show full text]
  • How to Make Free Resolutions with Macaulay2
    How to make free resolutions with Macaulay2 Chris Peterson and Hirotachi Abo 1. What are syzygies? Let k be a field, let R = k[x0, . , xn] be the homogeneous coordinate ring of Pn and let X be a projective variety in Pn. Consider the ideal I(X) of X. Assume that {f0, . , ft} is a generating set of I(X) and that each polynomial fi has degree di. We can express this by saying that we have a surjective homogenous map of graded S-modules: t M R(−di) → I(X), i=0 where R(−di) is a graded R-module with grading shifted by −di, that is, R(−di)k = Rk−di . In other words, we have an exact sequence of graded R-modules: Lt F i=0 R(−di) / R / Γ(X) / 0, M |> MM || MMM || MMM || M& || I(X) 7 D ppp DD pp DD ppp DD ppp DD 0 pp " 0 where F = (f0, . , ft). Example 1. Let R = k[x0, x1, x2]. Consider the union P of three points [0 : 0 : 1], [1 : 0 : 0] and [0 : 1 : 0]. The corresponding ideals are (x0, x1), (x1, x2) and (x2, x0). The intersection of these ideals are (x1x2, x0x2, x0x1). Let I(P ) be the ideal of P . In Macaulay2, the generating set {x1x2, x0x2, x0x1} for I(P ) is described as a 1 × 3 matrix with gens: i1 : KK=QQ o1 = QQ o1 : Ring 1 -- the class of all rational numbers i2 : ringP2=KK[x_0,x_1,x_2] o2 = ringP2 o2 : PolynomialRing i3 : P1=ideal(x_0,x_1); P2=ideal(x_1,x_2); P3=ideal(x_2,x_0); o3 : Ideal of ringP2 o4 : Ideal of ringP2 o5 : Ideal of ringP2 i6 : P=intersect(P1,P2,P3) o6 = ideal (x x , x x , x x ) 1 2 0 2 0 1 o6 : Ideal of ringP2 i7 : gens P o7 = | x_1x_2 x_0x_2 x_0x_1 | 1 3 o7 : Matrix ringP2 <--- ringP2 Definition.
    [Show full text]
  • The Grothendieck Spectral Sequence (Minicourse on Spectral Sequences, UT Austin, May 2017)
    The Grothendieck Spectral Sequence (Minicourse on Spectral Sequences, UT Austin, May 2017) Richard Hughes May 12, 2017 1 Preliminaries on derived functors. 1.1 A computational definition of right derived functors. We begin by recalling that a functor between abelian categories F : A!B is called left exact if it takes short exact sequences (SES) in A 0 ! A ! B ! C ! 0 to exact sequences 0 ! FA ! FB ! FC in B. If in fact F takes SES in A to SES in B, we say that F is exact. Question. Can we measure the \failure of exactness" of a left exact functor? The answer to such an obviously leading question is, of course, yes: the right derived functors RpF , which we will define below, are in a precise sense the unique extension of F to an exact functor. Recall that an object I 2 A is called injective if the functor op HomA(−;I): A ! Ab is exact. An injective resolution of A 2 A is a quasi-isomorphism in Ch(A) A ! I• = (I0 ! I1 ! I2 !··· ) where all of the Ii are injective, and where we think of A as a complex concentrated in degree zero. If every A 2 A embeds into some injective object, we say that A has enough injectives { in this situation it is a theorem that every object admits an injective resolution. So, for A 2 A choose an injective resolution A ! I• and define the pth right derived functor of F applied to A by RpF (A) := Hp(F (I•)): Remark • You might worry about whether or not this depends upon our choice of injective resolution for A { it does not, up to canonical isomorphism.
    [Show full text]
  • Vector Bundles on Projective Space
    Vector Bundles on Projective Space Takumi Murayama December 1, 2013 1 Preliminaries on vector bundles Let X be a (quasi-projective) variety over k. We follow [Sha13, Chap. 6, x1.2]. Definition. A family of vector spaces over X is a morphism of varieties π : E ! X −1 such that for each x 2 X, the fiber Ex := π (x) is isomorphic to a vector space r 0 0 Ak(x).A morphism of a family of vector spaces π : E ! X and π : E ! X is a morphism f : E ! E0 such that the following diagram commutes: f E E0 π π0 X 0 and the map fx : Ex ! Ex is linear over k(x). f is an isomorphism if fx is an isomorphism for all x. A vector bundle is a family of vector spaces that is locally trivial, i.e., for each x 2 X, there exists a neighborhood U 3 x such that there is an isomorphism ': π−1(U) !∼ U × Ar that is an isomorphism of families of vector spaces by the following diagram: −1 ∼ r π (U) ' U × A (1.1) π pr1 U −1 where pr1 denotes the first projection. We call π (U) ! U the restriction of the vector bundle π : E ! X onto U, denoted by EjU . r is locally constant, hence is constant on every irreducible component of X. If it is constant everywhere on X, we call r the rank of the vector bundle. 1 The following lemma tells us how local trivializations of a vector bundle glue together on the entire space X.
    [Show full text]
  • Derived Functors and Homological Dimension (Pdf)
    DERIVED FUNCTORS AND HOMOLOGICAL DIMENSION George Torres Math 221 Abstract. This paper overviews the basic notions of abelian categories, exact functors, and chain complexes. It will use these concepts to define derived functors, prove their existence, and demon- strate their relationship to homological dimension. I affirm my awareness of the standards of the Harvard College Honor Code. Date: December 15, 2015. 1 2 DERIVED FUNCTORS AND HOMOLOGICAL DIMENSION 1. Abelian Categories and Homology The concept of an abelian category will be necessary for discussing ideas on homological algebra. Loosely speaking, an abelian cagetory is a type of category that behaves like modules (R-mod) or abelian groups (Ab). We must first define a few types of morphisms that such a category must have. Definition 1.1. A morphism f : X ! Y in a category C is a zero morphism if: • for any A 2 C and any g; h : A ! X, fg = fh • for any B 2 C and any g; h : Y ! B, gf = hf We denote a zero morphism as 0XY (or sometimes just 0 if the context is sufficient). Definition 1.2. A morphism f : X ! Y is a monomorphism if it is left cancellative. That is, for all g; h : Z ! X, we have fg = fh ) g = h. An epimorphism is a morphism if it is right cancellative. The zero morphism is a generalization of the zero map on rings, or the identity homomorphism on groups. Monomorphisms and epimorphisms are generalizations of injective and surjective homomorphisms (though these definitions don't always coincide). It can be shown that a morphism is an isomorphism iff it is epic and monic.
    [Show full text]
  • The Calabi Complex and Killing Sheaf Cohomology
    The Calabi complex and Killing sheaf cohomology Igor Khavkine Department of Mathematics, University of Trento, and TIFPA-INFN, Trento, I{38123 Povo (TN) Italy [email protected] September 26, 2014 Abstract It has recently been noticed that the degeneracies of the Poisson bra- cket of linearized gravity on constant curvature Lorentzian manifold can be described in terms of the cohomologies of a certain complex of dif- ferential operators. This complex was first introduced by Calabi and its cohomology is known to be isomorphic to that of the (locally constant) sheaf of Killing vectors. We review the structure of the Calabi complex in a novel way, with explicit calculations based on representation theory of GL(n), and also some tools for studying its cohomology in terms of of lo- cally constant sheaves. We also conjecture how these tools would adapt to linearized gravity on other backgrounds and to other gauge theories. The presentation includes explicit formulas for the differential operators in the Calabi complex, arguments for its local exactness, discussion of general- ized Poincar´eduality, methods of computing the cohomology of locally constant sheaves, and example calculations of Killing sheaf cohomologies of some black hole and cosmological Lorentzian manifolds. Contents 1 Introduction2 2 The Calabi complex4 2.1 Tensor bundles and Young symmetrizers..............5 2.2 Differential operators.........................7 2.3 Formal adjoint complex....................... 11 2.4 Equations of finite type, twisted de Rham complex........ 14 3 Cohomology of locally constant sheaves 16 3.1 Locally constant sheaves....................... 16 3.2 Acyclic resolution by a differential complex............ 18 3.3 Generalized Poincar´eduality...................
    [Show full text]
  • LECTURE 6: FIBER BUNDLES in This Section We Will Introduce The
    LECTURE 6: FIBER BUNDLES In this section we will introduce the interesting class of fibrations given by fiber bundles. Fiber bundles play an important role in many geometric contexts. For example, the Grassmaniann varieties and certain fiber bundles associated to Stiefel varieties are central in the classification of vector bundles over (nice) spaces. The fact that fiber bundles are examples of Serre fibrations follows from Theorem ?? which states that being a Serre fibration is a local property. 1. Fiber bundles and principal bundles Definition 6.1. A fiber bundle with fiber F is a map p: E ! X with the following property: every ∼ −1 point x 2 X has a neighborhood U ⊆ X for which there is a homeomorphism φU : U × F = p (U) such that the following diagram commutes in which π1 : U × F ! U is the projection on the first factor: φ U × F U / p−1(U) ∼= π1 p * U t Remark 6.2. The projection X × F ! X is an example of a fiber bundle: it is called the trivial bundle over X with fiber F . By definition, a fiber bundle is a map which is `locally' homeomorphic to a trivial bundle. The homeomorphism φU in the definition is a local trivialization of the bundle, or a trivialization over U. Let us begin with an interesting subclass. A fiber bundle whose fiber F is a discrete space is (by definition) a covering projection (with fiber F ). For example, the exponential map R ! S1 is a covering projection with fiber Z. Suppose X is a space which is path-connected and locally simply connected (in fact, the weaker condition of being semi-locally simply connected would be enough for the following construction).
    [Show full text]