You are on page 1of 123

Elementary Applied Topology

Robert Ghrist University of Pennsylvania

EE

Contents
Preface 1 Manifolds
1.1 Manifolds . . . . . . . . . . . . . 1.2 Conguration spaces of linkages . 1.3 Derivatives . . . . . . . . . . . . 1.4 Vector elds . . . . . . . . . . . 1.5 Braids and robot motion planning 1.6 Transversality . . . . . . . . . . 1.7 Signals of opportunity . . . . . . 1.8 Stratied spaces . . . . . . . . . Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii
2 3 5 6 7 9 12 14 16 20 22 23 23 25 25 26 27 29 30 32 34

1

2

Complexes

2.1 Simplicial and cell complexes . . . . . . . . . . . 2.2 Vietoris-Rips complexes and point clouds . . . . . 2.3 Flag complexes and networks . . . . . . . . . . . 2.4 Phylogenetic trees and links . . . . . . . . . . . . 2.5 Witness complexes and the sampling problem . . 2.6 ƒech and random samplings . . . . . . . . . . . . 2.7 Nerves and neurons . . . . . . . . . . . . . . . . 2.8 Strategy complexes and planning with uncertainty 2.9 Decision tasks and consensus . . . . . . . . . . . 2.10 Discretized graph conguration spaces . . . . . . 2.11 State complexes and reconguration . . . . . . . Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 3.2 3.3 3.4 3.5 Counting . . . . . . . . . . . Curvature . . . . . . . . . . . Obstructions to vector elds Index theory for vector elds Tame topology . . . . . . . . iii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

19

3

Euler Characteristic

38 39 40 41 42

37

. . . . . . . . . . . . . . . . . . . . . . . . . . . . Exact sequences Pairs . . . . . . . . .8 5. . . . . Exercises 5 Sequences 5. . . . . . . . . . . . . . . . . . . Homology of a relation Functoriality . . . . . . . . . . . . . .13 5. . . . . . . . .5 5. . . . . . . . . . . . . . . . . ƒech homology . .9 5. . . . . . Borsuk-Ulam theorems . . . . . . . . . . . . . . . . . . . . . . .6 5. .2 5. . . . . . . . . . . . . . . . . . . . . . . . . . . .8 4. . . . . . . Gaussian random elds . .8 3. . . . . . 93 94 94 98 . . . . . . . . . . . . . . . . . . . . . . .1 6. . . . . . . . . . . .3 Dual spaces . . . . . . . . . . .9 4. . . . . . The space of natural images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inverse kinematics Winding number and degree . . . . . . . . . . . . . . . . . . . . Exercises 6 Cohomology 6. . .2 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12 5. . .12 Chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 72 73 75 75 76 78 79 80 80 81 82 84 86 88 90 Mayer-Vietoris Degree and local computation . . . . . . . . . . . . . . . . . . . .3 5. . . . . . . . . . . . . . Exercises 4 Homology 4. . . . . . . . . . . . . . . Homological coverage Persistent homology . . . . . . . . . . . . Relative homology Local homology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lefschetz index . . . . . . . . .11 4. . . . . . . . . . . . . . . . . . . . . . . . . . Euler characteristic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 3. . . . . . . . Euler integral transforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 5. . . . . . . . . . . . . . . . .9 3. . . . . . . . Poincaré duality . . . . . . .1 5.11 5. . . . . . . . . . . . . . . . . . . . . . . . . . . Cochain complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Homology examples Coecients . . . Zigzag persistence . . . . . . . . . . . . . . . . . . . .3 4. . . . . 44 45 46 48 50 51 54 Target enumeration The Fubini Theorem . . . Intrinsic volumes . . . . . . . .7 4. .10 3. . . . . . .1 4.2 4. .11 Euler calculus . . . . . . . . . . . . . . . . 55 56 57 59 60 61 62 62 63 63 64 66 67 69 Singular homology Reduced homology . . . . . . . . . .10 4. . . . . . . . . . .14 Homotopy invariance . . . . . . . . . . . . . . . .10 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6 3. . . . . . . . .iv Contents 3. . . . . . . . . . . . . .4 5. . . . . . . . . . . . . . . . . . . . . .4 4. . . . . . Equivalence of homology theories Coverage in sensor networks . . . . . . . .6 4. . . . . . .5 4. . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . 8. . . . . . . . . 7. 8. . .6 Convex geometry . .3 Operations on sheaves . . . . . . . . . . . . . . . . . . 7. . . . 8. . . . . . . . . . . . . . . . . . . . 8. . . . . . . . .9 Bundles and sectional category . . . . . . . .4 Stratied Morse theory . . . . . 8. . . . . . . . . . . . . . .8 Diagrams and nerves . . . . . . . . . . . . . . . . 152 157 149 Bibliography . . . . . . . . . . . . .2 Examples of sheaves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 L-S category .6 Combinatorial Morse theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. . . . . . . . . . . . . . . . . . . . . . . . . . . .10 Circular coordinates in data sets . . . . . . . . . . . . . . . . . . . . . . . . . 99 99 101 101 105 106 108 109 111 7. . . . . . . . . . . . . . . . . . .4 Alexander duality . . . . . . . . A. . . . . . 8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. . . . . . . . . . . . . . . . . .1 Presheaves . . . . . . . . . . . 7. . . . . . . . . . . . . . .10 Clustering functors . . . . . . . . . . . . . . . . . . . . . . .9 Laplacians . . . . . . . . . .4 Sheaf cohomology . . .6 Network coding . . . . . . . . . . . . . . . . . . . . . . . . . . .9 Functors . . . . . . . . . . . . . . . . . . . . 149 On linear and abstract algebra . . . . . . . . . . . . . .8 Unimodal decomposition in statistics . . . . .7 Categories . . . . . . . . . . . . . . . 7 Morse Theory v . . . . . . . . . . 7. . . . .Contents 6. . . . . . . . . . . . . . . . . . . . . . . . 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. . . . . . . . . . . . .3 Qualitative signal processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. . . . . . . .1 Classical Morse theory . . . . . . . . . . . . . . . . . . . . . . . . . . 8. . . . . . . 7.2 4-valued Euler integration . . . . .11 Natural transformations 8. . . . . .7 Forms and de Rham cohomology . . . . . . . . . . . . . .1 A. . . 112 114 115 117 117 120 122 123 125 126 129 132 134 136 137 139 140 141 143 143 144 146 146 147 8 Sheaves & Other Functors 131 A Background On point-set topology . . . . . . . . 8. . . . . . . . . . .10 Topological complexity of path planning Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 A. .5 Conley Index . . 150 On homotopy theory . . . . . . . . . . . . . . . . 6. . 6. . . . Exercises . . . . . . . . . . . . . . . 7. . . . . . . . . . . . . . 7. . . . . . . . . . .5 Euler integration over ad hoc networks 6. . . 8. . . . . . . . . . . . . .8 Currents . . . . . . . . . . 6. . . . . 7. . . . .12 Sheaves and categories Exercises . . . . . . . .5 Pursuit and evasion . . . . . .

vi Contents .

2HAB=?A .

or obstructions take the form of algebraic-topological quantities. 3. it is rife with principles and tools (Mayer-Vietoris. the genus of a surface. Implicit are the notions of proximity and continuity. even when the answers to the problems themselves are hard to compute. Certainly. degrees. Integration: Topology is the premiere tool for converting local data into global properties. Counting the number of connected components of a space is a simple topological invariant. all about arrows: maps. Proximity in a space is controlled by its topology: the collection of open neighborhoods of all points in . Obstruction: Topology often provides tools for answering feasibility of certain problems. Topology explores deformations of maps and spaces. We having maps write   for homotopic maps. As such. invariants. Homotopic spaces are those Id. For example.viii What topology is Preface 6 f 'f g X!Y opology is the mathematics of spaces and maps between them. A homotopy of maps   : is a continuous 1-parameter family of maps J : interpolating  and  . where data is never not noisy. Continuation: Topological properties are usually robust. Nested systems of neighborhoods communicate proximity without necessarily requiring a metric distance. sequences. The number of components is not something that changes with a small error in measurement. This is important in extracting high-level features. from a certain modern perspective. 2. These characteristics.f X ! Y f f X'Y What topology can do Topology was built to distinguish qualitative features of spaces and mappings. sheaves) for integrating local data. The term map always means a continuous function between spaces. It is good for. give global characteristics important to classication. Topology is. Characterization: Topological properties extract qualitative features. . . : and : with Id: and Topology does its work by means of distinguishing spaces and maps to various degrees of resolution. X X h Y !X hg ' f X!Y gh ' f . A more general deformation arises at the level of maps. Excision. inter alia: 1. The simple gures adorning chapters are drawn with this dictum in mind. convergence. spectral sequences. classes. diagrams. and the basic equivalence relations of topology emphasize qualitative features. This is vital to applications in scientic disciplines. functors. 4. a homeomorphism (a bijective map with continuous inverse) counts as a topological equivalence. This thin notion of closeness suces to dene all the necessary notions of continuity. and connectivity familiar to students more at home with calculus in Euclidean spaces. A topological invariant of spaces [maps] is an assignment of some (usually algebraic) datum to spaces [maps] which respects the equivalence relation of homotopy: homotopic spaces [maps] are sent to the same invariant. or the number of connected components of an object.

though necessary. This reects the author's learning of the subject of topology: despite the best eorts of brilliant topologists (including Profs. in the hopes that it will lure the unsuspecting reader to greater depths and prepare for the eld as it will be. for the sake of making the exposition quick and painless. This is not a text in computational topology: the reader may look to several excellent sources [?. rather than according to application domain. Krsti¢. If the reader is a scientist or applied mathematician hoping that topological tools are a quick x. Experts will likely be exasperated with this text. is the impetus for future work: hard. for many reasons. The chapters are organized according to mathematical sub-topic. This book represents a partial collection of such cartoons. There is no recourse to tweaking epsilons should desiderata fail to be found. This raises an interesting philosophical question about the nature of applied mathematics: is it how dierent branches of mathematics embed in the physical world. this precludes depth. The goal is to skip through the formalisms in order to get to applications as quickly as possible. tools. The reward of reading this book with care may be limited to the realization of new questions as opposed to new answers. take this text with caution. is to make the text intentionally simplistic. no matter how cartoonish. perspectives. The best approach. and it seems certain that even a more detailed treatment of the examples given here would likely appear quaint in less than a decade. properly used. The intent is to span a broad array of ideas. ?] for the problem of algorithmic complexity of the topological objects explored in this text. emphasizing examples and avoiding the set-theoretic denitions which. What this text is This text is a quick tour of examples to motivate the subject of applied topology. This text. The text is .Preface What topology cannot do ix Topology is ckle. and Vogtmann). The interested reader will have to look elsewhere for more details. The subject is in its infancy. or is how dierent applications implicate and are aided by mathematics? The organization of the text reects the author's view that applied mathematics concerns the incarnation of mathematical objects and structures. Dranishnikov. The reader should not conclude that the subject is quickly or painlessly learned. slow. and applications. and fruitful. This is not a mathematics text: proofs are seldom given. the author never learned much of anything in the subject without rst nding some physical manifestation of the principle. perhaps. may overly discourage the interested scientist. The author would have called this text Cartoons in Applied Topology were it not for the resulting confusion. It is not uncommon that a new mathematical tool contributes to applications not by answering a desired question but by revealing a dierent (and perhaps more signicant) underlying principle. Hatcher. The text begins with the necessary background in spaces. both in the mathematics and in the applications. The author believes that the questions of What is it good for? and How do I compute that? are neither independent nor inseparable.

to point them in the direction of better texts to come. The vision and guidance of (chronological order) Philip Holmes. Bob Williams. The author has had the pleasure of collaborating with and learning from Yuliy Baryshnikov. and Rob Vandervorst during this last decade of development in the eld. Peter Bubenik and John Oprea above all The young eld of Applied Topology has numerous practitioners and supporters. typesetting by L . Dan Koditschek. Yasu Hiraoka. Thank you for your patience. and the AFOSR. The work behind and writing of this text was made possible through the generous support of DARPA. Vin de Silva. the primum mobile of the eld. All gures were drawn by the author using Adobe Illustrator. and their consistent support for the author are worth more than words can repay. the NSF. )TEX. Acknowledgements A debt of thanks is due to the organizers and attendees of the 2009 CMBS meeting at Cleveland State University. Michael Robinson. the ONR. and Gunnar Carlsson have been invaluable. Of particular note is Benjamin Mann. The author is also exasperated with a text that was written with suboptimal constraints on space and time. Konstantin Mischaikow. John Etnyre.x Preface meant not for experts but for students.

Chapter 1 Spaces: Manifolds .

or C 1 ) whenever U \ U.6 2 Chapter 1. A topological n-manifold is a space M locally homeomorphic to Rn . Manifolds he most intuitive and initially useful topological constructs are smooth manifolds and smooth functions between them. This consists of insisting that the maps 1. That is. one needs a smoothing of a manifold.1 Manifolds are smooth (innitely dierentiable. there is a cover U = fU g of M by open sets along with maps  : U ! Rn that are continuous bijections onto their images with continuous inverses. Manifolds are the extension of domains familiar from calculus  curves and surfaces  to higher-dimensional settings. In order to do dierential calculus.

The standard tools of multivariable calculus  the Inverse and Implicit Function Theorems  lift to manifolds and allow for a simple means of producing interesting examples. spectacular progress [?] that perches this dimension between the simple (2) and the wholly bizarre (4). The classication of 3-manifolds is a delightfully convoluted story [?]. are often immediately ignored. 6= . if so. ignoring whatever complexities black holes. Charts and atlases are rarely explicitly constructed. they generate a maximal atlas of charts which specify a smooth structure on M . with recent. The pairs (U .. Smooth curves are 1-manifolds. This is a simplication.  All . thus. seemingly. strings. The Classication Theorem for Surfaces states that any compact surface is dieomorphic to the orientable or non-orientable surface of some xed genus g  0. and other exotica produce. and the existence of holes or handles is captured in the invariant called genus. and. manifolds in this text should be assumed Hausdor and paracompact. The story for 2-manifolds  surfaces  introduces two more parameters. a 3manifold . Compact surfaces can be orientable or non-orientable. The sphere S2 is the orientable surface of genus zero.  ) are called charts. the torus T2 the orientable surface of genus one. The spatial universe is. compactness suces to distinguish the two. Any connected curve is dieomorphic (smoothly homeomorphic with smooth inverse) to either R or to the circle S1 . their nonorientable counterparts are the projective plane P2 and the Klein bottle K 2 respectively. The reader for whom these terms are unfamiliar is encouraged to ignore them for the time being. easily classied.

  1 :  (U \ U.

) ! .

(U \ U.

) .

A simple application of (topological) manifolds to robotics falls under a dierent aegis. Sn is a connected n as manifold dieomorphic to the compactication of R⨿ n follows. 3 Example 1. Sn . The 0-dimensional sphere S0 is disconnected  it is the disjoint union of two points. Consider a planar mechanical linkage consisting of several at. is dened as the n +1 space of all 1-dimensional linear subspaces of R . SO3 . with topology induced in like manner to Pn . P2 . For any n. Identify opposite sides of a square with edge orientations reversed. is a non-orientable surface dieomorphic to the following quotient spaces: 1.2. The projective plane. This abstract space is dieomorphic to Sn via a dieomorphism that sends the origin and ? to the south and north poles of the sphere Sn respectively. with the topology that says neighborhoods of a point in Pn are generated by small open cones about the associated line. is the set of points in Euclidean Rn+1 unit Pn . 2. Consider the quotient space obtained from R ?.2 Conguration spaces of linkages Applications of manifolds and dierential topology are ubiquitous in rational mechanics.1. Identify antipodal points on the 2-sphere S2 . The n-sphere is an n-dimensional manifold. The Grasmannian Gn k is dened as the space of all k -dimensional subspaces of Rn . One can use out-of-plane height (or mathematical license) to assert that interior intersections of rods are ignorable. where a : Sn ! Sn is the antipodal map a(x ) = x . Pn is dieomorphic to the quotient Sn =a. The Grassmannian is a manifold that specializes to +1 Pn = Gn .2 (Projective spaces) The real projective space. where ? is an abstract point whose neighborhoods consist of ? union the points in Rn suciently far from the origin. 1 1. P1 . Among the many possible extensions of projective spaces.1 (Spheres) The n-sphere. it appears obvious). ?]. Identify antipodal points on the boundary of the closed unit ball B  R2 . Projective 1-space. The space P3 is dieomorphic to the space of rotation matrices. Conguration spaces of linkages distance to the origin. That Pn is an n-manifold for all n is easily shown (but should be contemplated until Example 1. and mathematical physics and are well-covered in standard texts [?. Hamiltonian dynamics. the Grassmannian spaces arise in numerous contexts. the group of real 3-by-3 matrices with determinant 1. The conguration space of the linkage is a topological space that assigns a point to each conguration of the linkage  a relative positioning of the rods up to equivalence generated by rotations and translations in the plane  and which . rigid rods joined at their ends by pins that permit free rotation in the plane. 3. For n > 0. is dieomorphic to S1 .

Manifolds assigns neighborhoods in the obvious manner. Many other familiar manifolds are realized as conguration spaces of planar linkages (with judicious use of the third dimension to mitigate bar crossings). Four rods of lengths fLi g4 1 are linked in a cyclic chain. ?. The conguration space of a planar linkage is almost always a manifold. The second circle comes from taking the mirror image of the linkage along the axis of its xed rod in the plane and repeating the circular motion there: this forms an entirely separate circle's worth of conguration states. ?. a motor) into rocking motion (as in a windshield wiper). The conguration space is thus one-dimensional and almost always a manifold. used extensively in mechanical components. ⨿ The conguration space of the illustrated linkage is S1 S1 . the system is seen to have one mechanical degree of freedom. Theorem 1. Any smooth compact manifold is dieomorphic to the conguration space of some planar linkage. depending on the length between the xed ends of the linkage [?]. The realization question this exercise prompts has a denitive answer (albeit with a convoluted attribution and history [?. the coproduct or disjoint union of two circles. The linkage illustrated in the margin has conguration space a closed. one has a single short rod. When one rod is anchored. This remarkable result provides great consolation to students whose ability to conceptualize geometric dimensions greater than three is limited: one can sense all the complexities of manifolds by hand via kinematics. ?]): Example 1. say. This linkage is used to transform spinning motion (from. Example 1. orientable surface of genus g ranging between 0 and 4. The reader is encouraged to try building a linkage whose conguration space yields an interesting 3-manifold. A neighborhood of a conguration is all congurations obtainable via a small perturbation of the mechanical linkage. The reader is encouraged to build a few congurable linkages and to determine the dimensions of the resulting conguration spaces. or crank-rocker linkage. causing the opposing rod to rock back-and-forth.4 Chapter 1.5 (Robot arms) A robot arm is a special kind of mechanical linkage in which joints are sequentially attached by rigid rods. One end of the arm is xed . as illustrated. the dimension of which conveys the number of mechanical degrees of freedom of the device. then this rod can be rotated completely about its anchor. The undergraduate thesis of Walker [?] has many examples of orientable 2-manifolds as conguration spaces of planar linkages. If.4.3 (Crank-rocker) The canonical example of a simple useful linkage is the Grashof 4-bar.

including the map to R3 which records the location of the end of the arm. Among the most commonly available joints are pin joints (cf. an elbow) and rotor joints (cf. the cross product of n circles. Dierentiability is a prototypical example. pick-n-place. or the map to SO3 that records the orientation (but not the position) of an asymmetric part grasped by the end manipulator. Ignoring the (nontrivial!) potential for collision of an arm in a convoluted conguration. A map between manifolds f : M ! N is dierentiable if pushing it down via charts yields a dierentiable map. gradients. etc. There are natural maps associated with this conguration space. and more are familiar constructs of calculus that extend to arbitrary manifolds by means of localization.3 Derivatives Derivatives.1. where n is the number of rotational or pin joints. each with conguration space S1 . Derivatives 5 (mounted to the oor) and the other is free (usually ending in a manipulator for manufacturing. vector elds. whenever f takes p 2 U  M to f (p ) 2 V. Specically. the conguration space ∏ of this arm in R3 has the 1 n topology of the n-torus.3. grasping. 1. rotation of a forearm).). T = n 1 S .

one has .  N .

 f   1 a smooth map from a subset of Rm to a subset of Rn . The derivative of f at p is therefore dened as the derivative of .

An element of Tp M is of the form v = [ 0 (0)]. This notion is the rst point of departure from the calculus mindset  in elementary calculus classes there is a general confusion between tangent vectors (e. where [] denotes the . A deeper inquiry leads to a signicant construct in dierential topology. but this pictorial representation is dangerously ill-dened  in what larger space does this tangent space reside? Do dierent tangent spaces intersect? There are several ways to correctly dene a tangent space. The most intuitive uses smooth curves. and one must check that the choice of chart does not aect the result. but it is not satisfying. The tangent space to a manifold M at a point p 2 M .g. Two such curves.. where the origin of Tp M is abstractly identied with p itself. and ~ are equivalent if and only if 0 (0) = ~ 0 (0) in some (and hence any) chart.  f   1 . Tp M . It is tempting to illustrate the tangent space as a vector space of dimension dim M that is tangent to the manifold. Dene Tp M to be the vector space of equivalence classes of dierentiable curves : R ! M where (0) = p . It suces to use charts and coordinates to understand derivatives. from vector elds) and points in the space itself. is a vector space of tangent directions to M at p .

The vector space structure is inherited from that of the chart in Rn . t 2 R. 3. However. In this topology. Manifolds equivalence relation. s +t (x ) = t (s (x )) for all d  (x ) = (x ). a ready and compelling exemplar (in light of Example 1. The next step is crucial: one forms the disjoint union of all tangent spaces Tp M . v ). That this is so is not so obvious. Given  a vector eld on M . For M non-compact or  insuciently smooth. An isolated xed point may have several qualitatively distinct features based on stability. dt t x 2 M and s. v ) is a cross product of a neighborhood of v in Tp M with a neighborhood of p in M . it is not the case that T S2  = S2  R2 . Specically. A vector eld on M is a choice of tangent vectors (p ) 2 Tp (M ) which is continuous in p . 0 (x ) = x for all x 2 M . The dynamics of vector elds and ows is the fundamental overlap between topology and dierential equations. Such a map  is called a section of T M . Example 1. .4 Vector elds The formalism of tangent spaces and tangent bundles simplies the transition of calculus-based ideas to arbitrary manifolds. The natural topology on T M is one for which a neighborhood of (p. One thinks of t (x ) as determining the location of a particle at x following tangent to  for t units of time. one must worry about existence and uniqueness of solutions: such questions are not considered in this text.  : M ! T M is a (continuous) map satisfying    = IdM for  : T M ! M the projection map taking a tangent vector at p to p itself. the ow associated to  is the family of dieomorphisms t : M ! M satisfying: 1. vector elds can be integrated to yield a ow. the tangent bundle of a circle is dieomorphic to S1  R1 . Smooth vector elds on compact manifolds always yield smooth ows. In any particular chart. For example. Invariance with respect to charts implies that the derivative of f : M ! N at p 2 M is realizable as a linear transformation Dfp : Tp M ! Tf (p) N . where v 2 Tp M . 2. An element of T M is of the form (p. T M is a smooth manifold of dimension equal to 2 dim M .6 (Equilibria) The primal objects of inquiry in dynamics are the equilibrium solutions: a vector eld is said to have a xed point at p if (p ) = 0. A tangent vector coincides with the intuition from calculus in the case of M = Rn . a basis of tangent vectors may be chosen to realize Dfp as the Jacobian matrix of partial derivatives at p . p 2 M . into a self-contained space T M called the tangent bundle of M . 1. As in the case of dierential equations on Rn .9) is the topology and dynamics of vector elds.6 Chapter 1.

As a rst step at modeling such . such equilibria are fundamentally stable solutions. e.1) above. unstable. Periodic orbits are submanifolds dieomorphic to S1 .1) Example 1. dened by taking the limit t ! 1 in (1.5. controlled by an algorithm that either guides the robots from initial positions to goal positions (in a warehousing application). saddle-type.. or (3) innite. the blood-and-bones of periodic orbits and equilibria. or executes a cyclic pattern (in manufacturing applications). for typical systems.5 Braids and robot motion planning A dierent class of conguration spaces is inspired by applications in multi-agent robotics. however. W u (p ). such solutions are. W s (p ) is in fact a manifold. The existence of periodic orbits  in contrast to that of equilibria  is a computationally devilish problem. together with the musculature of their stable and unstable manifolds.g.g. 1. such equilibria are fundamentally unstable. or degenerate. The minimal such T > 0 is the period of the orbit. it is possible to nd smooth. On S3 . A sink is a xed point p whose stable manifold contains an open neighborhood of p . the Hopf eld) [?].. Braids and robot motion planning The 7 stable manifold of a xed point p is the set W s (p) = fx 2 M : tlim  (x ) = pg: !1 t For typical xed points of a typical vector eld. One may classify periodic orbits as being stable. These robots are costly and cannot tolerate collisions. A periodic orbit of a ow is an orbit ft (x )gt 24 satisfying t +T (x ) = (x ) for some xed T > 0 and all t 2 R. Consider an automated factory equipped with mobile robots. xed-point-free vector elds whose set of periodic orbits is. The dynamics of vector elds goes well beyond equilibria and periodic orbits (see. as is the unstable manifold. gives the basic structure for reasoning about the body of behavior. [?]). then the analogous circulatory system would be comprised of the periodic orbits. A common goal is to place several such robots in motion simultaneously. like the Nash equilibria they encompass. (2) empty [?]. A saddle equilibrium satises dim W s (p ) > 0 and dim W u (p ) > 0. respectively: (1) all of S3 (e. expressing all types of knotting of loops possible [?].7 (Periodic orbits) If a continuous-time dynamical system  a ow  is imagined to be supported by the skeleton of its equilibria. minimax solutions. A source is a p whose unstable manifold contains an open neighborhood of p .1. (1.

represents those congurations of n points in X which experience a collision  this is the set of illegal congurations for the robots.8 Chapter 1. which itself corresponds to moving the braid strands in such a way that they cannot intersect. on a factory oor). in some settings. is very relevant to mobile robot motion planning (e. 2 Example 1. however. UCn (X ) = Cn (X )=Sn : This space is given the quotient topology: a subset in UCn (X ) is open if and only if the union of preimages in C(X ) is open. the reader will reason correctly that there are innitely many inequivalent motion plans between starting and ending congurations. denoted Cn (X ).  = f(xi )n (1.g. anonymity is not detrimental  any camera will do. denoted UCn (X ). it is also among the most topologically interesting conguration spaces. There are applications for which labeling the points is important. and the conguration space Cn (R2 ) forms an acceptable model for robot motion planning on an unobstructed oor. Manifolds a system. Conguration spaces of points on a manifold M are all (non-compact) manifolds of dimension dim Cn (M ) = n dim M . an image of S1 in the conguration space. Of course. perhaps..2) 1 : xi = xj for some i 6= j g: 1 The set . UCn (R2 ). Consider the case of n robots which begin and end at xed congurations.) How many dierent ways are there for the robots to wind about one another en route from their starting to ending locations? The space-time graph of a path in conguration space is a braid. From the point of view of topology. if the robots are not labeled. a weaving of strands tracing robot paths. However. and near-collisions are unacceptable. is dened to be the quotient of Cn (X ) by the natural action of the symmetric group Sn which permutes the ordered points in X . it is a loop. The conguration space of n distinct labeled points on a topological space X . or. assume the location of each robot is a mathematical point in a space X (typically a domain in R2 or R3 ). and a few sketches. The space Cn (S1 ) is homeomorphic to (n 1)! disjoint copies of S1  Rn 1 . A deformation in the motion plan equals a homotopy of the path (xing the endpoints). with warehousing. removing a suciently small neighborhood of  gives a space equivalent to Cn (X ). This complex motion corresponds to a path in Cn (R2 ). (If the path has the same beginning and ending conguration. in which robots move specic packages. while UCn (S1 ) is a connected space. The unlabeled conguration space. the pairwise diagonal. being a prime example. is the space ( n ) ∏ n C (X ) = X  . in practice. such as mobile security cameras in a building.8 (Braids) The conguration space of points on R . From this. robots are not point-like. tracing out non-colliding routes in between.

5. Transversality 9 both in modeling complex mechanisms or automated guided vehicles. One reason for specifying a vector eld on the entire conguration space (instead of simply dictating a path from initial to goal states) is so that if the robot experiences an unexpected failure in executing a motion..  from Equation (1. One of many useful techniques for controlling robotic systems to execute behaviors and avoid collisions is to place a vector eld on the conguration space and use the resulting ow to guide systems towards a goal conguration.9 (Navigation elds) The use of conguration spaces in robotics is widespread. in automated guided vehicles [?]. Two intersecting surfaces in R3 generically intersect along curves. the vector eld automatically corrects for the failure. This approach to robot motion planning makes extensive use of dierential equations and dynamics on manifolds of conguration spaces. This technique has been successfully applied in visual servoing [?]. Coordinatize C2 (S1 )  T2 as f(1 . Additional features are also desirable: near the collision set (e. requiring explicit coordinate systems. so that (almost) all initial conditions will ow to and remain on the goal. 4. and insectoid. consider the problem of mobilizing a pair of robots on a circular track so as to patrol the domain while remaining as far apart as possible. 3.[?]. the vector eld should be pointed away from the collision set. A real square matrix A is generically invertible. Intuition is an acceptable starting point: consider the following examples of generic features of smooth manifolds and mappings: 1. Two curves in Rn generically do not intersect for n > 2. 2 ) . Programming such a vector eld is a challenge. in robots that juggle [?].1. 1 6= 2 g. 1. Three curves in R2 generically do not have a point of mutual intersection.6 Transversality Genericity is often invoked in applications.  _2 = 1 + sin(2 1 ):  All initial conditions evolve to this desired state of circulation on S1 . . _1 = 1 + sin(1 2 ) . construct an explicit vector eld on X having the goal as a global attractor.2)). where each leg position/momentum is represented by S1 . The task: given a conguration space X and a specied goal point or loop in X . Enlarging the problem from path-planning to eld-planning returns stability and robustness. This is an excellent model not only for motion on a circular track. but seldom explained in detail. and in hopping. Two intersecting curves in R2 generically intersect in a discrete set of points.6. 2. walking[?]. This can be done by means of a vector eld on C2 (S1 ). For example. but also for an alternating gait in legged locomotion.[?] robot locomotion.g. The reader may check that the following vector eld has f1 = 2  g as an attracting periodic solution and f1 = 2 g as a repelling set: Example 1.

Tp V  Tp W = Tp M 8 p 2 V \ W: (1. Consider a dierentiable between smooth manifolds. All are provably true with precise meaning using the theory of transversality. The roots of a polynomial are generically distinct. 7. The xed points of a vector eld are generically discrete. a map f is transverse to a submanifold W  M if and only if f J W . The conguration space of a planar linkage is generically a manifold. 9. written V J W . if and only if. This is equivalent to the statement that.10 6. Theorem 1.3) Otherwise said. Note that the absence of intersection is automatically transverse. For example.10 (Preimage Theorem). written f J g . Manifolds Critical points of a R-valued function on a manifold are generically discrete. Some of these seem obviously true. A point q 2 N is a regular value of f : M ! N if f J fq g. submanifold of map then M of dimension dim f :M!N f 1 (W ) is a (1. The notion of transverse maps is the rst of many examples of this principle. Two smooth maps f : are transverse. This is one hint that dierential tools are ecacious in the management of singular behavior. If f J W for W  N a submanifold. 8. Chapter 1. This mapcentric denition does not constrain the images of maps to be submanifolds. W in M are transverse.5) f 1 (W ) = dim M + dim W dim N: . 10.4) Dfv (Tv V )  Dgw (Tw W ) = Tp M 8f (v ) = g (w ) = p: This means that the degrees of freedom in the intersection of images of f and g span the full degrees of freedom in M . if and only if: V !M and g:W !M (1. two planes in R3 are transverse if and only if they are not identical. Likewise. A central theme of topology is the lifting of concepts from spaces to maps between spaces. Note that submanifolds V. Two submanifolds V. One benet of transversality is that localized linear-algebraic results can be pulled back to global results. the derivative is a surjection  the matrix of Dfp is of rank at least n = dim N . A generic map of a surface into R5 is injective. the tangent spaces to V and W span that of M . others less so. for each p 2 f 1 (q ). W  M are transverse if and only if the inclusion maps V : V ! M and W : W ! M are transverse as maps.

The xed point set of  is therefore the preimage  1 ( ). A subset of a topological space is said to be residual if it contains a countable intersection of open. for special singular values. the transversality condition is checked by showing that the mapping f has a derivative of full rank at the appropriate (regular) value.6. The proof of this theorem relies heavily on Sard's Lemma: the regular values of a suciently smooth map between manifolds are generic in the codomain. Recall from Ÿ1. Its dimension is n = 2n n + 0. this set is a submanifold of dimension dim Fix() = dim M + dim M dim T M = 0: . the set of smooth maps f : M ! N with f J W is residual in C 1 (M. Tnnis the inverse image f 1 (1. This intuition is the driving force behind the Transversality Theorem.1. which is therefore seen to be a disjoint union of two compact manifolds. For a generic perturbation of . 4. For M and N smooth manifolds and W  N a submanifold. The sphere Sn is the inverse image of 1 under f : Rn+1 ! R given by f (x ) = kx k. Example 1. 1) of the map f : Cn ! Rn given 2. 0)g of all zero vectors. Such regularity seems to fail rarely. Baire) spaces.6) This provides an eective means of constructing manifolds without the need for explicit charts: it is often used in the context of a regular value of a map. Transversality 11 The proof of the Preimage Theorem relies on linearization and a simple fact from linear algebra: for a linear transformation T : U ! V . 1. In the above examples. For reasonable (e.4 that a smooth vector eld is a section. and hence form a decent notion of topological typicality. The matrix group On of rigid rotations of Rn (both orientation preserving and reversing) can be realized as the inverse image f 1 (I ) of the map from n-by-n real matrices to symmetric n-by-n real matrices given by f (A) = AAT . The zero-section   T M is the set f(p. dim ker T = dim U + dim coker T dim V: (1. N ). Theorem 1.11 (Transversality Theorem).. then this set of transverse maps is both open and dense. It is a manifold of dimension n = (n + 1) 1 + 0. If W is closed. is a manifold of the same dimension as On . The 1 1 dimension of On is n2 2 n(n + 1) + 0 = 2 n(n 1).12 (Fixed points of a vector eld) The xed point set of a dieren- tiable vector eld on a compact manifold M is generically nite. A property dependent upon a parameter  2  is said to be a generic property if it holds for  in a residual subset of   even when that subspace is not explicitly given. dense subsets.g. The torus ∑ by f (z ) = k =1 kzk k. thanks to transversality. : : : . or smooth map  : M ! T M with  ((p )) = p for all p 2 M . as the inverse image of +1 under this restricted det. residual sets are dense. 3. As such. SOn . the space of all smooth (innitely dierentiable) maps from M to N . The determinant map restricted to On is in fact a smooth map to the 0-manifold S0 = f1g.

7 Signals of opportunity Applications of transversality are alike: (1) set up the correct maps/spaces. Proof. v3 ) = det [v2 v1 . but does not remove it. since dim R 1 + dim W dim C3 (R2 ) = 1 + 5 6 = 0: This. Example 1. Other consequences are not so obvious. v2 . N Any continuous function f :M! M is between smooth manifolds is generically perturbed to a smooth injection when dim N > 2 dim M . v3 v1 ] : This map has zero as a regular value. Manifolds A 0-dimensional compact submanifold is a nite point set. (2) invoke transversality. f (y )): . one expects collinearity at a discrete set of times. sinks. With more careful analysis of the meaning of transversality. generically. only source. a generic perturbation of the paths perturbs where the alignment occurs.14 (Whitney Embedding Theorem). y ) 7! (x. as in computing the generic intersection of curves and surfaces in R3 . How large? The critical bound comes. implies a stability in the phenomenon of collinearity: at such an alignment. This has some simple consequences. and saddles are generic xed points (cf. and Theorem 1. from the proper transversality criterion and a dimension count: Theorem 1. y. The conguration space C2 M = M  M a manifold of dimension 2 dim C2 f : C2 M 2 M. It follows from Theorems 1. The graph of f M of two distinct points on induces a map !C M N N : (x. it can be shown that the type of xed point is also constrained: on 2-d surface.10 that. 1.10 implies that the set of collinear congurations is a submanifold W = f 1 (0) of C3 (R2 ). Ÿ3.13 (Beacon alignment) Consider three people walking along generic smooth paths in the plane R . How often are their positions collinear? One considers the map f : C (R ) ! R which computes 2 3 2 the (signed) area of the triangle spanned by the three locations at an instant of time: f (v1 .11 and 1. A set of three paths is a generic map from R (time) into C3 (R2 ). f (x ). a submanifold as its image when the dimension of the target is large enough. after generic perturbation.3).12 Chapter 1. moreover. The following theorem states that any continuous map of a source manifold into a target manifold has. as it must. (3) count dimensions.

why other signals could not be used. this implies that a receiver can be localized to a unique position in a planar domain D using only a sequence of ve or more transmission signals. passive signals  arising naturally from TV transmissions. performing a reconstruction of the receiver location after-the-fact. Consider a scenario in which one wants to determine location in an unknown environment. all with nontrivial power constraints. implying unique channel response and the feasibility of localization in D via TOA. The following is a simple mathematical model for localization via signals of opportunity. nor does the receiver compute any distances-to-transmitters. There is a small but fascinating literature on the use of such signals of opportunity to solve localization and mapping problems. generic perturbations of f induce generic perturbations of C2 f .10 and 1.7. Certainly. and more.11. however. Such a device works by receiving signals from multiple satellite transmitters and utilizing known data about the transmitters to determine location. According to the appropriate! transversality theorem. it may be most feasible to realize a localization scheme a posteriori. do the received TOA (time of arrival) signals uniquely localize the receiver? Consider the signal prole mapping T : D ! RN which records the TOA of the (received. In fact. In practice. Thm. A simple application of Whitney's theorem. Assume there exist N transmitters of xed location which asynchronously emit pulses whose times of arrival can be measured by a receiver at any location in D. 4. For example. then the resulting perturbation embeds the domain D smoothly for N > 2k . within the tolerances of the signal reception. they are unreliable in urban canyons. globally readable from generically-placed transmitters. Note: it is not assumed that the signals are perfectly spherical. due to Michael Robinson [?] informs a problem of localization via signals. This sophisticated system requires many independent components to operate. including geosynchronous satellites. radio. Thus. Given a receiver located at an unknown point of D. If we assume that the generic placement of transmitters associated with this system provide a generic perturbation to T within C 1 (D. RN ).1. Consider a connected open domain D  Rk which is a k -manifold. even ionospheric waves induced by lightning strikes []  provide easily measured pulses with which to attempt to reconstruct location.13] . GPS devices are not universally available: they do not operate underwater or indoors. a GPS device would suce. Though wonderfully useful. identied. the generic dimension of the non-injective set is: dim C2 M + dim(C2 M  N ) dim(C2 M  N  N ) = 2 dim M dim N < 0: Thus. Signals of opportunity 13 The set of points on which f is non-injective is (C2 f ) 1 (C2 M  N ). where N  N  N is the diagonal. From Theorems 1. and the result is a smooth submanifold. the mapping T is generically injective. There is no reason. and ordered) transmitter pulses. ! This follows from the more general multi-jet transversality theorem [?. synchronized clocks. there are no self-intersections. the need for satellites and synchronized clocks can limit availability.

It follows from transversality that any reasonable signal space of sucient dimension preserves the ability to localize based on knowledge of the image of D under T . m . Such piecewise-manifolds are common in applications. Second. A k -manifold with boundary is a space locally homeomorphic to either Rk or Rk 1  [0.g. This does not change the dimension of the signals codomain: the system has the same number of degrees of freedom. building exteriors. and other structures that.14 Chapter 1. Yet this is not enough in practice: further generalization is needed. multi-bounce. First. A k -manifold with corners is a space.. Many of the tools and theorems of this chapter (e. is therefore a manifold of dimension k 1. that such a boundary is assembled from manifolds of various dimensions. The only change is that the codomain has certain well-mannered singularities inherited from the action of SN . The boundary of a manifold with corners no longer has the structure of a smooth manifold. : : : . Manifolds This result is greatly generalizable [?]. are piecewise-manifolds. An application of transversality theory shows that the . tangent spaces and transversality) apply with minor modications to manifolds with boundary. for some 0  m  k . The addition of signal noise further frustrates a dierential-topological approach. a manifold with boundary has m  1. In realistic settings. one may quotient out the signal codomain by the action of the symmetric group SN to model inability to identify target sources. and diraction conspire to make manifold theory less than applicable in this setting. at best.. derivatives. the assumption that D is a manifold is a poor one. A true manifold has m = 0 everywhere. Transversality and dimension-counting provide a critical bound on the number of signals needed to disambiguate position. one may modify the codomain to record dierent signal inputs. the domain has a boundary: signals are bouncing o of walls. each point of which has a neighborhood locally homeomorphic to { } x 2 Rk : xi  0. For example. where m may vary from point to point. suitably glued together. and other constructs are not dicult to generate. Signals do not propagate unendingly. independent of the types of signals used. The analogues of smoothings. with the usual compatibility conditions required for a smoothing. e. using TDOA (time dierence of arrival) merely reduces the dimension of the signal codomain by one and preserves injectivity for suciently many pulses. One approach to this last diculty is to enlarge the class of manifolds. is clear in the case of a cube I n = [0. for x 2 Rn . 1. as.g. The boundary of D. Finally. 1]n . i = 1. 1).8 Stratied spaces The application of Whitney's Theorem to signal localization in the previous subsection is questionable in practice. and the physical realities of signal reection/echo. tangent spaces. Note however. @ D. Consider the solution to a polynomial equation p (x ) = 0.

Typical examples of stratied spaces include singular solutions to polynomial or real-analytic equations. More physical examples are readily generated. Each Rn has a unique smooth structure except n = 4. Recall the setting of planar linkage conguration spaces. can nevertheless be decomposed into manifolds of various dimensions. they are. A manifold is orientable if it has an atlas such that all transition maps . The 4bar mechanism gives a 1-d manifold. Homeomorphic manifolds are not always dieomorphic. An intuitive denition of a stratied space is a space X .1. for generic choices of coecients of p . 2. such that each Xi is a manifold of dimension dim Xi = ni . even when not a true manifold. a manifold of dimension n 1. Innumerable applications call for the solution to a specic polynomial equation. Stratied spaces 15 solution set is. Upon building such a linkage. Spheres S7 have exactly 27 distinct smooth structures [?]. nature does not always deal out such conveniences. The null set of a polynomial. one can feel the dierence as it passes through a singular point. There is a hierarchy of such stratied spaces which deviate from the smooth regularity of a manifold. along with a nite partition X = [i Xi . glued together in a particular manner. In dimensions three and below. which has an uncountable number of exotic smoothings. except when the lengths satisfy L1 = L3 and L2 = L4 . however. These singularities have physical signicance: they correspond to congurations which are collinear. In this case. Notes 1. the conguration space is a pair of circles with singularities  either two or three depending on whether or not all the lengths are the same.8. Some control over how these manifolds are pieced together is needed but not integral to this discussion: a stratied space is a piecewise-manifold. It is unknown if S4 has non-standard smoothings.

wood. the author uses wooden dowels .   1 between charts preserve orientation (have determinants with positive derivative). 3. metal or plastic with pin joints works well. For planar linkages. It is instructive and highly recommended to build a complex mechanical linkage and investigate its topology à la main. The projective plane P2 and Klein bottle K 2 are well-known non-orientable surfaces. at cardboard. For 3-d linkages.

Various forms of the Transversality Theorem apply to stratied spaces [?]. the set of f : M ! N whose r -jets are transverse to W is a residual subset of C 1 (M. N ). N ). The Transversality Theorem works with the C r topology. partial dierential equations. some tuned to applications in mechanics [?. no higher spatial dimensions are available for resolving intersections between edges in spatial linkages. which allow perturbations to one domain while holding the function xed elsewhere.1. Probability theory oers complementary and. N ) of the jet bundle. The reader unfamiliar with calculus on manifolds should consult one of the many excellent introductions [?. the Taylor polynomial of f at p up to order r . integration of forms.] 1. 8. constrained by allowing motion only to turn or to move forward. 6.16 Chapter 1. Compute the homeomorphism types of C2 (R3 ) and C3 (R3 ). without trying to compute (past dimension 10. for r 2 N or r = 1. for any submanifold W  J r (M. The Jet Transversality Theorem says that. 5. This example is the basis for various musings about perceptions of musical chords [?]. for each p 2 M . incommensurate approaches. This j r f records. and more. A chord is considered to be a point in a conguration space of points on . and reections. What is the conguration space of a 4-bar linkage in R3 ? Remember to factor out the action of the Euclidean group (translations and rotations). ?].5. Tweaking the topology of the function spaces allows for relative versions. 1. Stokes' Theorem. Show that UC2 (S1 ) is homeomorphic to a Möbius band (without boundary). discernment seems to fail).4. With practice. the user can tell the dimension of the conguration space by feel. Other tools of advanced calculus follow in patterns similar to those of derivatives. only consider positions and orientations. Higher derivative data associated to a map f : M ! N is encoded in the r -jet j r f taking values in a jet bundle J r (M. often. Manifolds with latex tubing for rotational joints. There are some phenomena which are topologically generic. What happens in the degenerate cases having collinearities? 1. As always. the computations are done at the chart level and shown to be independent of coordinates. 7. walls.6. One must be careful in proving genericity results. ?]. Exercises 1. 1. Unfortunately. they are sucient to derive a form of Whitney's Theorem on embeddings and allow for unique channel response in a transmitter-receiver system outtted with corners. Can a planar mechanical linkage admit a conguration space that is non-orientable? 1. Caveat lictor. Transversality is a topological approach to genericity. Compute the dimension dim Gn k of the Grassmannian.3. One assumes a topology on the space of frequencies of notes yielding S1 (octaves are identied). as the specication of a topology on function spaces is required. but have small (even zero?) measure in a desired probability measure. including dierential forms. What is the conguration space of a unicycle riding on the plane R2 ? What is the conguration space of an airplane in R3 ? [Do not include velocities.2. The study (and even the denition) of stratied spaces is much more involved than here indicated. 4.

8.Exercises 17 a circle.12. Prove that the boundary of a manifold with boundary is a manifold of codimension one. How many transmitters must be heard (as a function of k ) to ensure disambiguation? What goes wrong in the case k = 1? [?] . naturally.10. 1.7. Use codimension and transversality (and some ridiculous physical assumptions) to argue which is the more dangerous situation in which to nd oneself: a mineeld. The codimension of a submanifold V  M is codimM W = dim M dim W . 1.9. Consider the conguration space of a ray in R3 based at the origin: it is homeomorphic to S2 . Restate Theorem 1. or a hurricane? 1.11.12 to show that a self-map of a compact manifold f : M ! M has. Instead of TOA or TDOA signals. a tornado storm. 1. 1. Remove K disjoint compact balls from R3 (none containing the origin) and consider the resulting conguration space of rays which may not intersect any of the balls. For musical chords. Modify the proof of Lemma 1. generically.13. Find examples for which the resulting conguration space is not connected. 1. these congurations are unlabeled. a nite number of xed points. 1.10 more concisely in terms of codimension. assume DOA (direction of arrival). in which the receiver in Rk detects the direction (point in Sk 1 ) of each transmitter. Use transversality to prove that polynomials with real coecients generically do not have repeated real roots.

Chapter 2 Spaces: Complexes .

1 (Statistical independence) Independence of objects occurs in multiple contexts. where V is the set of vertices (or 0-simplices) and E is a collection of distinct unordered pairs in V . The most familiar simplicial complexes are abstract graphs. These edges are the 1-simplices of the graph. simplicial complexes without simplices of dimension higher than one. The independence complex of a collection of random variables compactly encodes statistical dependencies in a topological manner. one should rapidly metabolize the formal definition and progress to drawing pictures. Any such system must be closed with respect to restriction (any subset of an independent set is also independent) and thus denes a simplicial complex: the independence complex IO is the abstract simplicial complex on vertex set O whose k -simplices are collections of k + 1 independent objects. One topologizes a simplicial complex X as . One class of elementary but illuminating examples occurs naturally in the form of independence of objects. of course. including linear independence of a collection of vectors or linear independence of solutions to linear dierential equations. three 0-simplex faces. Example 2. where j j = k + 1. s7 g is a 2-simplex with three 1-simplex faces.. Abstract graphs are ideal for expressing pairwise relations between objects. Relations of a higher order point to simplicial models. Each element  2 X is called a simplex. X  P(S ). s5 . closed from below by restriction: for each  2 X . an independence system is a collection of unordered subsets of O declared independent. As with the setting of manifolds. An abstract simplicial complex is a collection of nite subsets of S . E ). : : : . Given a nite collection of objects O = fx g. all subsets of  are also in X . Given a k -simplex . the subcollection fs2 . For example.1 Simplicial and cell complexes Let S be a discrete set and P(S ) its power set  the set of all subsets. if S = fsi gi . i.e. Complexes change in focus from spaces that resemble the smooth surfaces of advanced calculus to those assembled from simple pieces motivates the eventual transition to algebraic means of assembling spaces and invariants. 2. its faces are the simplices corresponding to all subsets of   S .) 20 Chapter 2. the probability density fX of the combined random variable (X1 . More subtle examples include statistical independence of random variables: recall that the random variables X = fXi gn 1 are statistically independent if their probability densities fXi are jointly multiplicative. An abstract graph is often presented as a pair (V. or rather a k -simplex. the empty-set face. Xn ) satises fX = ∏ i fXi . and.

X (k )  X (k 1) and one therefore identies X with 1 ∪ X = X (k ) : k =0 The abuse of notation is intentional. to be the quotient space: ( )/ ∪ ⨿ (k ) (k 1) k X = X  . and glues them together respecting faces. By construction. To topologize an abstract simplicial complex X . This ill-named geometric realization of X is performed inductively as follows. Digital cameras store data in pixels on a 2-dimensional cubical complex. Simplicial and cell complexes a quotient space built from topological simplices. A simplicial complex is one of many possible data structures for building a space. Thus. For example.1) where  is the equivalence relation that identies faces of k to the corresponding faces of  in X (j ) for j < k . 1]. 1)k +1 : k ∑ } (2. equivalent to S . attaching maps between faces identify cubes and are likewise clear. one may take the basic k -dimensional building block to be a k -cube I k . Dene the k -skeleton of X . via this restriction. The space X is given the weak topology: a subset U  X is open if and only if U \ X (k ) is open for all k . More general still is the class of spaces referred to in this text as cell complexes and in the topology literature as CW complexes. one sees how these faces are glued along the boundary of k . The faces of the cube are obviously smaller-dimensional cubes. The resulting class of cubical complexes are quite natural in many applications. one takes a copy of k for each k -simplex of X . X (0) is a discrete set. The 1-skeleton X (1) is a space homeomorphic to a collection of vertices and edges connecting vertices: a topological graph. the topological space X will be used interchangeably with the abstract simplicial complex X . Many applications in numerical and nite element analysis work well on 3-dimensional cubes (or voxels) arranged in a lattice. These have as their basis cells closed .2. The space: 21 standard k -simplex is the xi = 1 : i =0 The faces of k are copies of j for j < k via restriction of Rk +1 to coordinate subspaces.1. where I = [0. :dim =k k = { x 2 [0. k 2 N.

. There is no hope of VR and S being homeomorphic. as it is possible to arrange data points in Q so as to have a simplicial sphere Sk for arbitrary k > 1 which projects in S to a convex set: such bubbles are artifacts of the Vietoris-Rips complex and not intrinsic to the point cloud.2 Vietoris-Rips complexes and point clouds . it may be desirable to reconstruct the underlying submanifold from the point cloud. Complexes continuous functions from the boundary The reader may nd it simpler to work with cell complexes whose boundary maps are homeomorphisms of @B k to their images. as any child doing a dot-to-dot can attest. Consider a point cloud Q in Rn and the projection map  : VR ! Rn taking the vertices VR(0) to Q and taking a k -simplex of VR to the convex hull of the associated vertices in Q. Choose a constant  > 0. Note the diculty in choosing the correct  to best approximate the underlying structure. @Bk to X (k 1) . but a simplicial complex may improve matters. A graph based on a point cloud Q can be ecacious in discerning shape. The Vietoris-Rips complex of scale  on Q. prodsimplicial and -) see [?. If one receives Q as a collection of data points  a point cloud  sampled from a submanifold. 2. The remainder of this chapter is a menagerie of interesting simplicial and cellular complexes that have been found useful in applications. as the domain of the projection  is likely to have higher dimension than the codomain. Although this complex is based on concrete points in Rn . S. ?]. The degree to which a Vietoris-Rips complex VR of xed  accurately captures the topology of the point cloud is mixed. but the more general cell complexes are certainly omnipresent.22 balls B k . Even homotopy equivalence is too much to ask. This text falls back to the simplicial and cubical setting frequently. is the simplicial complex whose simplices are all those nite collections of points in Q of pairwise-distance  . The image of  in Rn is called the shadow. For other specialized types of complexes (e. of the Vietoris-Rips complex. and for attaching maps. arbitrary Chapter 2. for large-scale topological holes. Nevertheless. Consider a discrete subset Q  Rn of Euclidean space. VR (Q). there may be simplices of dimension much larger than n (when suciently many points are crowded together). the Vietoris-Rips complex seems to capture the topology correctly.g. but the reader should note the ubiquity and utility of CW/cell complexes in the literature.

Flag complexes also arise as higher-dimensional models of communications networks. the ag complex" is the maximal simplicial complex having the graph as its 1-skeleton. is an abstract simplicial complex on the vertex set L dened as follows. the witness complex provides a decent topological approximation to the underlying structure of the point cloud. higher-dimensional. A subcollection of points S  L is a simplex in the witness complex if and only if for each subset T  S there is a weak witness xT 2 Q with every point in T closer to xT than to any point in L T . whenever there appears to be the outline of a simplex. Consider a collection of points Q  R2 that represent the locations of nodes in a wireless communications " This is sometimes called a clique complex. it is necessary to list all the simplices. one lls it in. To encode a typical simplicial complex into a form that computational software can manage. especially in topological combinatorics. simplices are all accounted for: the graph determines the complex. note that the Vietoris-Rips complex is entirely specied by knowing the vertices Q and the edges (depending on ). suitable as an introduction. The remaining. Otherwise said. This is by now a multifarious subject. 2. The Vietoris-Rips complex is the ag complex associated with the distance- graph of a point cloud Q.4 Flag complexes and networks The Vietoris-Rips complex is an example of a more general class of simplicial complexes that ll a frame. Given a graph. More recent generalizations to strong and parameterized witness complexes span the gap between computability and rigorous recovery of the correct topological type [?]. The weak witness complex. The following is a elementary version. Simple examples indicate that for a reasonable choice of landmarks. with a variety of related denitions built to suit dierent data sets. The witness complex constructions of Carlsson and de Silva [?] greatly reduce size at the expense of topological precision. Given a (large) point cloud of nodes Q in a Euclidean space Rn . choose a (small) set of landmarks L  Q. Witness complexes and the sampling problem 23 2.2. WL . In contrast.3. even with the simplication to the input problem implicit in the ag complex structure.3 Witness complexes and the sampling problem The Vietoris-Rips complex of a large data set is too large to store and manipulate. . One of the advantages of this construction is its parsimonious representation with regards to input and storage of the data.

and Weinberger [?] have used this property of ƒech complexes to prove a result about randomly sampled point clouds. e. in the cross-polytopes above) appearing. it is sometimes a proper subcomplex. 2. Smale. Due to irregularities in transmission characteristics. In Chapter 4. A k -simplex of C lection of k + 1 distinct elements xi of Q such that the net intersection of diameter  balls at the xi 's is nonempty. Consider a smooth submanifold M  Rn . Given a point cloud Q in Rn and a length parameter   to be the simplicial  > 0. One way to circumvent this is to sort through the higher-dimensional simplices of the Vietoris-Rips complex more carefully. the communication links formed may not be solely a function of geometric distance between nodes.g. Niyogi.24 Chapter 2. received by nodes which are within range. These are compensated   is always homotopic to the union for by a topological accuracy: the ƒech complex C of diameter- balls about Q. Complexes network. these neighbors then establish communication links and in so doing assemble an ad hoc network. it is frustrating to have higher-dimensional features (as. Ignoring the details of the communications protocols. dene the ƒech complex C   is a colcomplex built on Q as follows. a method for addressing coverage-type problems is considered using these ag complexes.. the union of suciently sized balls about the points is homotopic to M : this union of balls can be captured by the ƒech complex of Q. It is intuitively obvious that for a suciently dense sampling. Nevertheless. forming the ag complex associated with these edges may recover some of the rough structure implicit in the global network. ambient characteristics of the domain.5 ƒech and random samplings The diculty of building topologically accurate simplicial models is classical. and signal bounce. In the context of Vietoris-Rips complexes. consider the following simple model of communication: nodes broadcast their unique identities. Let  denote the injectivity radius of M : the largest number such that all normal rays to M of length  are mutually . The ƒech complex is a subcomplex of the Vietoris-Rips complex VR . The ability of the ƒech complex to approximate M is conditional upon the density of sampling and the manner in which M is geometrically embedded in Rn . The disadvantages of a ƒech complex are its storage requirements (one cannot simply store the 1-skeleton and ll the rest in) and its construction (one must check many intersections to build the full complex). Let Q be a collection of points sampled at random from M .

g. convex subsets of Rn . Given a nite collection U = fU g of subsets of X . Theorem 2. Examples of nerves based on. Assume a collection where M Q of points on has injectivity radius that the minimal distance from any point of M to Q . =2 for  <  3=5. vertices of the nerve correspond to elements of U.2 (Niyogi-Smale-Weinberger [?]). In particular.3 (Nerve Lemma). the union of elements of If U is a nite collection of subsets of non-empty intersections of subcollections of U. Nerves and neurons nonintersecting. . edges correspond to pairs in U which intersect nontrivially. U contractible. Consider X an arbitrary topological space. suce to suggest the following classical generalization of the result for ƒech complexes:. This denition respects faces: the faces of a k -simplex are obtained by removing corresponding elements of U.6 Nerves and neurons As the Vietoris-Rips complex is a metric-ball version of the more general ag complex construction. The k -simplices of N(U) correspond to nonempty intersections of k + 1 distinct elements of U. Each such cell group is assumed to determine a specic location in space. so also is the ƒech complex the metric-ball version of a more general object. such as when Q is a sampling of points from are also available [?] and relevant to the statistics of point clouds.6.2. as follows. N(U). the authors consider the impacts of external stimuli on certain place cells in the dorsal hippocampus of rats. These cells experience dynamic electro-chemical activity which is known to strongly correlate to the rat's location in physical space. then NU X with all is homotopic to One of the more recent and interesting applications of nerves is in neuroscience. If the is less than M with noise. e. then the ƒech complex C Additional results. one builds the nerve of U. For example. a smooth compact submanifold density of  2 (Q) is homotopic to M . 25 Theorem 2. leaving the resulting intersection still nonempty. The work of Curto and Itskov [?] considers how neural activity can represent external environments.. 2. and the collection of such location patches. Q is √such M  RN .

Cell groups are identied by correlating spike-train rings that occur simultaneously (in practice. and have ring elds that are convex. forms a cover U satisfying the hypotheses of the Nerve Lemma. This. such spike trains being common data forms in neuroscience. This is an increasingly signicant constraint in biological systems. or other taxonomic data). These investigations suggest that rats may build a structured representation of their external environment within an abstract space of stimuli. n. : : : . metric tress. but the embedding of the tree into the plane is not part of the data. These time series show occasional bursts of activity. The relevant objects of study are rooted. omni-directional. assume that the label on the root is 0 and labels on the leaves are 1.7 Phylogenetic trees and links Mathematical biology is a prime target for topological methods. allows for the computation of the nerve of the place elds based on spike train correlation. Complexes or place elds. 2. The correlation of spiking activity provides the intersection data of the place elds. The following construction comes from the seminal paper [?]. The following experiment is proposed: place cells are monitored. and (2) all interior edges (those not connected to a lead) labeled with positive real numbers (and represented as the length of the edge). For simplicity. with the corresponding activity recorded as a time-series. within a window of time depending on some experimenter-chosen parameter). A tree is a graph without cycles. 2. Let Tn denote the following tree space. It suces to assume that place elds exist and are stable. The leaves of a tree are vertices of degree 1 (only one edge is attached). and a root is a dedicated leaf. In contrast to the neuroscience example above is an analogue of conguration spaces associated to genetic data in the form of phylogenetic trees  data structures for organizing and comparing taxonomic and genetic data. These trees can be readily illustrated as planar graphs. the space of rooted metric trees with n labeled (non-root) leaves. . end-labeled. Curto and Itskov argue that these assumptions are generally satised for place elds of dorsal hippocampal place cells recorded from a freely foraging rat in a familiar open eld environment. Of note is the lack of metric data  the construction of the physical environment is purely topological and can be eected without reference to coordinates. in turn. where metrics can be rare or needlessly articial. given the noisy nature of Nature. 3. This space has a natural cube complex structure as follows. All non-leaf vertices of the graphs are assumed to be essential in that the number of edges attached is greater than two. phenotypes.26 Chapter 2. A phylogenetic tree is a rooted tree with (1) leaves labeled by data (genes.

The link of v in X . reparameterize all interior edge lengths to the open interval (0. by shrinking each interior edge continuously to length zero. Tn may seem too simple: it is contractible to an origin. the radial graph with no interior edges.4 (Tree spaces) The tree space T3 is a letter Y. a nite cube complex Tn is obtained. and let v 2 X (0) be a 0-cell. `kX (v ). To compactify this to a closed cube. Let X denote a cell complex built from simplices. the leaf labels.7. Vogtmann [?]). Two trees will be considered isomorphic if there is a mapping between them that preserves the graph structure and the labels. Despite this simplicity. Phylogenetic trees and links 27 Two such trees are equivalent if there is a mapping between them that preserves the graph structure (takes vertices to vertices and edges to edges). and the interior edge lengths.2. origin in is a ag complex that has the homotopy type of (n 1)! The link of the 3 copies of Sn glued together at a single point. with faces inherited from faces of cells in X . As topological spaces. Those faces with one or more 0 factors may be thought of as having the corresponding interior edges collapsed (zero length edges). or other simple polyhedral basis cells. The best way to analyze this assembly is through a simplical model called a link. the manner in which the various cubes are assembled is topologically (and geometrically) of interest. Example 2. . These edge-collapse faces are identiable as trees of a dierent isomorphism class. 1) and add the boundary faces. cubes. The space T4 consists of 15 2-dimensional squares (corresponding to the 15 possible binary rooted trees with 4 labeled leaves) glued together so as to have a common corner (corresponding to the radial rooted tree with 4 leaves) and a link equal to the petersen graph. Tn Theorem 2. is the simplicial complex whose k simplices consists of (k + 1)-dimensional cells attached to v . A xed isomorphism class of phylogenetic trees is parameterized by the interior edge lengths: this yields an open cube of dimension equal to the number of interior edges. By gluing together all such cubes according to isomorphism class.5 (Billera-Holmes-Vogtmann [?].

stepping onto a ight bound for Newark does not guarantee arrival at Newark. G . as the abstract simplicial complex dened on the set of actions E (G). in which k -simplices consist of collections of k +1 disjoint actions (including all branches of any chosen nondeterministic action) whose unions contain . The question of path-planning within a nondeterministic transition graph is one amenable to simplicial data structures. The tree spaces Tn (though arising in other contexts in topology and algebraic geometry as certain moduli spaces) were generated in response to challenges in statistics associated to genetic data [?]. but once chosen. and whose edges E represent deterministic actions. 2. the outcome is one of several possibilities. however. the notion of a guaranteed strategy for reaching a goal state despite adversarial meddling seems dicult if not impossible. and this data does not spring forth fulled formed from the mind of the researcher: the data are often noisy and not perfectly repeatable. complexes also have broad applicability in engineering systems as spaces that collate states.28 Chapter 2. A collection of numerical or vector-valued data can be averaged. Planning on the transition graph is straightforward: nd a directed path from initial to goal. In such a setting. but only after the nondeterministic action has been chosen and executed. These possibilities are not necessarily chosen probabilistically. For concreteness. where vertices are locations and edges represent ights. Consider a transition graph. Ill weather and air travel does not exactly t that scheme. a directed graph G whose vertices V represent states of an abstract system. then execute the appropriate actions sequentially. Experimental data is used to generate phylogenetic trees. if something may go amiss. but what does it mean to average a sequence of phylogenetic trees in a scientically meaningful manner? This requires a geometry (and thus a topology) on the space of all possible phylogenetic trees. an adversary may determine any outcome among the terminal edges. Given a nondeterministic transition graph G. Complexes There are numerous reasons why scientists and engineers might want to work with spaces of objects rather than merely the objects themselves. However. Note that the state chosen by the adversary is known to the game-player. but other situations do exhibit adversarial forms of uncertainty for which. Worse still are the kinds of uncertainties precipitated by an adversary. it must be mitigated. subways.8 Strategy complexes and planning with uncertainty Many of the complexes of this chapter approximate manifolds or provide models of data. dene the strategy complex. trains. out of the control of the chooser. Erdmann [?] models adversarial uncertainty in the transition graph via branched edges. or other discrete enter-exit modes of transport. consider the example [?] of a transportation network. Other examples include motion-planning in robotics with discrete states (robots move from landmark to landmark). freeways. Actions may be deterministically chosen.

2. Hidden beneath these 1-dimensional models is a rich higher-dimensional topological structure underlying inuence. incorporation of stochastic nondeterminism.9 Decision tasks and consensus A related use of simplicial structures assists in multi-agent decision-making. Consider the problems of consensus or distribution among collaborative agents. else it is contractible.6 (Erdmann [?]). Each n-simplex encodes the setting in which n + 1 of the transmitters have cooperated to share specic chan- . but the path is not. Note: the goal is certain. This relation is closed under restriction (a smaller collection of such edges clearly also contains no directed cycle). Example 2.9. consensus. Decision tasks and consensus no directed cycles 29 . The obstruction to attainability lies in the strategy loopback complex: denote by G g the strategy complex of G augmented by deterministic arrows from g to each v 2 V (G) g . non-interfering channels. Build a simplicial complex on the vertex set corresponding to transmitter i choosing channel j . say that G has a complete strategy for attaining g if for N large there exists a mapping from V (G) to E (G) which associates to v an action based at v . engineering systems (network clocks). and robotics (cooperative action). and is sensible. where simplices (up to and including dimension N 1) correspond to several transmitters having made distinct choices. and decompositions of strategy complexes by means of deterministic subsystems. Additional results include realizability of simplicial complexes as strategy complexes. g 2 V (G) i G G possesses a g is homotopic to a sphere of The proofs of this and related theorems in [?] use nerves extensively. sociology (voting. Theorem 2. Recent interest in synchronization phenomena [] focus on graph-theoretical methods. The problem of collaborative consensus has a large literature: mathematical approaches depend sensitively on the modeling assumptions. as a directed cycle means that a skilled adversary can cycle the player through states ad innitum. The intuition is that G consists of trap-free strategies: each simplex of G is a strategy for progressing to one or more states.7 Consider the problem of N transmitters attempting to broadcast without interference over a frequency spectrum with K  N distinct. dimension A nondeterministic transition graph complete strategy for attaining a goal jV (G)j 2. 2. For g 2 V (G) a goal state. a common scenario in biology (animal ocking). so that one attains g with the execution of at most N actions. and division. beliefs).

30 Chapter 2. it is presumed that there are N independent processors Pi who share a common read/write memory element for communication purposes. dene the input complex I to be the abstract simplicial complex whose k -simplices are labelings of the states of k + 1 distinct processors. The output complex O is the abstract simplicial complex whose k -simplices are labelings of the states of k + 1 distinct processors which are consistent with a desired output. while the other N n 1 transmitters are viewed as having failed/declined to choose. A fault tolerant distributed algorithm is one which will terminate in nite time even in the event that some (but not all) processors fail irrevocably. The approached used in [?] begins with a simplicial model of the problem. then the putative consensus algorithm must of necessity output that precise label as the consensus state. These simplices form a complex using the natural boundary map derived from transmitters `forgetting' a choice. For example. A decision task consists of the input and output complexes. Because of the relationship between the simplices. The problem is whether there exists an algorithm for the processors. . For simplicity and concreteness. three transmitters choosing from four channels gives rise to a simplicial 2-torus T2 . The problem in reasoning about such systems is that processors which take a very long time to reach decisions are not readily distinguished from processors which have failed and will never respond. Given N processors with M possible initial states. faulttolerant computation. it is possible to reduce questions of existence of distributed fault-tolerant consensus algorithms in terms of the existence of a map from the input complex to the output complex. and failures. Complexes nels. along with a set of constraints. if all N processors begin in consensus with the same label. The intuition behind a k -simplex is that k + 1 of the processors exhibit a legal system state and the N (k + 1) remaining processors have failed to report. via read/write communication to the common memory. to come to deterministic nite-time consensus. a processor crashes and can engage no further in communication). In this setting. Because of this. The interesting catch is that all computations are asynchronous (each processor works at its own speed) and faulty (sometimes. faces. For example. the topology of the complexes (encoded by the simplex faces) correlates with failure of processors. assume that there are M > 1 possible labels or states and that each processor has an initial preference. Work of Herlihy and Shavit [?] and their collaborators provides a topological approach to consensus/decision problems in the context of distributed.

Note that Dn (G) is a subcomplex of the cubical complex Gn and a subset of Cn (G) (it does not contain partial cells that arise when cutting along the diagonal). This can be shown by other. Thus. Dene the discretized conguration space of G as: ~. removing an  neighborhood of ). Keeping track of vehicles or robots or tokens on G leads. Dn (G) is the set of congurations for which.10. one can think of the vertices (0-cells) of Dn (G) as discretized congurations  arrangements of labeled tokens at the vertices of the .e.2. and is. Equivalently. any wait-free fault-tolerant protocol would have to induce a surjective continuous map from I to O.. One of the key theorems [?] gives necessary and sucient conditions for consensus stated in terms of existence of maps (with details concerning simplicial subdivisions and colorings). two labels. with the proviso that initial consensus terminates the program immediately. With this natural cell structure. but rather as a physical.10 Discretized graph conguration spaces Consider a nite graph G. guidewires. instead of restricting tokens to be at least some intrinsic distance  apart (i. There is no distributed asynchronous fault-tolerant solution to the problem of binary consensus. perhaps more direct means. the largest subcomplex of Gn that does not intersect . This is impossible. The decision task is to come to consensus. The salient feature of the result is that the topology of the decision task can determine whether an algorithm exists. the Herlihy-Shavit theorem has resolved more complex consensus problems for which other solutions were unknown.. however. ?] are approximations to these spaces by cubical complexes. or optical paths [?]. the input complex I ' SN 1 is a simplicial sphere and ⨿ N 1 the output complex O =  N 1 is a disjoint pair of simplices: all zeros and all ones. the path contains at least one entire (open) edge.8 In the case of N processors trying to come to binary consensus (i. 2. These deeper examples rely not on common-sense notions of connectivity but rather on holes of higher order  the homology theory of Chapter 4.2) ~ denotes the set of all product cells in Gn = G  where     G whose closures intersect the topological diagonal . since I is connected and O is not. This type of implicit inference is a hallmark of topological methods. as in Ÿ1. given any two tokens on G and any path in G connecting them. in fact. however. one now restricts tokens on G to be at least one full edge apart. Discretized graph conguration spaces 31 Example 2. now thought of not as an abstract system of states. Cn (G) is in general neither a manifold nor a simplicial/cell complex. to conguration spaces. Thus.e. zero and one). geometric highway of paths. Cn (G) is important for motion planning in robotics when the automated vehicles are constrained by tracks.5. The following constructions [?. Dn ( G ) = ( G      G )  (2. The conguration space of n labeled points on G.

is a 2-dimensional cube complex. For each edge e incident to v . Example 2. the token not implicated by e can move to two of the three available vertices without interfering with e . providing discrete localization of droplets via electrowetting  a process that exploits current-induced dynamic surface tension eects to propel and position a droplet.10 (A nonplanar graph) There are several interesting examples of graph . Complexes graph. In digital microuidics. one can manipulate many droplets in parallel. The edges of Dn (G). arranged cyclically as shown. conguration spaces whose structures are illuminated by discretization. and D2 (K5 ) is thus a topological 2-manifold: a compact surface. thus. There are a total of six such 2-cells incident to v . Fix a single vertex v 2 D2 (K5 ): this corresponds to a pair of distinct labeled tokens on vertices of K5 .. the work of Fair in [?. With the tools of the next chapter at our disposal. it will be shown that this surface is of genus six. The discretized conguration space is an accurate cubical model of the topological conguration space. e is incident to exactly two 2-cells of D2 (K5 ). Example 2. or 1-cells. e. the complex looks the same at each vertex. Using the grid. The discretized conguration space D2 (K5 ) of two labeled points on a complete connected graph of ve vertices.g. Note. K5 . From this vertex v 2 D2 (K5 ) emanate six edges: each token can move to any of the three non-occupied vertices of K5 . The goal of this and related microuidics research is to create an ecient lab on a chip in which droplets of various chemicals or liquid suspensions can be positioned. mixed. ?]). however. Each 2-cell in Dn (G) represents two physically independent edges: one can move a pair of tokens independently along disjoint edges. The plates are embedded with a grid of wires. A k -cell in Dn (G) likewise represents the ability to move k tokens along k closure-disjoint edges in G in an independent manner. tell us which discrete congurations can be connected by moving one token along an edge of G. and then directed to the appropriate outputs. Thanks to the symmetry of K5 .32 Chapter 2. that collisions are not always to be avoided  chemical reactions depend on controlled collisions. Applying an appropriate current drives the droplet a discrete distance along the wire grid. The discretized conguration space of droplets on the graph of the electrical grid captures the topology of the multi-droplet coordination problem.9 (Digital microuidics) An excellent physical instantiation of the discretized conguration space is implied by work on digital microuidics (see. small (1mm diameter) droplets of uid can be quickly and accurately manipulated in an inert oil suspension between two plates.

used to label the vertices V (G). An alphabet A = f0. Local recongurations are supported on the closure of each edge in G. Then Dn (G0 ) is homotopic to Cn (G0 ) = 2. each generator  has supp = trace equal to an edge with local state a labeling of one vertex with 0 and the other 0 < i  N for some i .2. DN (G). The trace of a generator is the (nonempty) subset of the support trace  supp on which the local states dier. Fix a graph G and an alphabet of labels A. A simple familiar example is that of discretized motion-planning on a graph from Ÿ2. called state complexes. are cube complexes with interesting topological and geometric properties. so as to split each edge of Let connecting a vertex to itself ) and let G into n G be a nite topological graph (with no edges G0 be the subdivision of G that inserts vertices 1 edges.11. : : : . 1.10. This system has a well-dened notion of a conguration space. closed under the actions of a xed set of local recongurations. A collection of generators f g admissible at a given state u is said to commute if their supports and traces are pairwise non-intersecting: trace \ supp. labelings of the vertex set of supp by elements of A. the other local state reversing these labelings. The resulting conguration spaces. N g labels the vertices of G as empty (0) or occupied with a (numbered) token.11 (Abrams [?]). State complexes and reconguration 33 Theorem 2. that generalizes to any recongurable system as follows. A generator is admissible at a given state u if one of the generator's local states matches the restriction of u to supp .11 State complexes and reconguration Discretized conguration spaces of graphs are generalizable to a very broad class of systems  ranging from metamorphic robots to protein chains  which are recongurable based on local rules. each such labeling V (G) ! A dening a state u. or generators. Cn (G). Each generator  consists of a support subgraph supp  G and an unordered pair of local states. A recongurable system is a collection of states fu g.

= ? 8 6= .

This arm is positive in the sense that it may extend up and to the . etc. That is. the 1-skeleton is the graph whose edges pair two states which dier by the local states of a generator. For recongurable systems that admit only nitely many admissible generators at any given state. the state complex S is a locally compact cubical complex. Example 2. 2-cells correspond to commutative pairs of generators acting on a state. the 0-skeleton consists of the states in the recongurable system.: This means that the changes implicated by a generator (its trace) do not impact the applicability of the other generators (their supports): commutativity encodes physical independence. Dene the state complex of a recongurable system to be the cube complex S with an abstract k -cube for each collection of k admissible commuting generators.12 (Articulated planar robot arm) This recongurable system models the position of an articulated robotic arm with one end xed at the origin and which can (1) rotate at the end and (2) ip a 2-segment corner from up-right to right-up and vice versa.

and its generator corresponds to sliding the robot Ri from one end of the edge to the other. let the underlying ⨿ graph be the disjoint union G = i Gi . let G be a linear graph on n vertices with an alphabet A = fU. Consider a collection of n planar graphs (Gi )n 1 . Specically. although the transition graph (the 1-skeleton of S) for this system is complicated. each embedded in the plane of a common workspace (with intersections permitted). The state complex in the case n = 5 has cubes of dimension at most three. To model this with a recongurable system. trace( ) = . For each edge 2 E (Gi ). Generators commute when the associated subwords are disjoint. Complexes right only. Supports (and traces) are (1) a closed edge. Generators are of two types: (1) exchange the subwords UR and RU along an edge. Rg encoding up and right respectively. and (2) the terminal vertex. the state complex itself is contractible.34 Chapter 2. A grammatical model is simplest: states are length n words in two letters. coordination problems. there is exactly one generator  . The generators for this system are as follows. On each Gi . In this case also. respectively. The ∏coordination space of this system is dened to be the space of all congurations in i Gi for which there are no collisions  the geometric robots Ri have no overlaps in the workspace. supp( ) consists of this trace union any other edges . The trace is the (closure of the) edge itself. The support. (2) change the terminal letter in the word. a robot Ri with some particular xed size/shape is free to translate along edges of Gi .

in Gj (j 6= i ) for which the robot Rj sliding along the edge .

The resulting state complex is a cubical complex that approximates the coordination space. can collide with Ri as it slides along . this recongurable system has S = Dn (G). Any state of this recongurable system has all vertices labeled with zeros except for one vertex per Gi with a label 1.13 (Robot path coordination) The following example is inspired by robot . It appears to be dicult to characterize which cubical complexes can arise as Example 2. except for a single 1 at the boundary vertices of . The alphabet is A = f0. In the case where Gi = G is the same for all i and the robots Ri are suciently small. 1g and the local states for  have zeros at all vertices of all edges in the support.

2 have been applied to computing the topological type of data sets in reaction-diusion systems [ ]. 4. They may also be viewed as weak versions of the very interesting class of combinatorial cell complexes known as ? high-dimensional automata of Pratt hom complexes. The literature would seem to benet from the language of simplices. 8. State complexes are undirected versions of the ? [ ]. ? ?] that a suciently dense sampling recovers the topology of The bounds in Theorem 2. it is impossible to obtain a sphere Sn for n > 1. 7. Notes 1. State complexes and reconguration 35 state complexes for some recongurable system: the problem is closely entwined with deep ideas in geometric group theory [?]. Theorem 2.1. prior to the work of ƒech. This statement is true but should propel the reader to more sophisticated methods rather than to despair. The Nerve Lemma is usually attributed to Leray [ ]. Many results about the topology of state complexes are highly dependent on their geometric features [ ]. The Nerve Lemma is ubiquitous in topology and its applications. Alexandrov investigated spaces like ƒech complexes in [ ]. A few results are known: e. 3. One popular approach is the alpha complex of Edelsbrunner [?] which restricts a ƒech complex by a Delaunay triangulation to produce a topologically accurate simplicial model with computationally small footprint. there are numerous ways to improve on the Vietoris-Rips construction. indeed. 6. is much deeper than that stated.2 is remarkable for its explicit computation of density bounds  it was previously known [ . anything resembling a sphere. # The name apparently is meant to connote graph-based as opposed to pictorial. One virtue of topological methods is their coordinate-free nature: it is not necessary to know exact locations. the directedness associated to conditional independence may explain the convoluted terminology there employed. ?]). however. A recent survey article [ ] gives a broad treatment of applications of geometric and topological combinatorics to a broad class of problems in phylogenetics. . The monograph of Koslov [ ] has ? a wealth of information on this latter class of spaces. the submanifold. 5. however.11. whose version.14 ([?. that It is often stated ? U be comprised of convex sets. Theorem 2. This is a very active branch of applied topology and geometry.2. Work on ? ? ? ? graphical methods# is an attempt to encode statistical independence that is reminiscent of Example 2. one does possess strong geometric data in the form of coordinates.g. as these and all nonempty intersections thereof are contractible. concerning sheaves. or. n > 1 into S is homotopic to a constant map State complexes are aspherical: Sn !? any map of a sphere Sn for . It will sometimes be mentioned that the Nerve Lemma breaks down if the sets involved are not contractible. If. which echoes antecedents in the work of ƒech [ ] and Alexandrov [ ]. Theorem ? ? ?? follows from some simple observations combined with the powerful geometric tools of Gromov [ ]. 2..

3. Is it possible to have a nondeterministic graph whose every action is nondeterministic but which nevertheless has a complete strategy for attaining a goal state? 2. n = 6 as the 1-skeleton of a polytope.6.4. for any length robot G arm. Complexes Exercises 2. R 2 its whose interiors intersect 2. . the discretized conguration space of complete bipartite graph K3.3. and whose Draw this edges are given by ipping the diagonal edge of two incident triangles. 4. for the hexapod insectoid RHex [ ]. this starts to look like a conguration space of a smooth curve in the plane with nonnegative partial derivatives. Dene an abstract simplicial complex for a legged robot based on legs o-the-ground and compute what this complex is in the case of RHex. two labeled points on the 2. Show that 2. 5g. Show that the ƒech and Vietoris-Rips complexes t together: for any point set Rn . Show that the state complex of Example ?? is always contractible.9. is a topological 2-manifold. Consider the following recongurable system: ve edges (a pentagon). with the desired outcome being neither consensus (all the same label) nor diversity (all three labels represented among choices) but in between (exactly two labels are chosen by the processors).11. 2.12. To remain in a stable standing position. Consider four compact convex sets in R2 having the property that any three sets Does intersect nontrivially.13. Consider the graph whose vertices are triangulations of a regular graph in the case n-gon. with local states that exchange labels on vertices. is a cyclic graph with ve nodes and and each state is a bijection of nodes to generators have support and trace equal to an edge of the graph. For example. Assume a legged robot may lift each leg independently. 2. 2. Show that if in their ? R2 and let VR be the Vietoris-Rips complex and S S image. Q    (Q)  VR2 (Q)  C  2 (Q): C What improvement to the constants can be made if n = 2? (see. 2 Show that D (K3. and that every simplicial complex can be subdivided to give a cubical complex. Consider four points in shadow.. In the limit as length goes to innity.5.8. Show that every cubical complex can be subdivided to give a simplicial complex. VR contains two disjoint closed edges then VR contains at least one 2-simplex. What are the possible nerves of this collection of sets? the Nerve Theorem help you deduce whether or not all four sets intersect nontrivially? 2. 2. 2. A.g. Let Y denote a Y-graph (a graph with three edges emanating from a central vertex). [ ]) ? 2. Show that C 2 (Y ) embeds in R3 and draw a picture of it. A = f1. D3 (Y) cannot be homotopic to C3 (Y).36 Chapter 2. 2.1. e.10. D2 (Y)? How does it compare to the discretized conguration space 2. What are the input and output complexes associated to 3 processors and 3 labels. certain combinations of legs must remain on the ground.3 .2. there is one such generator per edge and per vertex labeling.3 ). Show that the state complex of this system is an orientable surface.7. at least three legs with at least one per side must remain on the ground.

Chapter 3 Euler Characteristic .

1]) = 1. Euler Characteristic he simplest and most elegant non-obvious topological invariant is the Euler characteristic. the Euler characteristic is its cardinality (X ) = jX j. This intuition of counting with weights inspires the combinatorial denition of Euler characteristic. such is a tree (a contractible graph) if and only if  = 1. Example 3.1) This quantity is well-dened for nite cellular objects and independent of the decomposition of X into cells:  is a homeomorphism invariant... Given a space X and a decompo⨿ sition of X into a nite number of open cells X =  . for certain examples. discrete set X . e. an integer-valued invariant of spaces. it distinguishes ((0. suggesting that 2-cells be weighted with +1. 3. the addition of an edge does not change the count of connected components. This chapter is comprised of observations and applications of this invariant. 1)) = 1 from ([0. Connect two points together by means of an edge (in a cellular/simplicial structure). e. the Euler characteristic is decremented by one. Removing k disjoint convex open sets from such a convex superset results in a compact space with Euler characteristic 1 k ( 1)n . Note that this occurs precisely when a cycle is formed. it is clearly $ This is sometimes called the combinatorial or geometric or even motivic Euler characteristic. It is not a homotopy invariant for non-compact cell complexes. Given a nite. as will be shown in Chapter 5. Euler characteristic is also sharp invariant among connected compact orientable 2-manifolds:  = 2 2g . This intuition of counting connected components works at rst. meant to serve as motivation for the algebraic tools to follow that justify its remarkable properties. it is a homotopy invariant. . Continuing inductively. where each k -cell  is homeomorphic to Rk . To ll in such a cycle in the gure with a 2-cell would return to the setting of counting connected components again.1 (Conguration spaces of graphs) Recall from Example 2. however.g. Any compact convex subset of Rn has  = 1. Among compact nite cell complexes. Euler characteristic can determine the homotopy type of a compact connected graph. as.10 that the discrete conguration space D2 (K5 ) of two points on a complete connected graph of ve vertices is a cubical 2-manifold. the Euler characteristic$ of X is dened as ∑ (X ) = ( 1)dim  : (3. Being built from nitely many cells.38 6 Chapter 3.g. as the resulting space has one fewer component. where g equals the genus.1 Counting The Euler characteristic is a generalization of counting. the Euler characteristic counts vertices with weight +1 and edges with weight 1.

Curvature 39 compact. With this interpretation. 2.  as counting 3. (3. Thus. This change is local. not global. Theorem 3. (D2 (K5 )) = 20 60 + 30 = 10 and this surface has genus g = 1 1  = 6. perhaps with boundary a piecewise-smooth curve(s). as the reader may check. piecewise smooth over a cell structure. To determine the genus of this conguration space.2) ∫ M equals M . but the curvature measure. but stretching and deformation of M change . 2(M ) equals the integral of a curvature function over M . 1. d means kg ds . On 2-cells of M . like M . 2 The remainder of this chapter complements the interpretation of with more geometric. Edges correspond to pairs of one closed edge and one disjoint vertex: there are (2)(10)(5 2) = 60 such. is stratied. plus the integral over 0-cells of angle defect equals 2(M ). Note that the curvature is a geometric quantity: rigid translations and rotations leave it invariant. the integral over the 2-cells of Gauss curvature plus the integral over 1-cells of geodesic curvature. Faces correspond to ordered pairs of closure-disjoint edges in K5 .3) M M (0) M (1) M (2) i.2 (Gauss-Bonnet). one computes the Euler characteristic.3. the Gauss-Bonnet formula can be ∫ ∫ ∫ ∫ d = d + d + d = 2(M ). oriented surface in respect to area on R For 3 M a compact smooth (3. There are several corollaries of this result relevant to discrete and dierential geometry: written as .e. 3. the integral of Gauss curvature with  dA = 2(M ): That this is the beginning of a much larger story is evidenced by the following mild generalization. It is also orientable and connected. dynamical..2 Curvature A blend of Euler characteristics and integration is prevalent in geometry. On 1-cells of M . The classical result that initiated the subject is the Gauss-Bonnet Theorem. On 0-cells of M . Gauss curvature times the area element.2. of which there are (5)(4) = 20. d means  dA. Let M be an oriented surface in R3 . The Gauss map is the map : M ! S2 that associates to each point of M the direction of the unit vector normal to M in R3 . and analytic perspectives. The Gauss curvature  = detD is the determinant of the derivative of the Gauss map. geodesic curvature times the length element. Vertices of D2 (K5 ) correspond to ordered pairs of distinct vertices of K5 . d means the angle defect. Let M be a smooth surface embedded in R3 . of which there are (10)(3) = 30. Again.

this does imply that there are no xed-point-free vector elds on a 2sphere. closed surface has total (integrated) Gauss curvature constant. A smooth. Gauss-Bonnet implies that the 2 surface has Euler characteristic  = 20(  )=2 = 10. from Example 2.3 Obstructions to vector elds Euler characteristic has many uses. Recall. Euler Characteristic 1. all faces and edges are at. if A connected compact manifold M without boundary possesses a nonvanishing vector eld if and only (M ) = 0. locally at but for the vertices.10. Theorem 3. since all compact manifolds in these dimensions have Euler characteristic zero. to wit. Invoking transversality and rening as needed. Placing a at Euclidean metric on each square yields a space with an intrinsic geometry. you can't comb the hair on a ball at. one of which is as an obstruction to the existence of certain vector elds. 2. As the cells are suciently small and the vector eld  is nonvanishing. since (S2 ) = 2. The reader may rightly suspect whether an embedding in R3 is at all needful. 3. 2 minus the sum of the face angles) equals 4 . Recall that a vector eld is a continuous (or smooth) assignment of a tangent vector to each point in a manifold. On each n-dimensional cell  . This is not terribly useful in dimensions 1 and 3.3. d vanishes along the geodesic edges and the sum of the angles of the triangle equals  plus the integral of Gauss curvature over the triangle face.  is approximately linear in a neighborhood of each cell. consider the set of faces % This also is the source of the moniker Hairy Ball Theorem. . Assume that M is a compact closed n-manifold that admits a smooth nonvanishing vector eld . The easiest proof without invoking advanced tools involves imposing a dense cellular mesh and working cell-by-cell. 3. perturb the cell structure so that the vector eld  is transverse to all cells. However. (M ) = (S2 ) = 2. Place a smooth cell structure on M with all cells dieomorphic to convex polytopes of suciently small diameter. (sketch) The interesting direction (only if) reveals the obstructive nature of . no matter how the surface is deformed. the cubical 2-manifolds built from arranging six squares around each of 20 vertices. assuming smoothness. For a geodesic triangle. a fact manifestly clear to meteorologists. again.% Proof.40 Chapter 3. This recovers classical notions of angle-sums for triangles on spheres and other curved surfaces. Since each vertex has angle defect 2 6  =  . For M the boundary of a convex polyhedron in R3 . and the sum of the vertex angle defects (in this case.

12. sources and sinks have index +1. y ) coordinates based at p . let D be a disc whose boundary is @D = . IV (p ) is dened to be the following line integral: IV (p ) = where (V ) is the angle made by V in local coordinates. Then.4 (Population dynamics) Consider the following dierential equation model x _ = 3x x 2 2xy . This index is well-dened and independent of (a suciently small) Bp . 2 (3. 2 _ +y _2 Example 3. ∑ IV ( ) = IV (p ): (3.5) p2D\Fix(V ) of competing species x (t ). saddle points have index 1. One argues (in a manner not unlike those used in Green's Theorem or in contour integration) that index is additive over Fix(V ). y _ = 2y xy y 2 : (3. Index theory 41 By transversality. Let Bp denote a suciently small ball about p with boundary = @Bp . For  R2 Fix(V ) a simple closed curve.4 Index theory Consider. it will be possible to drop the assumptions about smoothness and manifold structure: see Theorem 5.4) extends to dene an index IV ( ) to any closed curve which avoids Fix(V ). y (t ): 1 d(V ). The index of V at p . Let p 2 Fix(V ) be an isolated xed point of V . It is clear that Equation (3. for simplicity. With better tools (from Chapters 4-5). I () on which  points to the interior of . thanks to Green's Theorem.3. Thus.4) x _ dy y _ dx d = x .4.6) . the vector eld is of the form V = (x. This invocation of transversality is admittedly glib. since the sets I ( ) partition all cells of M . a vector eld V on an orientable 2-manifold . More specically. (I ()) = ( 1)n +( 1)n 1 = 0 and (M ) =  (I ( )) = 0. Among the nondegenerate xed points. the union of which (by linearity of ) is homeomorphic to the open disc ∑ Dn 1 . _ y _ )T . then the integrand d is so that IV (p ) represents the (signed integer) number of turns the vector makes in a small curve about p . 3. I ( ) is nonempty for each top-dimensional cell  and contains both the interior cell of  (homeomorphic to an open disc Dn ) along with some boundary faces. if in (x.

say. Though this is homeomorphic to an interval. 3. The following classical theorem is a hint of these deeper connections: Theorem 3. Besides the lack of an explicit cell structure. Consider. The closure of this set in R2 is not homeomorphic to the closed .7) The proofs of Theorems 3. eld V For a continuous vector with isolated xed points on a compact manifold ∑ Fix(V ) M. Linearization reveals that the coexistence solution is a saddle point. 2) [species y dominates].4) equal IV ( ) = +1. worse things can occur when presented with a space ab initio. some (sub)spaces do not come with an externally imposed cell decomposition. To complete this integration theory requires an excursion to consider exactly which sets have a well-dened Euler characteristic. 1) = 1. 0) [species x dominates]. and (1. Though an explicit cell structure is often present in. This integrative technique motivates an extension of Euler characteristic from sets to certain functions over sets: an integration theory. must surround a collection of xed points whose indices sum to +1. by Equation (3. the simplicial setting. (3.5). thus. 1) [coexistence]. or the level set of a smooth function f : Rn ! R.5 (Poincaré-Hopf).42 Chapter 3.3 and 3. a conguration space of points in a domain. Euler Characteristic The equilibrium solutions consist of (0. (0.5 compute  locally (on cells and xed points respectively). 0) [mutual death]. since is everywhere tangent to the vector eld. The rationale for this equation (and the proper denition of the index of a vector eld in all dimensions) will become clearer in Chapters 4-5.5 Tame topology Euler characteristic is well-dened for spaces having a decomposition into a nite number of cells. The additivity present in Equation (3. The index of a periodic orbit must by Equation (3. IV (p ) = (M ): (3. then it cannot intersect the xed point set or the x or y axes. and the only remaining enclosable xed point has negative index. as with the graph of sin(1=x ) for 0 < x < 1. then add up this local data to return the global . It is clear (from the equations and from the interpretation) that the x and y axes are invariant sets. Is there a periodic orbit? No. If (t ) is a nontrivial periodic solution. for example.5) is reminiscent of the Euler characteristic. However. I(1. it is a wild type of equivalence.

Tame mappings are likewise easily dened: a (not necessarily continuous) function between tame spaces is tame (or denable) if its graph (in the product of domain and range) is a tame set. This result implies that tame sets always have a well-dened Euler characteristic. Given a xed o-minimal structure. O is closed under axis-aligned projections Rn ! Rn 1 .g. disjoint union of open standard simplices. as well as a well-dened dimension (the max of the dimensions of the simplices in a triangulation). more properly. or the closure of another open simplex in the triangulation. semialgebraic sets.6 (Triangulation Theorem). O1 consists of all nite unions of points and open intervals. To repeat: denable homeomorphisms are not necessarily continuous. . one can work with tame sets with relative ease. one wants to rule out such oddities and focus on spaces (and mappings) that are for all intents and purposes tame. 1]. The text of Van den Dries [?] is a beautifully clear reference. In applications. describable in terms of ane sets and matrix inequalities. in turn coming from model theory. The intersection of this set with the x -axis is an innite set of points. e.& An o-minimal structure O = fOn g is a sequence of Boolean algebras An of subsets of Rn (families of sets closed under the operations of intersection and complement) which satises certain axioms: 1. denable sets. & The term derives from order minimal. they are complete invariants. The surprise is that these two quantities are not only topological invariants with respect to denable homeomorphism. piecewise-linear sets form an acceptable o-minimal structure. Such a convention makes the following theorem concise: Theorem 3. Logicians have created an axiomatic reduction of such classes of sets in the form of an o-minimal structure.3.: piecewise linear (PL) spaces. expressible in terms of a nite set of polynomial inequalities. 2. Some deeper results of o-minimal geometry require having at least all algebraic sets [?]. but for our purposes. Elements of O are called tame or. Canonical examples of o-minimal systems include semialgebraic sets and subanalytic sets. dened in terms of images of analytic mappings [?]. O is closed under cross products. Note the last axiom: this niteness is the crucial piece that drives the theory. 3. Dierent mathematical communities focus on. A denable homeomorphism is a tame bijection between tame sets.5. Tame topology 43 interval [0. Any tame set is denably homeomorphic to a nite The intersection of the closures of any two simplices in this denable triangulation is either empty. or subanalytic sets.

the reader will not be surprised to learn of a deep relationship between tame sets. The Euler integral is dened to be the homomorphism X : CF(X ) ! Z given by: ∫ 1 ∑ h d = s(fh = s g): (3. and integration. This result reinforces the idea of a denable homeomorphism as a scissors equivalence. until the homological tools Chapters 4-5 are available. in the proof of Theorem 3. morphic if and only if they have the same dimension and Euler characteristic.4.8) and the denition of a measure is not a coincidence.8) (A [ B) = (A) + (B) (A \ B): for A  X denable. using the denition of tame sets. Recalling the utility of such scissors work in computing areas of planar sets.8. and in the additivity of the xed point index of Ÿ3.7 ([?]). the abuse of terminology is to prompt the reader to think explicitly in terms of integration theory.10) X s= 1 Alternately.1.11) h d = c ( ) = c ( 1)dim  X ' The proper term is a valuation. then ∫ ∑ ∑ (3. in the Gauss-Bonnet Theorem of Ÿ3. One is permitted to cut and rearrange a space with abandon.2. one may write h 2 CF(X ) as h = ∑ c  . The integral in this calculus depends on the following additivity principle: Lemma 3. the Euler characteristic. The reader may prove this via triangulation and induction. For A and B denable sets. Following ideas that date back to Blaschke (and likely earlier).44 Chapter 3. Denote by CF(X ) the set of bounded compactly supported ∫ constructible functions on X . The measurable functions in this integration theory are integervalued and constructible.9) X .3. (3.6 Euler calculus It is possible to build a topological calculus based on Euler characteristic. one denes a measure' d over denable sets via: ∫ A d = (A). where c 2 Z and f g is a decomposition of X into a disjoint union of open cells. Euler Characteristic Two tame sets in an o-minimal structure are denably homeo- Theorem 3. meaning that for h : X ! Z. all level sets h 1 (n)  X are tame. not a measure. (3. The similarity between Equation (3. 3. This additivity has been foreshadowed in the imagery of  as counting in Ÿ3.

then . This is not as easy as it sounds. and it is even less easy when X is not a continuum space. each of which observes some subset of W and counts the number of targets therein. # = N X h d ∫ (∑ ∫ If h : X ! N where h is a counting U of uniform Euler characteristic (U ) = N 6= 0 for all Proof. independent of knowing a decomposition of the counting function h. The sensors will be assumed to be distributed over a region so densely as to be approximated by a topological space X .9 ([?]). Target enumeration 45 That this sum is invariant under the decomposition into denable cells is a consequence of Lemma 3. This counting function h : X ! Z is.8 and Theorem 1. Consider a nite collection of targets. Assume that the sensors are additive but anonymizing: each sensor at x 2 X counts the number of targets in W detectable and returns a local count h(x ). determine the total number of targets in W . but rather a discretization thereof. In a particular system of sensors in X and targets in W . x ) 2 S i a sensor at x 2 X detects a target at w 2 W . under the usual tameness assumptions.12) For contractible supports (such as in the setting of beacons visible on star-convex . let the sensing relation be the relation S  W  X where (w. The horizontal and vertical bers (inverse images of the projections of S to X and W respectively) have simple interpretations.3. function for target supports 1 . represented as discrete points in a space W . Assume a eld of sensors. The horizontal bers  sensor supports  are those targets observable to a given sensor in X . It is therefore remarkable that a purely topological solution exists.7. constructible.7 Target enumeration A simple application of Euler integration to data aggregation demonstrates the utility of this calculus. ∫ ) X h d = X U d = ∑∫ X U d = ∑ (U ) = N # : (3. Theorem 3. A natural problem in this context is to aggregate the redundant anonymous target counts: given h and some minimal information about the sensing relation S. 3. The vertical bers  target supports  are those sets of sensors which detect a given target in W . but the identities of the sensed targets are unknown.

Let F :X!Y ∫ be a tame mapping between tame spaces. y )dx dy = f (x. This familiar result is the image of a deeper truth about integrations and projections: Given F : X ! Y . Real-analysis students learn to pay attention to ner assumptions on measurability (cf. This solves a problem in the aggregation of redundant data. If X is homeomorphic to U  Y and F is projection to the second factor. the target count is. Since the solution is in terms of an integral. the lemma holds via the exponent rule since  + dim  . Euler Characteristic domains). (X  Y ) = (X )(Y ). like volume. The product X  Y has a denable cell structure using products of cells from dim(   ) = dim X and Y.13) Proof. The assertion that d is an honest measure is supported by this fact and its corollary: the Euler integration theory theorem. then integrate the resulting function over the projected base Y . in practice).10.10. it is nontrivial to aggregate the redundancy. cf. the result follows from Lemma 3. Then for X h d = ∫ (∫ Y F 1( y) h(x ) d(x ) d(y ): ) (3. in the absence of target identication (an expensive signal processing task. 3. all h 2 CF (X ).46 Chapter 3. y )dy dx . Proof. one can integrate over the bers. Notice that the restriction N = 0 is nontrivial. If h 2 CF(R2 ) is a nite sum of characteristic ∫ functions over annuli. simply. of F rst. For X and Y denable.8 The Fubini Theorem Euler characteristic is like a volume in another aspect: it. Calculus students know ∫∫ admits a Fubini ∫∫ that f (x. Lemma 3. area is base-times-height. Given this. The o-minimal Hardt Theorem [?] says that Y has a denable partition into tame sets Y such that F 1 (Y ) is denably homeomorphic . it is a fundamental obstruction to disambiguating sets. the Euler integral of the function. The rest follows from additivity. or level sets. it is not merely inconvenient that 42 h d = 0. is multiplicative under cross products. local and distributed computations may be used in practice. For cells   X and   Y . since many nearby sensors with the same reading are detecting the same targets. tameness assumptions). Theorem 3. it is all the more remarkable that sets with  6= 0 can be enumerated easily.11.

Proof. This number is precisely the sensor count h(x ) (the number ∫ of times a sensor detects a vehicle coming into range). Specically.9. each of which moves along a smooth curve i : [0. Euler integral transforms to U  Y for U denable. Consider a collection of vehicles. T ]. or any ancillary data. Consider the projection map p : R2  [0. imaging. where h(x ) is the number of times x has entered a support. T ]. T ].9 Euler integral transforms Euler integration admits a variety of operations which mimic analytic structures of particular relevance to signal processing. g 2 CF(V ). varying tamely in t . N .3. in this case.14 (Convolution and Minkowski sum) On a real vector space V . Since p 1 (x ) is fx g  [0. 47 F : U  Y ! Y acts via projection. T ]. Example 3. Given f . the sensors do not identify vehicles. and inverse problems. by Theorem 3. Additivity plication of the Fubini Theorem to time-dependent targets. one denes (f  g )(x ) = ∫ V f (t )g (x t ) d(t ): (3. By the Fubini Theorem.9. Each such tube has  = 1. do they record times. The integral over R2  [0.13 ([?]). t ) for t 2 [0. which equals N . T ] of the sum of the characteristic functions over all N tubes is. 42 h d must equal the integral over the full R2  [0.14) . Example 3. assume that each vehicle possesses a footprint  a support Ui (t )  R2 which is a compact contractible neighborhood of i (t ) for each i . directions of approach. nor. the number of targets. At the moment when a sensor x 2 R2 detects when a vehicle comes within proximity range  when x crosses into Ui (t )  that sensor increments its internal counter. T ] 3. T ].12 (Enumerating vehicles) The following example demonstrates an ap- Proposition 3. the integral over p 1 (x ) records the number of (necessarily compact) connected intervals in the intersection of p 1 (x ) with the tubes in R2  [0. Over the time interval [0. ∫ to 42 h d. and that of the integral completes the proof. T ] ! R2 . T ] ! R2 in a plane lled with sensors that count the passage of a vehicle and increment an internal counter. Each target traces out a compact tube in R2 [0. The number of vehicles is equal given by the union of slices (Ui (t ). the sensor eld records a counting function h 2 CF(R2 ). As before. a convolution operation with respect to Euler characteristic is straightforward.

48

Chapter 3. Euler Characteristic

There is a close relationship between convolution and the Minkowski sum: for A and convex, A  B = A+B , where A + B is the set of all vectors expressible as a sum of a vector in A and a vector in B [?, ?]. This, in turn, is useful in applications ranging from computer graphics [?] to motion-planning for robots around obstacles [?]. For non-convex shapes, convolution of indicator functions can take on nonzero values besides 1.

B

One of the most general integral transforms is the Radon transform of Schapira [?, ?, ?]. Consider a locally closed denable relation S  W  X (that may or may not come from sensing), and let W and X denote the projection maps of W  X to their factors. The Radon transform with kernel S is the map RS : CF(W ) ! CF(X ) given by ∫ (RS h)(x ) = h  W d(w ): (3.15) X 1 (x )

Example 3.15 (Target enumeration) Consider the sensor relation S  W  X , and a nite set of targets T  W as dening an atomic function T 2 CF(W ). Observe that the counting function which the sensor eld on X returns is precisely the Radon transform RS T . In this language, Theorem 3.9 is equivalent to the following: assume that S  W  X has vertical bers W1 (w ) \ S with constant Euler∫characteristic∫N . Then, RS : CF(W ) ! CF(X ) scales integration by a factor of N : X  RS = N W .
A similar regularity in the Euler characteristics of bers allows a general inversion formula for the Radon transform [?]. One must choose an `inverse' relation S0  X  W.

X  W have bers Sw and S0w satisfying (1) (Sw \ S0w 0 ) =  6=  for all w 0 6= w 2 W . Then for all
(RS0  RS )h = (

Theorem 3.16 (Schapira inversion formula). 

W X (Sw \ S0w ) =  w 2W h 2 CF(W )
Assume that

S

and

S0 

for all ,

; and (2) 

)h + 

(∫

W

h W :

)

(3.16)

Recall that the sensor counting eld h : X ! Z is equal to RS T , where T  W is the set of targets. If the conditions of Theorem 3.16 are met, then the inverse Radon transform RS0 h = RS0 RS T is equal to a multiple of T plus a multiple of W . Thus, one can localize the targets by performing the inverse transform. It is remarkable that counting data alone can yield not only target counts but target positions as well. By changing T from a discrete set to a collection of (say, contractible) compact sets, one notes that Radon inversion has the potential to not merely localize but recover shape. This topological tomography is the motivation for Schapira's incisive paper [?].

Example 3.17 (Medical imaging) Schapira's paper outlines an application in imag-

ing most natural in the setting of MRI. Assume that W = R3 and one scans a compact subset T  W by slicing R3 along all at hyperplanes, recording simply the Euler

3.9. Euler integral transforms

49

characteristics of the slices of T . Since a compact subset of a plane has Euler characteristic the number of connected components minus the number of holes (which, in turn, equals the number of bounded connected components of the complement), it is feasible to compute an accurate Euler characteristic, even in the context of noisy readings. This yields a constructible function on the sensor space 3 X = AG 3 2 (the ane Grassmannian of 2-planes in R ) equal to the Radon transform of T . Using the same sensor relation to dene the inverse transform is eective. Since Sw  = P2 and Sw \ Sw 0  = P1 , one has  = (P2 ) = 1, 1  = (P ) = 0, and the inverse Radon transform, by (3.16), yields T exactly: one can recover the shape of T based solely on connectivity data of black and white regions of slices.

Example 3.18 (Sensor beams) Let the targets T be a nite disjoint collection of points in W = Dn , the open unit disc in Rn . Assume that the boundary @W is lined with sensors, each of which sweeps a ray over W and counts, as a function of bearing, the number of targets intersected by the beam. The bearing of a ray at a point p 2 @W lies in the open hemisphere of the unit tangent bundle to W at p . This open hemisphere projects to the open unit disc in Tp @W . Thus, the sensor space X is homeomorphic to T  (@W )  = T S n 1 , the tangent bundle of @W . The sensing relation will be seen to be self-dual: S = S0 . Any point in W is seen by any sensor in @W along a unique bearing angle. Thus, the sensor relation S has vertical bers (target supports) which are sections of T Sn 1 and hence spheres of Euler measure  = (Sn ) = 1 + ( 1)n . Any two distinct vertical bers project to X and intersect along the subset of rays from @W that pass through both points in W : this is a discrete set of cardinality (hence )  = 2. For n even, one has  = 2,  = 0, and Equation (3.16) implies that the inverse Radon transform gives 2T +2(#T )W . Thus, on even-dimensional spaces, targets may be localized with boundary beam sensors. One may likewise assume that the target set T is not a collection of points but rather a disjoint collection of compact sets in W ; target shapes will then be recovered, assuming the ability of beams to measure Euler characteristic of slices exactly. Example 3.19 (Fourier transforms and curvature) One relationship between Euler
and Lebesgue measures is encoded in the Gauss-Bonnet theorem. This, too, can be lifted from manifolds to denable sets and then to CF(Rn ). The mechanism is the

50

Chapter 3. Euler Characteristic 

2 Sn

microlocal Fourier transform. Given h 2 CF(Rn ), a point x 2 Rn , and a unit vector
1

, the microlocal Fourier transform is dened via: ∫  h d(y ); (M F h)x () = lim+ !0 B (x ) (y x )0

(3.17)

where B denotes the open ball of radius . Like the classical Fourier transform, M F takes a frequency vector  and integrates over isospectral sets dened by dot product. For Y a compact tame set, M FY () is the constructible function on Y that, when averaged over , yields a curvature measure dY on Y implicated in the Gauss-Bonnet Theorem (as shown by Bröcker and Kuppe [?]): for any U  Rn open,

U

dY = Vol Sn

1


1


1

5n

U

M FY

d d:

The Euler-calculus interpretation of the Gauss-Bonnet Theorem says, ∫ ∫ ∫ 1 M F  d d = Y = (Y ); Y Vol Sn 1 5n 1 Y Y (3.18) and this extends via linearity of M F to all of CF(Y ).

3.10 Intrinsic volumes
The humble combinatorial denition of  has evolved in this chapter to the more analytic construct of a measure. This has a grand history in the subject of integral geometry. There is a classical family of measures on Euclidean Rn that mediate Euler and Lebesgue measure and entwine topological with geometric data. The k th intrinsic volume k is characterized uniquely by the following: for all A and B tame in Rn , 1. 2. 3. 4.

Additivity: k (A [ B ) = k (A) + k (B ) k (A \ B ); Euclidean invariance: k is invariant under rigid motions of Rn ; Homogeneity: k (  A) = k (A) for all   0; and Normalization: k of a closed unit ball in Rn equals 1.

These measures generalize Euclidean n-dimensional volume (n = dvoln ) and Euler characteristic (0 = d). There are several equivalent denitions, all revolving about the notion of an average Euler characteristic. One way to dene the intrinsic volume k (A) is in terms of the Euler characteristic of all slices of A along ane codimension-k planes: ∫ ∫ ∫ k (A) = n k (A \ P ) d(P ) = n k A d d(P ); (3.19) Pln Pln P 
Intrinsic volumes are also known as Hadwiger measures, quermassintegrale, Lipschitz-Killing curvatures, Minkowski functionals, and, likely, a few more names unknown to the author.

3. A valuation is compact-continuous if it is continuous with respect to the Hausdor metric on compact convex sets. Thus. metals.Mullins formula states that the area A(t ) of a cell in an ideal dynamic 2-d cell structure changes as: ( ) dA j C0 j = 2M 1 .20 (Hadwiger Theorem). A classical theorem of Hadwiger [?] characterizes such valuations: Theorem 3. the space of ane (n k )-planes in Rn . Euclidean-invariant valuations on Rn has fk gn The vector space of compact-continuous 0 as a basis.20) is in fact the Euler characteristic  This measure is derived from the Haar measure on the Grassmannian /n n k and Lebesgue measure on the orthogonal 4k . while others shrink unto disappearance (neglecting pops). is a surface tension constant. The key to the extension was to note that the  1 in Equation (3. (3.20) dt 6 where M is a mobility constant. cells with fewer than six corners shrink: those with more grow. See [?] for a complete discussion. The same processes are ubiquitous in the microstructure coarsening of froth. The proper extension of this formula to three-dimensional cell structures was recently discovered by MacPherson and Srolovitz [?]. and jC0 j equals the number of corners (vertices) of the cell. The details of this construction are left to the literature: the point here stressed is that all the intrinsic volumes are certain Lebesgue-averaged Euler integrals. These amalgamations of Euler and Lebesgue measure are the key to several interesting applications. The von Neumann . and some ceramics [?]. Consider the vector space of Euclidean-invariant valuations  additive functionals on tame sets. It has long been known how idealized cells evolve and coarsen in the 2-dimensional setting. Example 3. .10.21 (Microstructure coarsening) Anyone who has observed foam in a pilsner glass knows that the cell walls of foam evolve over time so that some cells grow. Intrinsic volumes 51 where  is an appropriate measure on Pln n k . These intrinsic volumes are more than isolated examples.

21) dt 6 where 1 is the intrinsic 1-volume of the cell and jC1 j is the total length of all the edges (1-dimensional boundary curves) of the cell. A real-valued stochastic process is. A random eld is a stochastic process whose arguments vary continuously over the parameter space X . Taylor [?]. 3. for each k 2 N and each k -tuple fxi gk 0  X . measures the variance of directional derivatives of h with . so are the simplest random elds. One of the principal results in this area is that. both examples having as parameter space a geometric domain of dimension two (or three. compact) cell. attention will be restricted to Gaussian distributions with the following features: 1.21 is in fact an elegant relationship between dierent (intrinsic) volumes. astronomy. Efh(x )g = 0. which. and others have led in the exploration of Gaussian random elds from a geometric perspective. in brief. Euler Characteristic of the (contractible. = 2M 1 (3. 3. jx y j. To lift from 2-d to 3-d requires lifting the relevant interior volumes from 0 (= ) to 1 . with the addition of time). There is implicit in this work no small amount of topology. roughly. A Gaussian random eld with parameter space X  Rn is a random eld h over X k such that. Worsley [?]. one might wish to know the expected number of peaks in a eld. For example. Centered: all h(x ) have zero mean. although the These last two assumptions imply a well-dened constant . a random eld may be used to model the height of water on a noisy sea or the magnitude of noise in a sea of cell-phone towers. Knowing as much as possible about the tomography of the eld h is important in medical imaging. The MacPherson-Srolovitz formula for the volume V (t ) of a cell a in 3-d structure is: ( dV jC1 j ) .52 Chapter 3. For simplicity. Common variance: all h(x ) have xed variance 2 > 0. For example. the left hand side implicating 3 and the right implicating 1 . the collection fh (xi )g0 of random variables has a multivariate Gaussian distribution. with Euler characteristics featuring prominently.11 Gaussian random elds Most readers will be comfortable with the utility (if not the details) of stochastic processes. be a stationary random Gaussian eld. When contemplated in the light of intrinsic volumes. As the simplest random variables are Gaussian (having well-dened nite mean and variance with the usual normal distribution). or some other qualitative properties associated to the excursion sets fh  s g. 2. Stationary-isotropic: the covariance Efh(x )h(y )g is a function of distance respect to x . and a host of other applications [?]. it is clear that Equation 3. Adler [?]. a collection fh(x ) : x 2 X g of random variables h(x ) parameterized by a space X . the second spectral moment. x 2 T  Rn . Let h(x ).

11. . and sucient  regularity. and Gaussian. however. ?] for (many) details. the simplication that  introduces and its commensurability with integral techniques yields to computational eort. In addition. a stationary. ?]). isotropic. It appears very difcult to understand the expected shape or even topological type of these excursion sets. but computed and commensurate with experimental data [?]. the expected number of up-crossings at h = s  locations where h increases past s  is given by E#fh % s g = e The observation that the number of upcrossings #fh % s g is in fact the Euler characteristic fh  s g of the upper excursion set presages the deeper work of [?] for elds over higher-dimensional domains. stationary. The following result is a highly illustrative example: s2 2 2  T: 2 p (3.3. second spectral moment .23) Hek is the degree intrinsic volume of T. Assume that E(fh  s g) = e where s2 2 2 dim X ∑ k =0 (2 ) k +1 2 ( p )k   Hek 1 (s )  k (X ). almostsurely continuous Gaussian process with nite variance  2 and second spectral moment . The expected Euler characteristic of the upper excursion set fh  s g n over a compact denable subset X  R is Theorem 3. See [?. (3. T ]. x 2 [0. Gaussian random elds 53 ne details of the expected eld excursion sets cannot be computed. isotropic 2 Gaussian random eld with variance  .22) h(x ). The following classical result of Rice from the 1940s [?] helped initiate the subject: for h(x ). k Hermite polynomial in one variable and k (X ) is the k th Additional work [?] allows one to relax the assumptions of the eld being stationary. the expected Euler characteristics are not merely computable. zero mean.  This consists of nondegeneracy assumptions on the joint distributions of the rst and second derivatives of f and regularity assumptions on the covariance functions of the second derivatives. expectations for intrinsic volumes k of excursion sets are likewise derivable. x 2 Rn is a centered.22 ([?.

- It is to be suspected that a full theory of 3. 6. Higher perspectives are more illuminating. The scissors equivalence implicit in Theorem is the hint of deeper structures. Research in geometric random elds has the reputation for being very dicult to understand. Intrinsic volumes are well-dened on all of ? sets  thanks to the close relationship with CF (Rn ) ∫ d.1. Euler integration is here. community. as in [ ]. Exercises 3. Chapter 8 and [ ] give an interpretation in terms of sheaf theory. One way around this depressing result is to extend valued measure of sequences of spaces. where one encodes the Euler characteristics of ? The monograph [ ] details a theory of Gal ?  to a polynomialt. but requires a more delicate function space topology. [ ] on Euler characteristics associated to conguration spaces. The issue of numerical Euler integration  how to approximate an integral with respect d based on sampled data  is extremely interesting.3. ?] allows.54 Chapter 3. It is possible and protable to perform Radon transforms (and inversion thereof ) with weighted kernels: see [ ]. the resulting series is the Taylor series of a rational function. as it would likely subsume many of denable sets in the statement of Theorem the nondegeneracy assumptions. Euler Characteristic Notes 1. functions from algebraic geometry [ . an approach in terms of conormal cycles and related geometric measure theory. as it removes much of the mystery as to the natural appearance of  in this work. Euler of great current interest in algebraic geometry. Use Gauss-Bonnet to show that any sphere in R3 built from regular hexagons and This result has pentagons must have exactly 12 pentagons: no more and no less. integration with respect to UCn (X ) as coecients of a series in a formal variable For Similar algebraic manipulations for Euler characteristics of sequences of spaces appears in work on d[t ] over ltered spaces awaits development. and relevant It appears to have received little treatment from the Mathematics to applications. Caveate. and the sums of indices of the enclosed (assumed isolated) xed . presented as a combinatorial theory. 2. The literature on ? normal cycles [?. implications constraining the possible arrangements of carbon into buckyballs. Use the Poincaré-Hopf theorem to show that any periodic orbit in the plane encloses at least one xed point. This is perhaps due more to the typical mutual ignorance of researchers in probability and in topology.  not merely compact denable Continuity of these measures is possible. It may help matters to employ (1) the language of Euler integration. relying on currents (see Chapter 6). 7. Since the Euler measure is the universal motivic measure over topological spaces. thanks to results of Kashiwara [?]. technical. to 4. 5. 3. X tame. any other scissors-invariant group-valued measure on denable topological spaces must factor through . As a rst step. nor have I re-tooled Adler's proof (and all of probability theory) to the o-minimal setting.2. I have taken the liberty of invoking ??: this is not what Adler proves. A periodic orbit of a vector eld is a owline that is homeomorphic to S1 . A few other integral transforms will appear in Ÿ7. ? ? ?]. and (2) the o-minimal framework. characteristic and integration over CF is an elementary version of motivic integration.

55 X under a denable map f : X ! Rn is 3. Compute DA for A a compact convex Rn .11.14. In the Radon inversion of Example 3. Computing Euler integrals via Equation (3.7.10. a manifestation of the Fundamental Theorem of Integral Calculus.8. Why does the Fubini theorem for Euler integration not have the customary caveats about checking that the level sets of 3. There are some nite collection of distinct targets in [i Ui . is often more stable: ∫ X Prove this formula.18.9. Compute the (Euler) convolution of the characteristic functions of two round circles of radii r1 and r2 in the plane. Consider a collection n + 1 video cameras surveying a region. Prove directly that the image of a tame set again tame. show that h d is the total number N ∫ V U F are measurable sets? k = fUi gn of convex sets in R . but can determine which targets are seen in common by multiple cameras. Compute the microlocal Fourier transform of P for P a polygon in the plane. explain from physical reasoning why in the 3.5.3. thought of as regions seen by 0 of targets visible. Show that for A open and denable in Rn . 3.13. that the resulting -average yields the angle defect concentrated at vertices.24) 3. k (A) = ( 1)n k k (A) . since the level sets are usually not compact. but the cameras cannot identify them. B (x ) is a small open ball about x . What familiar quantity do you subset of obtain? 3. 3. 3. Given the function h : N(U) ! N which assigns to each (open) simplex of the nerve N(U) the number of targets in the ∫ corresponding intersection of elements of U.12. n = 1 there is no hope of inverting the Radon transform.6.Exercises points equals one. The following formula. How do these change as a function of radii? 3. Show that convolution in Euler integration is multiplicative: f g d = ∫ V f d ∫ V g d. 2 Compute 1 of a submanifold of R with boundary. 3. n n The following integral transform D : CF(R ) ! CF(R ) is from [?]: case Show Dh ( x ) = where lim !0 ∫ + B (x ) h d. 3.10) can be tricky. Show that DD = Id: this is why D is called the dual transform.4. 3. h d = 1 ∑ s =0 f h > s g f h < s g : (3.

Chapter 4 Homology .

and canceling according to parity of dimension. this objection will have been forgotten. in this case by the dimension of the simplices. The building blocks of a rudimentary homology for X are as follows. certainly. Counting in this eld is analogous to ipping a light switch. n+1 . and the next. This leads to the notion of homology. counting. Dene k -chains Ck as the vector space over the eld F2 = f0. 4. each with coecient 1).1 ( Consider ) . in the obvious basis. each cell of which is outtted with a metaphorical light switch. The initial discussion proceeds using the language of linear algebra and passes to more general algebraic constructs. Homology ecomposing a space into cells.1) The chain complex is graded. probably. 1g is the eld with two elements. More rened invariants arise upon enriching this data with linear-algebraic sensibilities. The reader who is weak in either linear or abstract algebra is encouraged to consult the Appendix (at least) for a review before proceeding.1 Chains The crucial construction of this chapter is that which converts a decomposition of a space in terms of simple pieces into a collection of vector spaces (or modules) and linear transformations (or homomorphisms): an algebraic version of a simplicial complex. consider a nite cell complex X .. For simplicity. It is benecial to denote the chain complex as a single object C = (C . To build intuition. 58 Chapter 4. 2. The resulting chain complex has Ck of Example 4. Recall that F2 = f0. The top-dimensional boundary map @ is. 1g with basis the k -cells of X . The collection of chains and boundary maps is assembled into a chain complex: @0 /0:  / Ck @ @k / Ck 1 k 1 /  @2 / C1 @1 / C0 (4. when the geometric meaning of a boundary is clear? By the end of this chapter. 1. @ ) and to write @ for the boundary operator acting on any chain of unspecied grading. a dimension k n +1 a single n-simplex n . Consider the boundary maps  the linear transformations @k : Ck ! Ck 1 which send a basis k -cell to its boundary (as an abstract sum of basis (k 1)cells. is ecacious for dening one topological invariant. mod-2 arithmetic is used initially. It seems at rst foolish to algebraicize the problem in this manner  why bother with vector spaces which simply record whether a cell is present (1) or absent (0)? Why express the boundary of a cell in terms of linear transformations.

1]n . Homology examples 59 1-by-n matrix with all entries 1. the n-dimensional cube with cell structure inherited from the (n) interval I = [0. It suces to show that @ 2 = 0 on each basis simplex  . for I n = [0. The graded boundary operator @ : C ! C is thus a formal ⊕ sum of face operators @ = 2V D . In contrast. the chain complex has Ck of dimension 2n k k .2) @ 2  = D. do nothing.2. Computing the composition in terms of face operators. 1] with two endpoints. Lemma 4. Note that in all cases. the boundary map @0 is the zero map. else. @k = 0 for all k . The following lemma is deep. The boundary of a boundary is null: @k 1  from the simplex's list.2. one obtains: ∑ (4. if present.4.

D : 6=.

Each (k 2)-face of the k -simplex  is represented exactly twice in the image of D.

D over all 6= .

For simplicity. As @0 = 0. Homology is an equivalence relation on cycles of C. Elements of H (X ) are homology classes.3) To repeat: Zk = ker @k is the subspace of cycles in Ck . and Bk = im @k +1 is the subspace of boundaries. The homology of X is the sequence of quotient vector spaces Hk (X ) over F2 . i. is a sequence of F2 -vector spaces built from the following subspaces of chains.e. for k 2 N. Two cycles in Zk = ker @k are said to be homologous if they dier by an element of Bk = im @k +1 . the sum over this pair is zero. H (C).2 Homology examples Example 4.. given by: Hk (X ) = Zk =Bk = ker @k / im @k +1 = cycles/ boundaries: (4. Thanks to F2 coecients. 4. The homology of C. im @k +1 is a subspace of ker @k . consider the case of an abstract simplicial complex on a vertex set V = fv g. The face operator D acts on a simplex by removing the vertex v Corollary 4. an element of ker @ . .3. A cycle of C is a chain with empty boundary. Proof. all 0-chains are 0-cycles: Z0 = C0 . Homology inherits the grading of C and will be therefore denoted H (X ) when a dimension is not specied.4 (Graphs) Consider G a topological graph  a compact 1-d cell complex. For all k . The boundary subspace B0  Z0 .

and bases. both of dimension one. In particular. As a graph is a 1-dimensional complex. and dim Hk (D2 ) = 0 for k > 1. 1-cycles are homologous if and only if they are identical. on an orientable surface. 5.5 (Surfaces) For Sg a simplicial complex homeomorphic to a connected 2g dim Hk (Sg ) =  1   0 It is worth stressing that H1 (X ) does not measure whether a cycle in X is contractible  whether it can be shrunk to a point in X continuously. since all cells are of dimension less than three. The dimension of H0 (G) is therefore the number of such path components. dim Hk (M ) = dim Hn k (M ). any loop (1-cycle) which divides the surface into two connected components is nullhomologous. Correspondingly. For Dn n 2 ) = n . dim H1 (Dn n 4. there are no higher dimensional chains with which to form homology classes: for graphs. The relation to genus is as follows: Example 4. spans. if and only if they lie in the same connected path-component of G. For a compact manifold M of dimension n. 2 a disc in the plane with n disjoint discs removed. coecients. dim H0 (D2 ) = 1. Hk (X ) = 0 for all k > 0. it follows that any two basis elements are homologous if and only if they are the endpoints of an end-to-end sequence of edges in G  that is. For X contractible. 2. dim H0 (X ) is equal to the number of path components of X . 3. H1 (G) collates the number of linearly independent cyclic chains of edges. complete with the notions of sums.! ! Let the reader beware: this result requires the use of . with an even number in each connected component of the graph. For example. The sphere Sn has Hk = 0 for all k except H0 and Hn . and Hk (G) = 0 for k > 1. Note that homology gives the collection of 1-cycles in G the structure of a vector space. Any element of Z1 (G) consists of a nite union of cyclic end-to-end sequences of edges. Homology consists of nite unions of vertices. If one chooses the vertex set of G as a basis for Z0 . the homology is concentrated in dimensions less than three. Hn (M ) has dimension one if and only if it is connected. though it may not be contractible.4) 1. A few facts (presented without proof) may help the beginning reader build up an intuition for what it is that homology measures:  1    : : : : k =0 k =1 : k =2 k >2 (4. compact orientable surface of genus g . 2 .60 Chapter 4.

3. vk ]. with coecients ck := dim Hk (X ). The simple version of the Künneth Theorem states that Pt (X  Y ) = Pt (X )Pt (Y ). in contrast to F2 's involutive switch. Kirchho's current rule states that the current owing through the edges of the circuit satises a conservation principle: at each node. Each k -simplex of X is specied as k = fvo . There is an important development when going beyond F2 coecients: 1 6= 1. Such coecients permit expressiveness: R is eective in describing simplices' intensities. For example. Example 4. this is naturally encoded as directedness. from which the computation for Tn follows. the sum of the incoming currents equals the sum of the outgoing currents.) located at certain vertices. v0 . F ). one may construct chains Ck for a cell complex X as F-vector spaces. Hk (T n ) = ( ) ⨿ Example 4. .3 Coecients The homology introduced in Equation (4. v1 . other elds may be used. Recall that an abstract simplicial complex consists of simplices on a formal vertex set V . Note the unambiguous need for R coecients. [vo . These additional adjectives are necessary or H 2 due to the plethora of homology theories in existence. The eect of this algebraic unfolding is the need for an orientation associated to simplices.7 (Kirchho's current rule) Consider an electric circuit as a 1-dimensional cell complex. For any eld F. An oriented simplex is a simplex with an ordering of vertices: [vo . 4. since current is measured as a real quantity. etc. Coecients 6. For a disjoint union of spaces 61 X = A B. v1 . instead of the Ck being vector spaces over a eld F2 . In the algebra of chains. Consider the Poincaré polynomial of X . @ ) of vector spaces and linear transformations with the property @ 2 = 0 is a chain complex suitable for building a homology theory. Pt (X ) = k ck t k . via∑ the the coecients of (1 + t )n .3) should be denoted the cellular homology of X with F2 coecients. The Poincaré polynomial of S1 is 1 + t . : : : . Hk (X ) = Hk (A)  Hk (B). cell (X . This yields a well-dened multiplicative action of f 1. Any graded sequence C = (C . thus. v2 . v1 ] = [v2 . v2 ] = [vo .6 (Cross products) A torus Tn = (S1 )n has homology satisfying dim n k : The reader will rightly suspect a relationship with polynomial algebra. : : : . In the language of this chapter: current is a 1-cycle. one uses a twist of the hand for orientations on curves and surfaces: a more formal approach is needed for the present setting. vk g for some choice of vi 2 V . An orientation on  is an equivalence class of orderings up to even transpositions.4. v1 . In calculus class. v1 ]. with circuit elements (resistors. capacitors. +1g on oriented chains: multiplication by 1 is the chain analogue of reversing orientation. For 1-chains.

Z) over a cell complex X with integer coecients as recording a nite collection of simplices with orientation and multiplicity. still.62 Chapter 4. This indicates that the meridian and longitudinal curves on K 2 . where Z2 = Z=2Z is the quotient of Z by the subgroup of evens. Ultimately.8 (Coecients) Using dierent coecients can lead to genuinely dierexample. Note also the presence of torsion in H1 (K 2 . Field denoted H (X . ent homology groups. Consider SO3 . there is a torsional Example 4. K 2 is a non-orientable surface.2): @ = k ∑ i =0 ( 1)i Di : (4. the presence of a torsional element in H (X . two chains of opposite orientation cancel.). etc. ple. belts. One visualizes chains C (X . The boundary operator incorporates orientations as follows (using the face operators from the proof of Lemma 4.9 (Rotations and projections) Rotations in R3 oer a fascinating. F2 ) =  F F   2 Example 4. H (K 2 . F). in surface examples. The resulting homology is cell (X . As one can demonstrate using a variety of physical devices (plates. F). are qualitatively dierent: sliding a meridional curve along a longitude reverses orientation. or. the Klein bottle K 2 satises: 0 : : : :  k =0 : k =0 k = 1 : . Z) is indicative of some type of twisting in X which. one works with a chain complex over an R-module (for R a ring) with the boundary maps module homomorphisms.5) The reader should check that the use of ( 1)i 's suces to keep Lemma 4. manifests itself in non-orientability. Homology as with contour integration. but it does in F2 : thus. The Klein bottle has no 2-cycle (a boundaryless nonzero chain) in Z coecients.2 in eect. sim- . Z coecients yield dierent ranks of homology groups at grading one and two. fermions. not the full story. In general. For  F2    2  F2 Hk (K 2 . unlike those of the torus T2 . H coecients and linear algebra are. and useful example of torsional phenomena where coecients matter. Z ) =  Z Z  Z 2 : k =1 : k k =2  0 : k >1 k >2 In this example. the group of real 3-by-3 matrices with determinant 1. more formally. Z)  = Z  Z2 . The generalization to Z coecients  in which each Ck is a Z-module  is particularly relevant. These are precisely the orientation-preserving rotations of Euclidean 3-space.

The reader will have observed that cellular homology is.. where the generators of Ck are continuous maps k : k ! X from the standard simplex k to X . In this theory. auxiliary structures imposed). The . These simplices do not t together to form a triangulation of X : there are far too many. Notice that there is a decoupling between the grading and the dimension of the image of the singular simplex. metric nite cell complexes). with no sensible notion of dimension other than in its preimage. dim Hk (P .g.4. The reason that Vietoris' theory is mostly forgotten lies in the ecacy of singular homology.g. Z)  = Z2 .. F2 ) = 1 for all 0  k  n. The boundary of a singular simplex is the restriction of the map to @ k . The singular chain complex of a topological space X is the complex Csing = (Ck . the image of a singular simplex may be very convoluted. However. the homology of the Vietoris chain complex stabilizes as  ! 0+ . The story for integer coecients is quite dierent:   Hk (Pn .. Z) =  Z2 : 0 : Z : k = 0 or k = n.4. this is an example of the n homology of projective spaces. Though the generators of C may be simplices or cells of a space X .4 Singular homology A denition with sucient strength has emerged: a chain complex C = (Ck . which requires neither a metric space nor the explicit limiting process. @k ).g. The next several sections delineate a few major homology theories (with others to follow in later chapters). The rst such example is singular homology  the most common and generally useful theory. but rather by limits to uncomputable abundance. @k ) is any sequence of R-modules Ck with homomorphisms @k satisfying @k  @k +1 = 0. like Euler characteristic. Indeed. One of the rst to take such an approach was Vietoris. one xes an  > 0 and considers the chain complex is generated by k -tuples of distinct points in X of pairwise distance  . Since SO3  = P3 . Singular homology 63 core writhing within: H1 (SO3 . who introduced his homology theory for metric spaces [?]. there are many dierent objects worthy of being counted and compared. independent of how the space is decomposed into cells. To be noted are the commonalities (e. In general. notions of dimension associated to the objects counted) and the distinctions (e. and this agrees with the cellular homology of a cell complex. The best method of proof comes not from parsimonious combinatorial renements of cell structures. for reasonable spaces (e. odd 0 < k odd < n : else 4.

(4. For two spaces that are homeomorphic. Reduced homology simplies certain results: it is convenient to have ~ = 0.6) 0 :k < 1 with all boundary maps as above. Let X be a topological space and U = fU g a collection of open sets whose union is X .64 Chapter 4. H =H =H The reader may think of the codomain of  as C 1 .6 ƒech homology Any homology theory counts objects with grading and cancelation. The 0-chains C 0 have as Consider the following chain complex C  basis the elements of U. there is an equivalence between their singular chain complexes that guarantees equivalent homologies. the reduced ~ (C) satises Hk (C)  ~k (C) for k > 0 and H0 (C)  ~0 (C)  R. While the objects and gradings are ultimately indicative of cells and dimensions. dene the reduced chain complex C  / Cn @n / Cn 1 @n 1 /  @2 / C1 @1 / C0  / where  : C0 ! R sends each basis 0-chain to the sum of the coecients. Proposition 4. the correspondence may be disguised. having a single basis element at grading 1. C  R /0. @ k ).  : C0 ! . unordered) elements of U.  (U) = (C k . @ ) is enormous  certainly of uncountably innite resulting chain complex (C dimension except in the most trivial cases. as will be shown in the next chapter.7) @ ( 1)i  U j  : i =0 j 6=i " For . The yet more useful and general homotopy invariance of singular homology is the true reward for the unwieldy bulk of the singular complex.5 Reduced homology It is sometimes convenient to modify a homology theory in grading zero. homology For a nonempty chain complex C with coecients in G . Given a ~ chain complex C over an R-module. However. Consider the case of the ƒech homology of a cover. in this lies its exibility. and the k -chains Ck have as basis non-empty intersections of k + 1 (distinct. 4. as such an acyclic space plays the role a nonempty contractible space have H of a zero in homological algebra. . The resulting chain complex is outtted with the obvious boundary operator:   k ∑ ∩ k (U 0 \    \ U k ) = (4. where k 2 Z and: advantageous to write C   Ck : k  0 ~k = R :k= 1 ." It is ~k for the chain groups. Homology sing .. 4.-vector spaces.10.

A homological version of the Nerve Lemma (Theorem 2. with the induced boundary maps @ on quotients. A). 4. the signs are superuous.7 Relative homology Homology relative to a subcomplex is an important variant. is well-dened. Theorem 4.13 (Excision). rather. as appropriate) and C(X ) a chain complex (singular. independent of the cover.11. Hk (X. the complex C geometric nerve.7. Thus. one checks that @  (U)). by taking a limit of ner and ner covers. The more subtle construct is the chain complex of the quotient: let C(X. they are equivalence classes of chains in X which are identical o of A. This is particularly useful for characterizing holes in non-tame spaces that the more well-behaved singular theory cannot abide. The resulting relative homology H (X. Example 4. The reader will be reminded  U := H (C homology of the cover. a multiple of p and a zero-cycle. There are two natural chain complexes implicating A. Let A  X be a subset (or subcomplex. fp g)  = Hk (X ) for k > 0 and H0 (X. since the subspace corresponding to the homology class [fpg] has been quotiented out. A U )  = H (X. fpg) has rank one less than H0 (X ). precisely. The only nontrivial chain in X whose boundary is a nonzero multiple of p is. etc. consists of subgroups Cn (A) < Cn (X ) with the obvious restriction of @ . The rst. This chain complex yields the homology of A. interior of A. Then Let U  A  X with the closure of U H (X U. [?]. C(A). e.. A) = Cn (X )=Cn (A) of chains on X modulo chains on A. The details of this limiting procedure require a few tools from Chapter 8 and are not here detailed: see. U are acyclic ( If all nonempty intersections of elements The ƒech homology of a space X is well-dened. A) denote the quotients Cn (X. by denition.4.3) follows: Theorem 4. Relative homology 65  @  = 0 and the ƒech In F2 coecients. These relative chains do not vanish on A. then the ƒech homology of U H  (U)  agrees with the singular homology of X : H = H (X ). cellular. A) collates homology classes of relative cycles (chains in X whose boundaries lie in A). of ~  0).) on X . H  (U) is the algebraic shadow of the of the nerve complex of Chapter 2.12 (Reduced homology) The reduced homology of a space X is isomorphic to the homology of X relative to a basepoint p 2 X . contained in the .g. One of the foundational theorems about homology concerns relative homology.

In the singular setting. say. in keeping with one's experience from calculus class. . Example 4. A local orientation at p 2 M is a choice of generator for the n-dimensional local homology at p (in Z coecients). one sees H (Dn . This leads to a painless denition of orientation. From Corollary 4. Z). A global orientation on a compact n-manifold M is a choice of generator for Hn (M . is a consistent choice of local orientations. Homology Excision implies that for the pair (X. A). a concept that usually requires recourse to a dextrous sleight of hand. a closed subset of Q). truly. H For complex of a cell complex A  X a closed subX . Apply Excision. H The local homology of p 2 X is the (singular) homology H (X.17 (Local orientation) It follows from Example 4. X p ). The correctness of this intuition follows from excision: local homology is. with U = X V and A = X p. what happens inside of A is irrelevant. This is a little hard to visualize. A)  = The need for a subcomplex is superuous.15 (Relative homology) The homology of a closed disc Dn relative to its boundary @Dn has as a nontrivial generator Dn itself. X p)  = H (V. Corollary 4. local. V p): that is.14 is inapplicable. Proof. it suces to have A closed and possessing an open neighborhood which deforms continuously onto A (unlike. Example 4.16.8 Local homology Corollary 4. @Dn )  = ~ (Sn ). as a relative n-cycle. but a moment's thought yields the intuition that it measures something of the features of X in an arbitrarily small neighborhood of p . Let open neighborhood of p 2 X be a (closed) point and V an p.66 Chapter 4. local homology can be computed from an arbitrarily small neighborhood of p. ~ (X=A). H (X. 4.14 and the fact that Dn =@Dn is homeomorphic to Sn . Then H (X.14. A global orientation. given that Corollary 4.15 that the local ho~ (Sn ) and thus nontrivial of rank one in mology of an n-manifold is isomorphic to H dimension n.

H (R).19 (Ad hoc communications networks) An ad hoc communications network among a nite set X of devices is generated when devices broadcast their identities (assumed unique) and any other devices within range hear the signal and establish a link. This result was originally used to prove an equivalence between ƒech and Vietoris homology theories. xi )g  R. is dened as follows: 4. For R  X  Y a relation. and R records which points in X are within  > 0 of points in Y . Following the constructions of Theorem 4. Y is a dense net of points in X . there are two possible directed graphs: one encoding hears and one encoding heard by. dene a chain complex C(XR ) as follows: generators of Ck (XR ) are unordered (k + 1)-tuples of points in X related to some xed y 2 Y . the sensing modality used for target enumeration in Chapter 3 is a relation between sensors and targets collating which sensors detect which targets. y ) 2 R. For example. this case permits topologizing the network as a graph whose edges are symmetric communication links. The nerve of the cover of X by metric balls about Y has homology H (Y R ).18 (Dowker [?]).18. By convention. The communications network can be encoded in a relation R  X  X . the complex on X determined by horizontal bers of R  any collection of devices heard by device xi forms a simplex of XR . and the Vietoris homology of X is captured by H (XR ). Other applications are more salient to this text. One may wish or assume or impose that R is symmetric (devices are heard by those they hear). in the case where X is a metric space. In a nonsymmetric case. These two simplicial . xj ) 2 R if and only if messages sent from device xi are heard by device xj .9 Homology of a relation Theorem 4. the diagonal  = f(xi . A point x 2 X is related to a point y 2 Y if and only if (x. Homology of a relation 67 A relation between two sets X and Y is a subset R  X  Y . Dually. Higher-dimensional simplicial models are also available. The homology of the relation.4. Dual to this is XR .9. let X R be the simplicial complex on X determined by vertical bers of R  any collection of devices which hear device xi 2 X forms a simplex of X R . It is possible and protable to build homologies based on relations. H (R) := H (XR )  = H (Y R ). where (xi . one constructs generators of C(Y R ) from vertical bers of R  unordered tuples of points of Y related to a xed x 2 X . Example 4. The boundary operators for C(XR ) and C(Y R ) forget points in the tuples with sign. mimicking the boundary operator of a simplicial complex (the property of being related to a common element is closed under subsets).

Homology X have. f @ = @f . dene H(f )[ ] = [f ] = [f  ].18. an open tank of uid whose surface properties are experimentally measured and imaged. there is a graded induced homomorphism f : C (X ) ! C (Y ) at the level of (cellular or singular) chains given by f () = f   (where  is a basis k -chain). Similar situations arise in MRI data. but can only be approximated by imprecise pixellated images of fh  0g. If the induced homomorphism H (f ) is known. a hole. inferences can be made. Given f : X ! Y . by Theorem 4. One suspects that issues of consensus and convergence under evolution are more delicate for systems with H (R) nontrivial. so that f = f2  f1 with f1 : Y1 ! X and f2 : X ! Y2 . 4. the induced chain map respects the boundary operation. where the structure of a tissue of interest can be imaged as an approximation. It is sometimes the case that what is desired is knowledge of the homology of an important but unobserved space X . 2.10 Functoriality Homology gives a natural measure of the qualitative features not only of spaces but also of maps between spaces. Functoriality means that induced homomorphisms on homology are an algebraic reection of the properties of continuous maps between spaces. One implication in the sciences is inference. 3. or lack of connectivity) true? . This map is as it must be: for a cycle. is functoriality.20 (Experimental imaging data) The problem of measuring topological features of experimental data by means of imaging is particularly sensitive to threshold eects. the topology of the set A = fh  0g must be discerned. 1. This property. H (g  f ) = H (g )  H (f ). The identity map Id : X ! X induces the isomorphism Id : H X ! H X . Example 4. Y2 . Perhaps the region of interest is the portion of the uid surface above the ambient height h = 0. Given a reasonably close image. is an observed topological feature (say. Consider. and it is from this principle the power of the subject emerges. which are related by a map f : Y1 ! Y2 that factors through a map to X . Given a (cellular or continuous) map f : X ! Y . e.. Given f : X ! Y and g : Y ! Z . There is hardly a more important feature of homology than this. at the heart of algebraic topology. Passing to homology yields the induced homomorphism H(f ): H X ! H Y . then.68 structures on Chapter 4. For a continuous map. isomorphic homology: H (X R )  = H (R)  = H (XR ): One imagines this could be useful in controlling certain dynamical systems over networks where nodes carry information that evolves under the inuence of neighboring nodes (neighborhood connoting a heard or heard by relation).g. although H (X ) is hidden from view. the observed data comprises the homology of a pair of spaces Y1 . H (f ): H X ! H Y is a (graded) homomorphism.

In many applications in manufacturing. Let. Given a part orientation. Inverse kinematics 69 If the goal is to accurately capture the topological features of an exact but unmeasurable set. The available motions are controlled by the N rotational joints. then one has a simple diagram of the form: H ( ) * / H A / H A+ . the mechanisms are considered insubstantial. A = fh  0g be the desired but unseen set.11 Inverse kinematics Other applications of homology are obstructive.5 consisting of N rotational joints and rigid rods. one must perform part placement  manoeuver the arm so as to locate the part in the correct location and in the correct orientation. 4.14. H A where the map H () is the induced map on the inclusion  : A ! A+ that itself factors through inclusion to the invisible desideratum A. This simple observation is greatly extendable to the concept of persistence. is there a sequence of rotations to realize it? Certainly. Any nonzero element in the image of H () must factor through a nonzero homology class in A: one can discern the presence of a true hole with two imprecise observations and a map between them.11.12-5.4. Given a xed part orientation. as will be seen in Ÿ5.g.. a global version is not. . are all nearby part orientations realizable via small changes in the rotations required? This local problem seems to be a reasonable assumption. The critical issue is that of understanding and inverting the kinematic map.  is onto for a choice of axes which span R3 . It is the inverse kinematic map which is problematic. e. functoriality may assist. and no thought is wasted on the problem of self-intersection. As is common when mathematicians study robot arms. If one can measure approximants A  A  A+ from below and above and then match common features of these images. One topologizes the problem as per [?]: consider the kinematic map  : TN ! SO3 taking the ordered sequence of rotations to the net orientation of the part at the end of the arm. Consider the idealized robot arm of Example 1. grasping a part at the end of the arm.

8)  Note that 0 factors through  as 0 : Tn ! SO3 ! S2 by enacting  and ignoring all but the last axis in the frame. Then on H1 Z coecients one has. thanks to functoriality.6 and 4. there is no map Proof. Consider the case where the part grasped by the robot arm is rotationally symmetric about the last axis.12 Winding number and degree It is helpful to think of an induced map on homology as akin to a winding or linking: to what extent are the holes of X wound about the holes of Y via f : X ! Y ? The . since H2 (SO3 )  = 0 while H2 (S2 ) 6= 0. there is a diagram: Id H2 (S2 ) H (s 0 ) / H2 (TN ) K + 2 / 0 ) H2 (O S ) H (  KKK KKK H() KKK% H2 (SO3 ) (4. s 0 does not exist. This yields a contradiction: since ZN has no nonzero elements of nite order. however. from (/ ! TN with Id Z2 H (s ) / ZN H() Z2 . in . There is no analogous contradiction in assuming that the map H (0 )  H (s 0 ) is the identity. one is dependant on luck. functoriality reveals what individual homology groups do not. s : SO3 ! TN satisfying   s = Id. This decidedly non-intuitive result is a direct consequence of the algebraic topology of conguration spaces. The desideratum is no longer an orientation in SO3 but rather in S2 . Again.21 with the map 0 : TN ! S2 leads to frustration. One does not conclude that the inverse kinematic map s 0 : S2 ! TN necessarily exists.21.70 Chapter 4. Repeating the argument of Proposition 4. In fact. Note. an identity map. The top-row composition H (0 )  H (s 0 ) cannot be the identity map on H2 (S2 ). Consider a putative section s : SO3 Examples 4. Homology There is no continuous section to Proposition 4. H (s ) = 0 and H()  H(s ) = 0. the composition of which is H ()  H (s ) = H (  s ) = This result means that a continuous assignment of robot arm rotation angles as a function of part orientation is impossible. That is. in general. such as a bolt or peg to be inserted. since it is the identity homomorphism on H1 (S2 ) = 0.9. as revealed by H1 . that. At the level of H2 . Id. 4. the direction in which the last axis points. as H1 (S2 ) = 0. s = Id. There is no continuous assignment of rotation angles to a directional axis for a robot arm manipulating a part.

4.12. Winding number and degree

71

full complexity of homomorphisms between abelian groups gives an algebraic picture of the wrapping and winding that maps can execute. Indeed, induced homomorphisms are the right way to express the classical notion of winding numbers. A continuous simple closed curve : S1 ! R2 in the plane separates the plane into two connected regions: dim H0 (R2 (S1 )) = 2. The mod-2 winding number of about a point p 2 R2 not in the image of is, intuitively, the number which represents whether p is inside (1) or outside (0) the image of . Of the many denitions the reader may have seen (either involving integrals or the clever counting of intersections), the best is via homology. For : S1 ! R2 fp g, consider the induced homomorphism:

H( ): H1 (S1 ) ! H1 (R2 fpg):
Both the domain and codomain of H ( ) are of rank one, and the map H ( ) is therefore multiplication by a constant: that constant deg( ) is the winding number (perhaps in F2 or Z depending on the coecients used). There is no need to restrict to non-selfintersecting curves: any : S1 ! R2 fp g determines a homology class. If, instead of F2 coecients, the integers Z are used, then the class in H1 (R2 fp g; Z) is the winding number of about p. The winding number measures the algebraic number of times the image of wraps about p , with a choice of orientations (both of and R2 fpg) determining the sign. From the fact that Hn Sn  = Z, any self-map of Sn induces a homomorphism on Hn which is multiplication by an integer: this is the degree of the map. The resulting degree theory is an important classical topic, some of the important points of which are as follows: 1. 2. 3. 4. 5.
deg is a homotopy invariant. deg(f  g ) = (deg f )(deg g ). deg Id = 1. deg a = ( 1)n for a : Sn ! Sn the antipodal map a(x ) = deg f = deg g if and only if f ' g [the Hopf Theorem].

x.

Example 4.22 (Index theory and vector elds) The index theory for vector elds introduced in Ÿ3.3 is better understood in terms of degree. Let V be a vector eld on a manifold M with isolated xed point p , and let Bp be a suciently small ball about p. The index of V at p, IV (p), is dened to be the degree of the map V j@Bp : @Bp  = Sn ! Rn f0g ' Sn , where V is represented in local coordinates on Bp . This index is well-dened, since p is an isolated xed point. By the properties of degree, the index is independent of the neighborhood Bp chosen, so long as its intersects Fix(V ) in p alone. Example 4.23 (Linking number) Degree is well-dened for any map between oriented compact n-manifolds, since Hn  = Z for such. One example of this form of degree is in classical knot theory, the study of embeddings S1 ,! S3 . One intuits

72

Chapter 4. Homology

that there are many inequivalent ways to tie a simple closed curve in R3 . Degree can be used to characterize how two disjoint oriented knots link together. To dene the linking number of two disjoint oriented knots, K1 and K2 , parameterize each knot as 1 ; 2 : S1 ! S3 and dene `k (K1 ; K2 ) to be the degree of the map  : T2 ! S2 given by:

The reader may observe that this quantity is invariant of the parametrization chosen, so long as orientation is respected. Linking number has been found useful in a variety of contexts, from owlines in uid- and magnetohydrodynamics [?] to closed DNA strands [].

( ) ( )  : 1 ; 2 7! k 2 (2 ) 1 (1 )k : 2 2 1 1

(4.9)

which interpolate between ordered and disordered structure. Besides being of great commercial interest, liquid crystals oer fascinating observable defects whose classication is inherently topological. For concreteness, consider the nematic liquid crystals, composed of axisymmetric rod-like molecules whose alignment (or director eld), under a continuum assumption based on average behavior, is a continuous function of position apart from certain defects or disclinations [?]. A topological classication of such defects is intricate [], but an initial step uses degree as follows. Motivated by Example 4.22, one sets up a map whose induced action on homology yields an appropriate degree.

Example 4.24 (Nematic liquid crystals) Liquid crystals are a mesophase of matter

In a nematic liquid crystal that is taken so thin as to be approximated by a plane region to which the molecules align (a so-called Schlieren texture), the director eld  assigns to points in R2 a planar direction in P1 , well-dened and continuous o the set of singular point-defects. To assign an index to such a defect at x 2 R2 , choose an oriented small loop encircling x and consider the restriction  : ! P1  = S1 . The resulting index I (x ) 2 Z is an undirected version of that used for vector elds. In the literature, one often divides this integer quantity by 2 so as to obtain agreement with the vector eld index when the line eld is orientable [?]. In the case of a nematic in R3 , the direction eld  takes values in P2 . Following the planar case, choose a suciently small ball B around an isolated point-defect x

4.12. Winding number and degree

73

and consider the restriction of  : @B  = S2 ! P2 and its induced map H (): H2 (S2 ) ! 2 2 H2 (P ). From Example 4.9, H2 (P ; Z) = 0, and the (integer-valued) degree is trivial; however, passing to F2 coecients for H2 yields a well-dened index I (x ) 2 F2 , since H2 (P2 ; F2 )  = F2 . For singularities occurring along a disclination curve, index is measured by a loop  = S1 locally linking the disclination. The resulting map on 1 homology H (): H1 (S ) ! H1 (P2 ) is again going to return a degree I 2 F2 , which, like that of a point-singularity, is sensitive to orientability of the direction eld about the disclination.

Notes
1. The reader will note, perhaps with displeasure, that computational issued have not been discussed: it is too large and uid a topic. See [ ,

? ?, ?]

for an introduction to

the many techniques available for fast computation of homology, some of which are distributable (a near-necessity for realistic scientic applications). The short version

seems to be that cellular homology is computable in time near-linear in the number of cells. The curse of dimensionality (high-dimensional complexes have many, many cells) remains a challenge, however. 2. The Künneth theorem of example 4.6 holds for singular or cellular homology in eld coecients, for spaces with nite dimensional homology groups. There is a more general version which accommodates more general coecients. In the case of

Z coecients
?

and spaces with torsion, it is exact but algebraically delicate: see, e.g., [ ]. 3. Ÿ4.5 is the rst hint that taking advantage of negative dimensional chains and homology may be permissible. There are homology theories in which

H

takes on signicant

meaning for negative gradings. 4. Theorem 4.18 on the homology of relations is not the strongest possible result: as with the Nerve Lemma, a homotopy equivalence between the two complexes exists. 5. This treatment of winding numbers begins with an invocation of the classic Jordan Curve Theorem, which, in the language of this text, says: Let homeomorphic to

Sk .

Then

~p (Sn H

A) is Z for p = n k

1

A

be a subset of

and

0

Sn

else. Note that

there is no assumption of tameness in this statement. 6. Linking numbers are just the beginning of the algebraic topology of knots and links  a vast subject. Knot theory has found some applications, notably in physics, biology, uids dynamics, and dierential equations. 7. The treatment of singularities in nematic liquid crystals in Example 4.24 is woefully incomplete, given the tools here used. Very interesting phenomena occur as the eld is continuously changed and disclination lines are enticed to collide or entangle, or as point-defects move so as to encircle a disclination line. The homological index is incapable of handling the resulting (non-abelian!) phenomena, and tools from homotopy theory (see ŸA.3) are required. Beyond the simple nematics are other types of

liquid crystals, with bi-axial nematic, cholesteric, and smectic phases being the most common. 8. There are recent applications of homology and linking numbers to problems of in robotics [ ,

D  R2 that prevents the object from being able to be moved 2 `to innity' by means of Euclidean motions (translations and rotations) in R D. The
the smallest discrete set

? ?],

where, given a xed geometric object in

R2 , one wishes to choose

caging

related problem of performing motion-planning in robotics about obstacles by means

c .16. that X (X ) = whose homology ∑ k k ( 1) dim Can you see any hint of non-orientability from your computations? 4. What happens to linking number when the orientation of one or both curves is reversed? 4.13.24 change if the molecules were disc-like.11.  : Tn ! SO Tn ) to still have a lack of section s ? 2 of a planar robot arm holding a at part in Is there a homological obstruction to the section Can you construct s? s : SO2 ! Tn with   s = Id? 4.10.12.14. n n 1 ). Construct a counterexample to the converse.2. ? Exercises 4. where Sn 1 is an equator of Sn . 4. F). Prove that the degree of a map f : Sn ! Sn is zero if f R 3 is not surjective. F) with eld coecients Hk (X .15. What is the relationship between local homology at v and the link of v (as dened in Ÿ2.7. F2 ) and H (M . 4. for each k 2 Z. Give an explicit construction of a map 4.1. Triangulate a Möbius strip M and compute the homologies H (M . 4.74 Chapter 4.7)? Let X be the set of all web sites and R  X  X the links-to relation whose vertical point from each factor and identifying them all to a single point quotient map. Consider the kinematic map R2 . Let f : T ! T be given by the linear transformation R H (R). X with k  n + 1 simplices of dimension n H n (X ) = 0. and d a b c d ] constants.4.11 if self-intersections of the robot arm are forbidden? What would you need to know about the resulting conguration space of the robot arm (now a proper subset of 4. What happens to the results of Ÿ4. as opposed to rod-like? What if the molecules were bi-axial (having the shape of an ellipsoid with no rotational symmetries)? . for every compact space H (X . The 4. Show that an abstract simplicial complex must have 4. p via a X ) in terms of the individual homologies H (X ). 4. Speculate on the topology of with 4. Verify. 4. Give an example of a pair of knots in f : Sn ! Sn with degree k . has been computed or stated. b. Do your guesses change if X is the set of Twitter users the follows relation? [ for a. Z).8. Compute H (S . S Let X be a simplicial or cubical complex and v a 0-cell of X .9. What is the induced map H(f ) on homology (say. 4.6.2 for the case of cube complexes with F2 and Z coecients.3. Give a proof of Lemma 4. whose linking number is zero but which cannot be separated (let the reader consider what the correct denition of separated should mean). How would the classication of singularities in Example 4. Compute wedge sum of spaces ∨ p 2 X H ( ∨ X is the quotient space obtained by choosing a single ber over a xed web page are other web pages to which the base page links.5. in R coecients)? 4. Homology of specifying homology classes leads to computable optimization methods [ ]. 2 2 2 2 2 Consider the torus T as the quotient R =Z .

Chapter 5 Sequences .

5. 1] ! Y which restricts to f on X  f0g and to g on X  f1g. As sing (X )  sing (Y ). . homotopies. indicating the upshift in the grading.2) One calls F a map of degree +1. and. since. Sequences omology takes as its input a chain complex  a hierarchical assembly line of parts  and returns the global features. abelian groups. The extension to homotopy is more subtle. The appropriate generalization of a homeomorphism to chain complexes is therefore an invertible chain 0 .  : C ! C0 is a homomorphism F : C ! C0 sending k -chains homomorphically to (k + 1)-chains so that @ 0 F F @ = ' :  @ / / Cn+1 @ / Cn /  Cn 1 @ { { { { { F {{{ ' F {{{ ' F {{{ ' {{{   {{  {{ {{ {{ {  }{{  }{{ F  }{{ }{{ 0 0 0 / Cn /  / Cn / Cn  +1 1 @0 @0 @0 (5. @ 0 F F @ . they assemble into commutative diagrams in homology. / Cn+1 @ / Cn @ / Cn  @0 / 0 Cn 1 @ /  / (5. When combined with functoriality.1) '  @0 0 / Cn +1 0 / Cn  ' ' 1 @0  Commutativity means that homomorphisms are path-independent in the diagram. Recall that such. depending on one's preferences). homeomorphisms. a homeomorphism map  one which is an isomorphism for all Ck ! Ck sing sing f : X ! Y induces a chain map f : CX ! CY which is an isomorphism. e..  also have analogues at the level of chain complexes. or modules. Clearly. exact sequences. especially.1 Homotopy invariance Homology begins by replacing topological spaces with complexes of algebraic objects (vector spaces. g : X ! Y are homotopic if there is a map F : X  [0. gives the dierence between the chain maps. the algebraic analogue of a 1-parameter family. via respect for the boundary operators. This chapter unpacks these tools and puts them to use. A chain map is a map ' : C ! C0 between chain complexes that is a homomorphism on chain groups respecting the grading and commuting with the boundary maps.0 76 Chapter 5. neighbors are sent to neighbors. This is best expressed in the form of a commutative diagram:   '  @ = @ 0  ' . Note the morphological resemblance to homotopy of maps: a chain homotopy maps each k chain to a k + 1-chain. A chain homotopy between chain maps ' . Further topological notions  continuous functions. H = H f . etc. chain maps. Chain maps are the analogues of continuous maps.g. The elemental tools for analyzing homology are a similar linelike devices: chain complexes. The dierence between the ends of the homotopy.

2.2 Exact sequences n. Exact sequences 77 Lemma 5. Consider [ ] ' and  are H(' )[ ] H(  )[ ] = [('  ) ] = [(@ 0 F + F @ ) ] = [@ 0 F ] = 0. Corollary 5. g : . Two abelian groups are isomorphic.1. 1] as a family of singular k -simplices parameterized by the homotopy. G  = H . since is a cycle and @ 0 (F ) is a boundary. Assuming chain homotopic maps from C to C0 .4 (Exact sequence) The following simple examples of exact sequences help build intuition: 1. if and only if the following sequence is exact: /H /0 /G 0 . Singular homology is a homotopy invariant. Csing X ! Csing Y Homotopic maps f. This prism is then triangulated into singular (k + 1)-simplices that encode the homotopy.5. 5.g : X ! Y induce chain homotopic maps f . 2 H (C). For each singular k -simplex  .3. Chain homotopic maps induce the same homomorphisms on homology. @ ) is exact when its homology vanishes: ker @n = im @n+1 for all Exact sequences are as ubiquitous as the nullhomologous spaces they mirror. The following theorem is proved by constructing an explicit chain homotopy: Theorem 5. Example 5. The idea behind the proof is simple. Proof. A complex C = (C .2. one considers the F -image of   [0. It is shown that this chain map P (called a prism operator [?]) is a chain homotopy.

the following sequence is exact: ' of 0 / ker ' / G ' / im ' /0 Such a 5-term sequence framed by zeroes is called a short exact sequence. The rst isomorphism theorem for groups says that for a homomorphism G . the second map is injective and the penultimate map is surjective. This one exact sequence compactly encodes many of the relations of rst-year undergraduate vector calculus. 0 induces the Any short exact sequence of chain complexes / A i / B j / C /0. long exact sequence: / Hn (A ) H(i ) / H (j ) / Hn (B ) Hn (C )  / Hn 1 (A ) H (i ) / : (5. These t together into an exact sequence. Sequences 2. on R3 . the long exact sequence is functorial: exact sequences and chain maps a commutative diagram of short 0 0 / A / B / C /0 /0 f  /A ~ g  /B ~ h  /C ~ . Consider dierentiable functions C = C 1 (R3 ) and X.78 Chapter 5. More generally. The most important examples of exact sequences are those relating homologies of various spaces and subspaces. In any such short exact sequence. all curl-free elds are gradients and all div-free elds are curls. (5.4) Moreover. 3. the kernel and cokernel of a homomorphism ' : G ! H t into an exact sequence: 0 / ker ' / G ' / H / coker ' /0 4. Theorem 5. This sequence encodes the fact that curl-of-grad and div-of-curl vanish.5 (Snake Lemma). and the initial R term in the sequence represents the constant functions on R3 .3) where r is the gradient dierential operator from vector calculus. as well as the fact that. the vector space of vector elds on R3 . C1 0 / R / C r / X r / X r / C /0. The critical technical tool for the generation of such weaves an exact thread through a loom of chain complexes.

2. The induced connecting homomorphism  : Hn (C) ! Hn 1 (A) comes from the boundary map in C as follows: 1. 5.5) / ~ ) Hn (A  H(f )  H (g )  H (h )   H (f ) ~ ) Hn (B ~ ) Hn (C ~ ) Hn 1 (A An exact sequence of chain complexes means that there is a short exact sequence in each grading. By exactness. 4.5.3. = j (. Fix [ ] 2 Hn (C). 3. and these short exact sequences t into a commutative diagram with respect to the boundary operators. Pairs induces a commutative diagram of long exact sequences 79 / Hn (A ) / / Hn (B ) / / Hn (C ) /  / Hn 1 (A ) / / (5.

) for some .

j (@. By commutativity. 2 Bn .

) = @ (j.

By exactness. ) = @ = 0. @.

6) The connecting map  takes a relative homology class [ ] homology class [@ ] 2 Hn 1 (A). A): / Hn (A) H (i ) / H (j ) / Hn (X ) Hn (X. This is perhaps best grasped via animation. 2 Hn (X. Doing so solidies the invaluable technique of diagrammatic argument. This yields the long exact sequence of the pair (X. a static illustration is a poor substitute. no matter how wedded to geometric intuition. 5. Any topologist. A) to the . (5. A)  / Hn 1 (A) / . A) /0. The reader should demonstrate that  is well-dened and that the resulting long exact sequence is indeed exact. but nevertheless conveys the critical shift in grading that exactness and weaving enacts.! (X. or a subcomplex in the cellular).! X is an inclusion and j : (X. the following short sequence is exact: 0 / C (A) i / C  (X ) j / C (X.3 Pairs Given A  X (a subset in the singular category. A) is an inclusion of pairs. must possess a thorough understanding of the connecting homomorphism. ?) . Set  [ ] := [ ] 2 Hn 1 (A). = i for some 2 An 1 . where i : A .

in the cellular case. C (A + B)  = C (X ). Sequences Example 5. exactness yields the recursion relation ~n (Sk )  H = Hn 1 (Sk 1 ). one requires X = int(A) [ int(B ). @Dk ): / Chapter 5. redux) The computation of H (Sk ) can be carried out via . @Dk )  / Hn 1 (@Dk ) H (i ) / Hn 1 (D k ) / . consists of those chains which can be expressed as a sum of chains on A and chains on B. with chain maps  : c 7! (c. @Dk )  = ~ (Dk =@Dk )  ~ (Sk ). @Dk  = Sk 1 .4 Mayer-Vietoris Another important sequence is derived from a decomposition of X into subsets (or subcomplexes) A and B . Consider the short exact sequence 0 / C (A \ B )  / C (A)  C (B )  / C (A + B ) / 0. Beginning with the explicit and trivial compu~ (S0 ).6 (Spheres) Computing the homology of the sphere Sk is a simple appli- Hn (Dk ) H(j ) / Hn (Dk . In cellular homology with A. for all n > 1. C (A + B ). In both settings. Mayer-Vietoris as follows. By denition. and  : (a.7 (Spheres. This sequence captures the integrative nature of homology  one builds from local to global. one inducts to show that H ~n (Sk )  tation of H = Z for  n = k . c ). one shows (via the techniques of ƒech homology) that H (A + B )  = H (X ). For all n > 0. Example 5. In the singular category. As the H =H rst and last terms in the sequence above vanish. By excision. Let A and B be hemispheres intersecting in an equato~ (A)  ~ (B ). The term on the right. the resulting long exact sequence yields the Mayer-Vietoris sequence: / Hn (A \ B) H() / H( Hn (A)  Hn (B ) ) / Hn (X )  / Hn 1 (A \ B) / The connecting map decomposes a homology class in X into portions in A and B . 5. in the singular category. and = 0 else. one obtains by exactness that  : Hn (Sk )  rial Sk 1 . B subcomplexes. then takes the boundary of one of these portions in A \ B . H (Dk . As H = 0  = H = k 1 Hn 1 (S ) for all n > 1 and all k .80 cation of the long exact sequence of the pair (Dk . b) 7! a + b. subcomplexes suce. Hn (Dk ) = 0.

Sn ?)  = Z for any point ? 2 Sn . a network of security cameras [?])? To know whether a curve surrounds a point in the plane. if desired) to intersect the image of in a nite set of points fqi gK 1 . one needs to know whether a cycle in an ad hoc nonlocalized network surrounds a node (as in.5 Degree and local computation Computing degree (Ÿ4. touching the ray and immediately turning back. The net winding number of about p is the sum of these local contributions to degree. qi ): Then deg (5. Sn Q) / Hn 1 (Sn Q) : / Hn (Sn p)  H (f )   / Hn (Sn ) = / Hn (Sn . For example. At each intersection point qi . the winding number of about p is easily computed as follows. q ) = deg H(f ) : Hn (V.5.8) (Sn . The validity of local computation can be shown using basic tools: Theorem 5. What the curve does elsewhere is irrelevant. the curve kisses or crosses the ray. From Theorem 5. Dene the local degree of f at q to be deg(f . Degree and local computation 81 5. What happens if. p has discrete inverse image f 1 (p) = f= K ∑ i =1 deg(f . Draw a ray from p and perturb it (via transversality. This intuitively simple procedure has a rigorous footing in transversality [?]. since both are oriented) or it osculates. the long exact sequences of the pairs (Sn . V q ) ! Hn (U.. (5. it suces to know the local behavior of the curve at a (small) nite number of points. Sn p) form a commutative diagram: Proof. Sn Q) and Hn (Sn Q) / Hn (Sn )  H (j ) / Hn (Sn . given a point p 2 R2 and a closed curve : S1 ! R2 fp g.12) is often possible by means of local computations. e. Sn p ) H (j ) H (f ) H (f ) Hn 1 (Sn p)  H (f ) . Assume f : Sn ! Sn and q 2 f 1 (p ) is an isolated point in the inverse image. U p). This simple example of a local computation rewards rumination. better still is the use of local homology. The local degree is an integer since Hn (Sn . Assume that for f : Sn ! Sn .5. Q = fqi gK 1 .7) for U and V suciently small neighborhoods of q and p satisfying f (U )  V .g. Each action contributes a local degree: 1 if crossing and 0 if kissing. instead of a curve in the plane.8. This is the pattern and stamp of topology. either the curve traverses (left-to-right or right-to-left.5.

even maps of Sn Odd maps of Sn have odd have even degree. by commutativity ( ) H(j ) H(f ) n n n n n deg f = deg Hn (S ) ! Hn (S . =(2) i i where the isomorphisms come from (1) excision. By exactness. The induced map H (f ) : Hn (Ui . key to which is a long exact sequence to track antipodes. S Q) ! Hn (S . Recall from Example 1. Sequences The rst terms in both rows and the last term in the bottom row vanish. /0. /0 0 0 / C  (P n ) / C  / C (Sn ) / C  / C (Pn ) f   / C (Pn ) (P n )  f   f (Sn ) . this map is a 2-to-1 local homeomorphism. so that f  a = a  f . It follows that ( ) ⨿ ⊕ n n (1)  Hn (S . the quotient space Sn =a consisting of equivalence classes of antipodal points is one denition of the real projective space Pn . the lower map H (j ) is an isomorphism. Assume that f is odd. Theorem 5. ( ) ⊕ ∑ n deg f = deg Hn (S ) ! Hn (Ui . S Q) = Hn Ui . qi ) ! Hn (V.82 Chapter 5. p) = deg(f . the F 2 coecients. Q  Hn (Ui . There is a commutative diagram of short exact sequences of chain complexes in Proof.6 Borsuk-Ulam theorems There are a number of theorems which fall under the name Borsuk-Ulam. The proof in the (harder) odd case is sketched. Let  : Sn ! Pn denote the quotient projection map. Thus. The following is the key step in these results. p ) is by denition multiplication by deg(f . Thus.9 (Borsuk-Ulam). qi ). written in the language of degree theory.2. all concern spheres and the antipodal map a : Sn ! Sn that sends x 7! x . S p ) : ⨿ Choose a small neighborhood V of p so that f 1 (V ) = i Ui is a disjoint collection of neighborhoods of the qi . qi ): i i 5. and (2) behavior of homology under disjoint unions. degree. Let f : P2 ! P2 be the induced map. qi ). qi ) ! Hn (V.

5.6. Borsuk-Ulam theorems

83

where  is the transfer map, lifting a chain in Pn to a pair of chains in Sn . The F2 coecients ensures that the sequence is exact. The corresponding long exact sequences yields:

0 0

/ Hn (Pn ) H( ) / 

Hn (Sn )

H ( ) / Hn (Pn ) 

/ Hn  / Hn

1

(P n )
H (f )

/0; /0

H (f ) H (f )   H ( ) H ( ) / Hn (Pn ) / Hn (Sn ) / Hn (Pn ) H (f ) 

1

(P n )

By exactness and knowledge of H (Pn ), one argues that in the above diagram,  and H ( ) are isomorphisms, while H ( ) = 0. By inducting on the dimension n and using the commutative square penultimate to the right, one shows that H (f ) is an isomorphism. By commutativity, this implies that H (f ) : Hn (Sn ; F2 ) ! Hn (Sn ; F2 ) is an isomorphism. This, in turn, is the mod 2 reduction of H (f ) : Hn (Sn ; Z) ! Hn (Sn ; Z); i.e., deg(f ) mod 2 = 1.

There are a number of famous corollaries of Theorem 5.9 whose proofs may be found in standard texts [?, ?]. 1. 2.

faces (simplices of @ n+1 ) whose images in Rn intersect. 3. Stone-Tukey: Given a collection of n Lebesgue-measurable bodies in Rn , there is a hyperplane which bisects evenly the volume of each body. n 4. Lusternik-Schnirelmann: Any cover U = fUi gn 0 of S by n + 1 open sets must have at least one element Uj containing an antipodal pair of points.

Borsuk-Ulam: Any map Sn ! Rn must identify some pair of antipodal points. Radon: Any map n+1 ! Rn of an n + 1-simplex has a pair of disjoint closed

Several of these theorems have physical interpretations. For example, it is common to express the rst corollary above as saying that (assuming meterological continuity) some antipodal pair of points on the earth have the same temperature and barometric pressure. Applications more relevant to this text lie in economics and fair division problems: see, e.g., [?] for applications of Stone-Tukey to fair division. The recent book of Matou²ek [?] contains a wealth of applications to combinatorics.

84

Chapter 5. Sequences

5.7 Euler characteristic
Sequences give an easy justication of the topological invariance of the Euler characteristic. This comes from lifting the notion of Euler characteristic from a cell complex to an arbitrary (nite) chain complex C = (C ; @ ): ∑ (C) = ( 1)k dim Ck : (5.9) k The reason for the alternating sum is to take advantage of cancelations that permit the following.

Lemma 5.10.

The Euler characteristic of a chain complex and its homology are

identical, when both are dened.

and @ 2 = 0, one has dim Zk = dim from which it follows that

Proof. The proof for eld coecients uses simple linear algebra. Since Hk = Zk =Bk

Hk + dim Bk

and dim Ck = dim Zk + dim

Bk

1

,

dim Ck = dim

Hk + dim Bk + dim Bk 1 : k
telescopes.

Multiply this equation by ( 1)k ; the sum over

Corollary 5.11.

For a nite compact cell complex 

(X ) =

X with subcomplexes A and B,
(5.10) (5.11) (5.12) 

(A [ B) = (A) + (B) (A \ B) (X A) = (X ) (A)
It follows that

k

( 1)k dim Hk (X ) 

is a homotopy invariant among this class of spaces.

These results follow from applications of Lemma 5.10 to (1) the chain complex for cellular homology; (2) the Mayer-Vietoris sequence; and (3) the long exact sequence of the pair (X; A) respectively, the last requiring a little excisive eort to relate C (X; A) to X A.

5.8 Lefschetz index
There is a generalization of Euler characteristic from spaces to self-maps. For any chain map ' : C ! C on a nite-dimensional chain complex C, dene the Lefschetz index ∑ as the alternating sum of the traces on the induced homomorphism,  (' ) = k ( 1)k trace (H(') : Hk (C) ! Hk (C)). This index, like the Euler characteristic which it mimics, is intimately connected to the question of xed points, not

5.9. Equivalence of homology theories

85

of vector elds, but of self-maps in general. For a self-map f : X ! X of a nite cell complex X , dene its Lefschetz index as that of f , or, equivalently, H (f ): ∑  (f ) =  (f ) = ( 1)k trace (H(f ) : Hk X ! Hk X ) : (5.13) k

Theorem 5.12.
point if 

(f ) 6= 0.

For

X

a nite cell complex, any map

f :X!X

must have a xed

Proof. The technical portion of the proof (omitted) is to show that f may be approximated by a map (also labeled f ) that respects a cell structure on X (maps the n-skeleton X (n) to itself for all n) without modifying the xed point set. Assuming this, if f is xed-point-free, then the cellular approximation takes no cell to itself. Thus, the trace of the chain map f : C ! C vanishes. The analogue of Lemma 5.10 holds: the alternating sum of traces of H (f ) equals the alternating sum of traces of f via the same telescoping sum argument. Thus,  (f ) = 0.

points that is a small perturbation of the identity map. The Lefschetz index of this map is  (Id) = (M ), since the identity map on H has trace equal to the dimension of H . Indeed, it follows that for any nite cell complex X with  6= 0, any map f : X ! X homotopic to the identity has a xed point.

M with  6= 0 has at least one xed point. If a nonvanishing vector eld existed, the time- map of the ow (for  suciently small) would be a map of M without xed

This provides a simple proof of Theorem 3.3, that any vector eld on a manifold

5.9 Equivalence of homology theories
The equivalence of the various homology theories discussed can be shown in a number of ways. One clean approach uses a diagrammatic lemma from homological algebra.

Lemma 5.13 (The Five Lemma).
of the form

Given a commutative diagram of abelian groups 

= 



/ 

=
/ 

/ / 

/ / 



/   

=
/ 



(5.14) 

=
vertical maps are iso-

whose top and bottom rows are exact, and whose four

outer

morphisms; the middle vertical map is an isomorphism as well.

Theorem 5.14.

On cell complexes, singular and cellular homology are isomorphic.

Proof. The standard proof inducts on k , the dimension of the skeleta of the cell complex X and uses the long exact sequence of the pair (X (k ) ; X (k 1) ) of skeleta in S ) and cellular (H C ) homologies. For a restricted class of regular cell comsingular (H  plexes (attaching maps are homeomorphisms), the proof is slightly simpler. Choose

Sequences a cell-by-cell assembly sequence for X and induct on this sequence order. The problem of coverage. By induction and previous computations of the homologies of balls and spheres. The reader of this text will no doubt be led to thinking of coverage as a homological problem. The cell e is glued to Xi along its boundary @e . is the question of whether there are holes in the sensor network  are there any regions in D which are not sensed? The reader is likely to have encountered failure of blanket coverage generated by a network of cell phone towers: Coverage lost implies a hole. C @e Hn  S @e Hn / / H C Xi n Ce  Hn /  CX Hn i +1 / / C @e Hn 1 / C X  HC e Hn n 1 1 i /  =  = S X  HS e Hn i n SX Hn i +1 /  S @e Hn 1   =   = S X  HS e Hn n 1 1 i with vertical arrows induced from the map converting a cellular chain to a singular chain. One simple-to-state problem is that of coverage.10 Coverage in sensor networks Sensors  devices which return data tied to a location  are ubiquitous. homeomorphic to the sphere Sk 1 . with the ƒech theory as a particularly good source for tools. in which one wants to determine whether a sensor network separates D or surrounds a critical region. blanket coverage. As opti- . The Mayer-Vietoris sequences for (Xi . Both of these actions are assumed to be local in the sense that individual nodes cannot extract sensing data from or communicate data over all of D. as individual sensors have specic ranges and geometrically-dened coverage domains.86 Chapter 5. the time-dependent problem familiar to users of robotic vacuum sweepers. The nodes have two functions: they (1) sense a neighborhood of their locale in D. or more precisely. most coverage problems fall under the aegis of computation geometry [?]. The 5-Lemma completes the proof. thanks to progress in miniaturization and wireless communications. Although many sensors commonly observed are stand-alone or global devices. However. and sweeping coverage. and (2) communicate with other sensors. four of the ve vertical maps are isomorphisms. Assume that e is a k -cell which when added to Xi yields Xi +1 . there is an increasing push to network multiple local sensors. the proof is trivial. e ) in cellular and singular homology t together in a commutative diagram. Other important coverage problems include barrier coverage. The problem of collating distributed pieces of sensor data over a communications network is a substantial engineering challenge for which the tools of topology and homological algebra seem strangely tting. For X0 a point. 5. Fix a domain D  R2 and consider a nite collection Q of sensors in a the plane R2 .

for clarity and ease of proofs: 1. geometric probability becomes important [?].15 ([?]). then all of D is covered by the sensors if. Given Q  R2 . but asymmetric systems are entirely permissible in this framework. Each sensor is assumed to have a unique identication which it broadcasts. communications errors  are not modeled. Specic geometric assumptions are kept to a minimum. 4. It is satised by systems with radially-symmetric communications networks (or unit disc graphs) and radially symmetric sensing regions with the proper ratio between sensing and communication [?. in the context of randomly-placed sensors. As stated. D is contained in the coverage region of the network. and motivates the use of the ag complex F of the communications graph G to model the topology of the sensed region. it is explicitly simplicial. false echoes. The assumption that k -tuples of nodes in pairwise communication have a convex hull in R2 which is sensor-covered is meant to capture the correspondence between local sensing and local communication. Where a topological approach is protable is in the setting where location coordinates and distributions are unknown. The following theorem gives a criterion for coverage based on homology. Without the use of coordinates. Sensors are modeled as a collection of nodes Q  R2 . Communication is symmetric and generates a communications graph G on Q. and D as above. some important considerations  e. 5. equivalently: . Theorem 5. F. it is problematic how to know whether such a 1-cycle bounds a hole or not. ?]. 2. certain neighbors detect the transmission and establish a communication link. The following simple application of homology to blanket coverage makes use of several of the constructs of this chapter.4) to search for coverage holes by looking for a long 1-cycle in the graph.11. however. 5. Sensor coverage regions are correlated to communications: the convex hull of any subset of sensors S  Q which pairwise communicate is contained in the union of coverage regions of S . The Under these assumptions. 3.5. one wants to know whether critical assumption is the fourth.. These assumptions are chosen to be weak enough to be applicable in realistic systems. C. connecting the communications and sensing of the network.g. Homological coverage 87 mization problems  where should one place the sensors for maximal coverage  these are known as art gallery problems [?]. node failure. One xes a cycle C  G whose image in R2 is a simple closed curve bounding a domain D  R2 . G. one is tempted (as in Ÿ2. time-variability.11 Homological coverage Looking at a picture of a sensor network communications graph.

@ D)  H()  / H ( ) / H1 (@ D) Functoriality implies that the diagram is commutative: H ( ) = H ( ) .16) oo o o o H() H ( ) ooo wooo   / H2 (R2 . @ D).17) 1 The rst term in (5. @ D) = 0. Equivalence of the two conditions comes from the long exact sequence of the pair (F. Chapter 5. @ D) However. as the long exact sequence of the pair (Rp .88 1. C)  / H1 (C) : (5.12) of @ D about p 2 R2 . C) to (R2 . Commutativity of diagram (5. Proof. By assumption.! R .  / H2 (F. where Rp = R p  R .17) is zero since R2 p is homotopic to S . and. By construction.15) H2 (R2 . C) with  [ ] = @ = C. This map induces the following diagram on long exact sequences of the pairs:   / / H2 (F. C) induced by the inclusion  : C .! F. Consider the simplicial projection map  : F ! R2 which sends vertices of the abstract complex F to the corresponding node points of Q  D and which sends a k -simplex of F to the (potentially singular) k -simplex given by the convex hull of the vertices implicated. acting as a homeomorphism on the second terms of the pairs. the image of 2  H() : H1 (@ D) ! H1 (Rp ) = Z is the winding number (Ÿ4. @ = C 6= 0. C)  / H1 (C) H (i ) / H1 (F) /  . @ D) = 0. Exactness of (5. hence.16) completes the proof. If the sensors do not cover some point p 2 D. Sequences 2. then p does not lie in the image of . @ D) reveals: / H2 (R2 p) / H2 (R2 p . Since p lies in the interior of D. Commutativity implies that H ( )[ ] 6= 0. @ D)  / H1 (@ D)  = / H1 (R2 p) / (5. and thus H ( )[ ] 6= 0. the winding number is 1. since ker H (i ) = im  by exactness. There exists [C] = 0 2 H1 (F). @ D)  / H1 (@ D) H2 (R2 p . the map  : F ! R2 is a 2 2 2 2 composition of maps F ! R2 p . 2 H2 (R2 p .  takes the pair (F. moreover.17) implies that H2 (R2 p . The next terms are H1 (@ D)  = H1 (R2 p ) = Z. and H () is an isomorphism.15) thus becomes: H2 (F. Diagram (5. H ( ) [ ] = H ( )[@ ] 6= 0. C)  / H1 (C) /   :  (5. [ ] 2 H2 (F. . which has vanishing  H2 . thus.

18) This is motivated by. 2.. !    . de Silva. e. As such. The motivation is that.g. This leads to replacing the ag complex F of the communications graph with the Vietoris-Rips complex of Q.5. 5. Vietoris-Rips complexes) modeling a pointcloud data set. those homology classes which persist over a larger parameter range have greater statistical signicance. The assumption on sensor coverage is conservative  it species that certain regions are guaranteed to be covered while passing no information about lack of coverage elsewhere. Consider a ltration of a space X by subspaces Xi  a sequence of inclusions X0 . The work described here spans contributions of Carlsson. ! X1 .. Persistent homology 89 1. and others.g. 4. This branch of applied topology is advancing very rapidly. A relative cycle 2 Z2 (F. Edelsbrunner. . see [?] for initial works.12. is a conservative criterion. for a parameterized family of spaces (e. When the homological coverage criterion fails. like the assumptions on which it is built. the homological coverage criterion cannot be if-and-only-if. Zomorodian. C (X0 ) ! C (X1 ) !    ! C (XN ) = C (X ). [?] for a recent survey.12 Persistent homology The capstone example of this chapter is a short survey of the exciting work being done in data analysis using sequences and homologies. ! XN = X: (5. choosing a basis for H1 (F) which is sparse (in the sense of implicating few nodes and edges) gives information about where the coverage holes may reside. 3. a sequence of Vietoris-Rips or ƒech complexes of a set of data points with an increasing sequence of parameters (i )i . This allows one to conserve power or establish a sleep-wake cycle by homological means. It is common to use the assumption that nodes communicate if and only if they are within a certain radius. and [?] for a comprehensive text on the subject. It. C) with @ = C suces to cover D: only those nodes implicated in are required to be actively sensing/communicating. This topological picture is converted to a sequence of chain complexes.

but rather a direct sum of free and cyclic submodules. For the general case. (5. j )-persistent homology of P. That is. Sequences then. i !j (P) is i < j .16.19) are in bijective correspondence with those persistent homology generators which come into existence at parameter ti and which persist for all future parameter values. Theorem 5.90 Chapter 5. As the ltering of P is via chain maps x . but.17 (Structure Theorem [?]). Fix a eld of coecients F and place a graded F[x ]-module structure on P with x . as explained in [?].j The Structure Theorem has a natural interpretation. At the chain level. one begins with a persistence complex: a sequence of chain complexes P = (Ci ). sj . H (P)  = ⊕  i x t i  F [x ]   ⊕  j x rj  (F[x ]=(x sj  F[x ])) . the shift map. together with chain maps x : Ci ! Ci +1 .i . . unlike the chain module. the (i. of course. The free portions of Equation (5. with eld F For a nite-type persistence module P coecients. For There is a good deal more algebraic structure in the interleaving of persistent homology groups. a unit monomial x n 2 F[x ] sends Ci to Ci +n via n applications of x . The classication of F[x ]-modules follows from the Structure Theorem for Principal Ideal Domains.) Note that each Ci = (C. one does not index the chain maps x . after passing to homology in eld coecients. Denition 5. The parameter intervals arising from the basis for H (P) in Equation (5. since the only graded ideals of F[x ] are of the form x n F[x ]. as multiplication.19) inspire a visual snapshot of H (P) in the form of a barcode. the Structure Theorem provides a birth-death pairing of generators of P (excepting those that are eternal or co-eternal). ti gi. (For notational simplicity. denoted H j i dened to be the image of the homomorphism H (x ) : H (Ci ) ! H (Cj ) induced by xj i. @ ) is a complex. The resulting homology H (P) retains the structure of an F[x ]-module. is not necessarily free. to a sequence of (graded) vector spaces with linear transformations between them: H (X0 ) ! H (X1 ) !    ! H (XN ) = H (X ): There is.19) for some collection of integer constants frj . One assumes a nite-type condition that each Ci is nitely generated as a F[x ]-module and that the sequence stabilizes in i in both directions (in the case of a bi-innite sequence of chain complexes). A barcode is a graphical representation of H (P) as a collection of horizontal line segments . P is free as an F[x ]-module. i 2 Z. The cyclic elements correspond to those homology generators which appear at parameter rj and disappear at parameter rj + sj . no need to restrict to the case of metric-based simplicial complexes.

Fix a positive integer k > 0. prompting judicious use of density ltrations.13 The space of natural images One recent example of discovering topological structure in a high-dimensional data set comes from natural images. ?]. For any point x in the data set. the data set M lies on a topological seven-sphere S7  R8 . H . A barcode is an infographic of a parameterized dim dim i !j (P) H is equal to the number of In particular. the more averaging occurs among neighbors.18 (Barcode Theorem [?]). 5. j ]. T ] the subset of M in the upper T -percent of density as measured by k . Lee.5. and Pederson [?] sampled this data by choosing at random 5000 three-pixel by three-pixel squares within each digital image and retaining the top 20% of these with respect to contrast.000 vectors in R9 whose components represent grey-scale intensities. Denote by M[k. The full data set consists of roughly 8. intervals in the barcode of H (P) containing the parameter interval [i. Theorem 5. k is a positive distribution over the point cloud which measures the radius of the ball needed to enclose k neighbors. Values of k are thus inversely related to the point cloud density. The larger a value of k used. for reasonable values of . For a xed value of k . By normalizing with respect to mean intensity and high-contrast images (those away from the origin). dene k (x ) as the distance in Rn from x to k th nearest neighbor of x in the data set. dim H (Ci ) is equal to the number of intervals containing i . and by utilizing a certain norm for contrast [?]. This is a two-parameter subset of the point cloud which. A codensity function is used in [?] as follows.13. A cursory visualization reveals points distributed seemingly densely over the entire S7 . Mumford. The space of natural images 91 in a plane whose horizontal axis corresponds to the parameter and whose vertical axis represents an (arbitrary) ordering of homology generators. blurring ner variations. A collection of 4167 digital photographs of random outdoor scenes was assembled in the late 1990s by van Hateren and van der Schaaf [?.000.

This indicates that the data set is diused about a primary circle in the 7-sphere. . represents an appropriate core. computing the barcode for the rst homology H1 reveals a unique persistent generator. Sequences and T . However.3) provide small enough spaces for computations to be done quickly. Rather. one obtains predictive insight to the structure of the space of high-contrast patches. there are two secondary circles which come into view at the lower density parameter. with 5000 points sampled at random from M[k.92 Chapter 5. T ]. by focusing on the generators and computing the barcode for H0 . An examination of the barcodes for the rst homology group H1 of the data set ltered by codensity parameter k = 15 and threshold T = 25 reveals a dierent persistent H1 . Combined with the basis of H1 generators. This is not surprising: the lowest order terms in any series expansion are always most easily perceived. it is observed that. A closer examination of the data corresponding to this primary circle reveals a pattern of 3-by-3 patches with one light region and one dark region separated by a linear transition. The barcode reveals that the persistent H1 of samples from M[k. The dierence between the two secondary circles lies in their bias for horizontal and vertical stratication respectively. there is indication of a persistent H2 generator (in F2 coecients) at certain settings of k and T . yet the two secondary circles are disjoint. this generator is dominant at the threshold and codensity parameters chosen. As noted in [?]. witness complexes (as in Ÿ2. As seen from the barcode. The reduction in k leads to less averaging and more localized density sensitivity. This curve between light and dark is linear and appears in a circular family parameterized by the angle of the transition line. Taking a density threshold of T = 25 at neighbor parameter k = 300. In practice. The rst interesting persistent homology computation on this data set occurs at the level of H1 . The barcodes for the second persistent homology H2 are more volatile with respect to changes in density and thresholding. A close examination of k these three circles reveals that each intersects the primary circle twice. This does not connote the presence of ve disjoint circles in the data set. T ]. besides the primary circle from the high-k H1 computation. each secondary circle regulates images with three contrasting regions and interpolates between these states and the primary circle. T ] has dimension ve. The structure of the barcode is robust with respect to the random sampling of the points in M[k. At certain density thresholds.

Given a large set of nodes Q. One means of performing comparisons is to build the sequence of . it is not the dimensions of the homology that matter. one could compare the resulting homologies. it may be infeasible to construct the full Vietoris-Rips complexes and compute persistent homology. as in Ÿ5. The primary and secondary circles appear with the appropriate intersection properties. For example. a small sampling of points is taken. arise in consistency tests for sampling point clouds as follows. each sampling detecting a dierent one? Given an ordered sequence of small subsamples. build simplicial complexes Xi based on them. perhaps random or according to a witness complex construction as in Ÿ2. other non-monotone sequences are possible and relevant to data management [?]. In the literature.    !  !  !  !    . To check for accuracy of the sampling. however. 5. where arrows connote inclusion of spaces or chains or the induced homomorphisms on homology.13. Is it the same hole? Or are there many holes in the true data set. However. A comparison of homology computations in F2 and F3 coecients resolves the ambiguity that H2 (K 2 .5. F2 ) and veries 2 2 that the persistent surface found is K and not T .3. sequences of the form    !   !  . but the correspondence. F2 )  = H2 (T2 .14. Zigzag persistence 93 the H2 barcode suggests a two-dimensional completion of the low-k persistent H1 basis into a Klein bottle K 2 .14 Zigzag persistence The icon of persistence is the inclusion sequence. Assume that several samples of the data all detect the presence of a single hole.

with the Colored Tverberg Theorem being one of the most general [ ]. like the sequence of inclusions in Equation (5. and its proof is often by means of convex geometry tricks. Sequences X1 X. the ability to infer global features from a small number of local measurements is greatly desirable in scientic and technological settings.  !  !  !  !  . 5. is linear. The deeper meaning of the Radon Theorem is. Such homology classes would be consistent over samples. The next step is to build a ? The subject of homological algebra C. ?].12 is deep and signicant: any map between CW complexes can be homotoped to a ? cellular map relative to a subcomplex on which the map is already cellular [ ]. The Cellular Approximation Theorem casually alluded to in the proof of Theorem 5. where each  is a vector space over F and each ! is a linear transformation going Notes 1. 3. outtted with homomorphisms that turn the pages. the arrows alternate. Were the analogue of Theorem 5.1 [ X 2 cH HH vv HH v v HH v HH vv v v X2 X. a bi-graded 2-d array of chains chain maps satisfying @@ 0 + @ 0 @ = 0. ? . The recent results of Carlsson and de Silva [?. In general. does not exactly make for colorful reading. In consistency tests. with horizontal @ and vertical techniques lead quickly to a spectral sequence. like that of the Helly Theorem (see Chatper 6).5.17 to exist.18). see [ . combined with the mantra that complexes are algebraic representations of spaces. 2. the resulting zigzag persistence and associated barcodes are extremely powerful. This chapter on exact sequences is just the beginning of diagrammatic homological algebra. It is to be suspected that such generalizations of the Borsuk-Ulam Theorem are useful  perhaps to economics most readily. but [ ].94 spaces and inclusions: Chapter 5. is the best place to start. In classical persistence. ? ?.3 [ X 4 cH HH vv HH v v HH v HH vv v v X4 This sequence. though notationally intricate. Such structures. The moral of these latter sections and of this chapter is that the homology of a sequence is worth more than a sequence of homologies.2 [ X 3 cH HH vv HH v v HH v HH vv v v X3 X. all arrows go to the right. The Radon Theorem (and all the Borsuk-Ulam type results) are greatly generalizable. are quite powerful. ?] use representation theory to classify indecomposable submodules of the homology of linear sequences of the form either to the left or to the right. The Snake Lemma and 5-Lemma are two of many wonderfully useful general results in homological algebra: for others. One should not underestimate the utility of local degree computations as in Ÿ5. topological in nature. it would classify homology classes which persist over the entire sequence. @0 double complex. Such a structure reminiscent of a book whose pages are double complexes. The Radon Theorem is usually stated in terms of convex hulls of points in Euclidean space. 4.

A persistent homology approach is given in [ ]. To perform computations. 7. As stated. In a triangulated n 1 Rn . fortunately. More desirable is a decentralized or distributed computation. and comparisons justify this reduction [ ]. the use of Vietoris-Rips complexes versus ƒech in point cloud data. the criterion for computing homological coverage in sensor networks is centralized. See the books by Edelsbrunner-Harer and Zomorodian and the book-in-progress by Blumberg. but less clean than the 2-d case presented here. boundary of the domain in question. 10. 8. in the sense that nodes must upload connectivity data to a central computer.5. ? ?. in which case a simple modication of the existing proofs suce. The available perspectives on persistent homology for an author to choose from are daunting. Algorithms for decentralized computation of 9. ? However nearly any homology computation of a realistically large and high-dimensional data set is nontrivial. Carlsson and de Silva use the representation theory of ? quivers prominently in classi- cation of zig-zag persistence. ? ? . cf. it can be ecacious to pass to a witness complex. but some applications may permit this condition. 12. Random sampling. multiple trials. From a classication theorem of Gabriel [ ]. it is shown that not only linear sequences. Multi-dimensional persistence is a pertinent and dicult generalization. solved [ ]. one must specify cycle: this seems awkward to the author. and Vejdemo-Johansson for proper details from several perspectives. performable by ? nodes communicating with neighbors. it is not dicult to image choosing a cycle or even establishing a fence of sensors. and the treatment in this chapter is necessarily elementary. Zigzag persistence 95 6. The extension of the coverage criterion to higher-dimensional networks is possible. Why was the ƒech complex of the sensor cover not used to obtain a coverage criterion? Determining the depths of overlaps of coverage sets is work-intensive. The computation of persistent homology is a well-motivated problem which is.14. 11. Carlsson. The paper of Carlsson and Zomorodian [ ] gives an excellent introduction to the diculties inherent in the subject. homology have just recently emerged [ . but any Dynkin diagram has a nice zigzag persistence classication theorem. The diculty resides in specifying the For regions in the plane. ?].

Chapter 6 Cohomology .

The duality implicit in this theory is a deep and meaningful construct whose utility is subtle and easily underestimated. given a real-valued function f : M ! R. t together to form a bundle of vector space over M . This symmetry in counting manifests itself in numerous numerical miracles. but as a ruler eld  a eld of rulers with direction. but the pairing is not well-dened in general. i =1 i where dxi is the dual to the xi unit tangent vector.98 6. The cotangent space to a manifold M at p 2 M is the vector space dual The cotangent spaces. The dual space of a real vector space V is V _ . It is best to imagine a gradient df not as a vector eld. to the fact that (S2n+1 ) = 0.# The number of ways to choose k items from n  k is exactly the same as the number of ways to (not) choose n k items. the gradient of f is the 1-form df which. . There is a corresponding notion of duality for linear transformations. The manner in which duality presents itself in Topology is best discovered through the familiar constructs of linear algebra and calculus. Note how the dual transformation reverses the direction of the map: it is the archetypical construct of this chapter. # That @f @x dxi . the vector space of all linear functionals V ! R. ordinates used to express it. the dual map or adjoint of f is f _ : W _ ! V _ given by (f _ ( ))(v ) =  (f (v )). For example. like their tangent space duals. The Chain Rule implies that df is independent of the local codf = n ∑ grammatical dualities cannot fail to precede even these is not a false statement. If f : V ! W is linear. orientation. and scale  along which tangent vectors are measured. It is a common mistake to conate the gradient of a function with a vector eld rf : this is permissible in Euclidean space. The dual space satises dim(V _ ) = dim(V ).1 Duals The rst form of duality one encounters in mathematics is combinatorial. Example 6. Cohomology ohomology mirrors homology with all the arrows reversed. the cotangent bundle T  M . The analogue of a vector eld is a 1-form: a choice of Tp M continuous in p. + Chapter 6. in local coordinates fxi gn 1 evaluates to Tp M = (Tp M )_ to the tangent space. from the geometric symmetry of Pascal's triangle. and (V _ )_  = V for V nite-dimensional.1 (Gradients) Dual vector spaces play an important role in calculus on manifolds.

The coboundary d is the dual of the boundary @ . freely generated by (oriented) simplices f g in a simplicial complex X .3 (Integration) Consider a chain complex . Cochain complexes 99 6. dene   to be the evaluation taking the value 1 if and only if  =  and 0 else. an annulus  then one can construct a similar 1-cocycle which is nonvanishing in H 1 . For k -cell. with cohomology class in H 1 dierentiating between those which are or are not globally expressible as a gradient of a potential. This cocyle is thus the zero cohomology class. Example 6. The astute reader will notice the implicit relationship between such cocycles and gradients of a local potential over the vertices.6. This is the coboundary of a 0cochain which labels vertices on the left with zero and on the right with one. The simplest means of constructing cochain complexes is to dualize a chain complex (C .2. the module of Given such a complex (with coecients in. This notation is illustrative. C with R-coecients. @ ). The natural pairing between C and its dual C_∫in these bases permits an integral interpretation: for basis elements  and  . _ . R). if one considers a surface with some nontrivial H1  say. Hk (C) = ker d k / im d k 1 : Cohomology classes are equivalence classes of cocycles in ker d . Two cocylces are cohomologous if they dier by a coboundary in im d . so that d  d = @ _  @ _ = (@  @ )_ = 0_ = 0: The coboundary operator d can be presented explicitly: (df )( ) = f (@ ). dene C k = Ck homomorphisms Ck ! R. On the other hand. The dual basis cochains can be thought of as characteristic functions f g. Using linearity and the integral notation.2 Cochain complexes A cochain complex is a sequence C = (C  . one has (d  )( ) =  (@ ). say.  a Example 6. The cohomology of a cochain complex is. Consider a triangulated disc with a 1-cocycle on edges using F2 coecients. d ) of R-modules C k and module homomorphisms d k : C k ! C k +1 with the property that d k +1  d k = 0 for all k . By denition of the coboundary d as the dual of the boundary @ .2 (Simplicial cochains) Examples in the simplicial category are illustrative. d implicates the cofaces  those (k + 1)-cells having  as a face.

A classical set of problems in combinatorial optimization concerns ows on G. The cut capacity is the sum over the boundary of this set of all incoming edge capacities. cuts and ows correspond to various chains and cochains. etc. A for each node except s and r. A cut is a partition of nodes of G into two sets. The cut capacity is the sum of the edge capacities of the removed edges. . Choose two nodes of G to be the sender (s) and receiver (r) nodes. capacitors.) located in the interiors of edges. equivalently: 1. A cut is an open subset of G containing r but not s. the sum of the in-pointing edge ow values equals the sum of the out-pointing edge ow values: this net sum is constant over G. The problem is constrained in that each edge is assigned a capacity that dominates the possible ow value on that edge. Example 6.4 (Kirchho's voltage rule) Consider an electric circuit as a 1-dimensional cell complex. N or R+ ) to edges of G so that. 2. ∫ c d = ∫ @c : (6. Example 6. In the language of this chapter: voltage is a 1-coboundary. say. 3. Motivated by problems in transportation and railway shipping. one containing s and the other r. and r. s [ r. Cohomology 2 C p and chains c 2 Cp+1 . where. The min-cut-max-ow theorem states that the maximum possible ow from s to r equals the minimal cut capacity.5 (Cuts and ows) Consider a directed graph G: a 1-d cell complex with each edge oriented. In the language of this chapter.1) This reformulation should motivate scientists and engineers who know from early education the utility of Stokes' Theorem. the classical max ow problem seeks to maximize this sum for a choice of ow. Kirchho's voltage rule states that the sum of the voltage potential differences across any loop in the circuit is zero. Z). The cut capacity is the sum of edge capacities over all edges which point from the s node-partition to the r node-partition. A ow is a choice of relative 1-cocycle in Z 1 (G. s. A cut is a subset of edges of G whose removal disconnects s and r (there are no longer directed paths from sender to receiver). where ow on G is an assignment of coecients (in. given G. with circuit elements (resistors. and assume that the graph is connected from s ! r (respecting direction).100 one has for all cochains Chapter 6.

a cut can be viewed as a nontrivial relative cocycle in H 1 (G. The beginner may be deated at learning that cohomology seems to reveal no new data. Cohomology theories: cellular. Z). By dualizing results of the previous two chapters. A) H (X ) H (A) / . via the Snake Lemma (Theorem 5. 2. one immediately obtains the following standard results: 1. 0 / C  (X. F )  = The situation is more delicate in Z-coecients. but it is nevertheless true that cohomology is determined by knowing the homology and certain algebraic properties of the coecient ring. F)_ . Cohomology 101 the sign of the Z-coecients is consistent with the edge orientations. reduced. though less than intuitive. / H n 1 (A) H (j ) / n H(i ) / n  / n H (X. Alternately. For example. the Mayer-Vietoris sequence. especially when viewing a ow in relative H1 and a cut in relative H 1 . A cut is a 1-coboundary in B 1 (G) via (1) above. in this text.3 Cohomology The denition of cohomology in terms of dualizing chain complexes. is ecient. as it alike to homology in every respect except intuition.3. the induced and connecting homomorphisms have the same notation in homology and cohomology and are distinguishable via context and direction. ƒech. 6. one should repeat the mantra that duality often simplies algebraic diculties. Students may wonder why they should bother with cohomology. . Cohomological long exact sequences of pairs. The short exact sequence of cochain complexes. This manifests itself in. The astute reader rightly suspects duality lurking in the min-cut-max-ow theorem. Theorem 6. Hn (X . at least as far as linear algebra can count. e.6. It must be stressed that to correctly interpret and use these results arrows must be reversed. where.6 (Universal Coecient).g.5). choosing a 0-cochain constant on the sender-receiver partition: the coboundary of this 0-cochain has support on all edges between partition elements. H n (X . s [ r. A) as follows. the ow may also be viewed as a well-dened relative 1-cycle in H1 (G. in keeping with the way that duals of linear transformations behave. becomes. Since the graph is nite. relative.. singular. Z). s [ r. all with arbitrary coecients. In the beginning. the long exact sequence. A) j / C  (X ) i /  C (A) /0. the long exact sequence of a pair (X. 3. Functoriality and the homotopy invariance of cohomology. local. For X a space and F a eld. the induced homomorphisms on cohomology twist composition: H(f  g ) = H(g )  H(f ). excision.

Its best explanation  cohomology  is less popular. a suitable basis consists of characteristic functions of connected components of X . tori. the drawn perspective is locally realizable  one can construct a local depth function. Perhaps this is why customers are asked to conduct rankings (e. Based on data from spheres. where n is the degree of the original 0-cell dual.102 Chapter 6. An even more cartoonish example that evokes nontrivial 1-cocycles is the popular game of Rock. by denition. a global depth function cannot be dened. using projective geometry for coordinates [?]). this duality has a geometric interpretation.g. A local gradient of rock-beats-scissors does not extend to a global gradient.8 (1-Cocycles) One cartoon for understanding this distinction between 6. asking which-of-these-two-is1 better?): nontrivial H is. in this setting. Note the dierences. The impossible tribar is a cartoon of a non-zero class in H 1 (properly speaking. compact orientable surfaces. H 0 = ker d 0 . local and global coboundaries is the popular optical illusion of the impossible tribar.. necessarily. Elements of H 0 (X ) are locally-constant functions on X .e. Paper. These subtleties mirror the distinction between homology and cohomology. due to arrow reversal. undesirable. has as its 2-cells neighborhoods of the original vertices. In homology. The proof in the case of cohomology is simpler than for homology. since.. H0 determines the number of path-connected components (homologous 0-cycles are connected by 1-chain paths) while cohomology H 0 measures connected components (as seen by functionals). like that of H0 . Note that these cell decompositions are truly dual and have the eect of reversing the dimensions of cells: k -cells generating Ck . and not as a 1-cochain (i. and. yields Example 6. dim Hk M = dim Hn k M . for which there are local but not global ranking functions.7 (Connectivity) Note that the dimension of H 0 . The Condorcet paradox  that individually consistent comparative rankings can lead to global inconsistencies  is a favorite topic in voting theory. Each dual 2-cell is a polyhedral n-gon.4 Poincaré duality Homology and cohomology of manifolds have a dimensional symmetry as an expression of duality. At the level of cellular homology. R+ ). Example 6. There is nothing to quotient out. one might guess that for M a compact orientable n-manifold and coecients in a eld. H 1 (S1 . Netix movie rankings or Amazon book rankings) as a 0-cochain. Cohomology connectivity data. This is true and is a version of duality due to Poincaré. However. yielding a chain complex C. When one looks at the tribar. and let C be the cellular chain complex with F2 coecients. where the dual cell structure places a vertex in the center of each original 2-cell. There is a dual polyhedral cell structure. Consider a compact surface with a polyhedral cell structure. and the Künneth theorem in Ÿ4.2. Scissors. and likewise with cohomology. has 1-cells transverse to each original 1-cell.

F2 )  = H n k (M .. in which cases it is of rank one. The coecients may be modied at the expense of worrying about orientability of the manifold and the torsional elements: see. cellular) cochain complex C on a space X .g. a better explanation for the symmetries present in (co)homology is more algebraic.) The coboundary map restricts k ! C k +1 and d 2 = 0. along with Theorem 6. ?] for details. ?. 2. Hk  = _  H .9 (Poincaré duality). Hk (M . and proceeds in a manner that adapts to non-compact manifolds as well. F ). [?. that the diagram is commutative. crucially. 4. Though this style of proof generalizes to higher-dimensional H2 k  H = 2 k = 2 k manifolds. Hc mology satises the following: 1. .  is not a homotopy invariant. The equivalence of singular and cellular (co)homology. one orders compact sets by inclusion and uses induced maps to take a limit. (In the simplicial or cellular setting.g. implies that. Given a (singular. k (Rn ) = 0 for all k except k = n. The dual complex C consisting of C := C k and d := @ entwines with C in a diagram: 0 0 / / C2 @ / C1 / @ / C0 / /0 /0 C0   = d   = d   = C1 C2 The reader should check that the vertical maps are isomorphisms and. omorphism) invariant.6. F2 ) ! Hc 2 For M an n-manifold.7.  Hc (X )  = H  (X. To unwrap this (and state a general symmetry result). PD : Hk (M .9).3]. 3. With a slight change of perspective (in Ÿ6. for a compact surface with F2 coecients. Poincaré duality 103 are in bijective correspondence with (2 k )-cells (on a surface) generating a modied  k _ _ cellular chain group C 2 k . but is a proper-homotopy (and hence a homeHc Compactly supported cohomology compactly expresses the duality implicit in manifolds: Theorem 6. [?. consider the subcomplex C c of cochains which are compactly supported: each cochain is zero outside some compact subset of X .. a more precise form of PD will be given. phism n k (M . $ This is subtle: in brief. 3. This cohocohomology with compact supports.4. this is equivalent to building cochains from a nite number of basis cochains. there is a natural isomor- For M a compact n-manifold. F2 ). Hc  (X )  Hc = H  (X ) for X compact. yielding a well-dened to d : Cc c  (X ). a modied cohomology theory is helpful. simplicial. e. See. e. X K ) for K a suciently large$ compact set.

U A)  =(3) H n k (Sn . Recall from Ÿ3.104 Chapter 6. then it is potentially dicult to approximate those Euler characteristics. Choose a family of small open neighborhoods of A that continuously deform onto A. For k > 0. and (5) denition of U . Taking the ag complex of the network can lead to the existence of . (2) properties of compactly supported cohomology. However.4) suggest that the estimation of the Euler characteristics of the upper excursion sets is an eective approach. Given an integrand h 2 CF(Rn ) sampled over a discrete set. there ~k (A) ! H ~n k 1 (Sn AD : H One proof of this result wields many of the tools of homology and cohomology. =(5) H where the nontrivial isomorphisms come from (1) Poincaré duality. Slight modications are required for k = 0. computational formulae such as Equation (7. and denote by U a suciently small representative of that family. (3) excision. however. (Sn A) (Sn U )) = H n k (Sn A. This assumption of integrating sensors over a continuous domain is highly unrealistic. if the sampling occurs over a network with communication links. (4) the long exact sequence of the pair. The condition on A is a form of tameness and is crucial. Theorem 6.10 (Alexander duality). is an isomorphism For A  Sn A ).6 Numerical Euler integration Alexander duality is a key step in developing a highly eective numerical method for performing integration with respect to Euler characteristic over a non-localized planar network. n k (Sn A) Hk (Sn A)  =(1) Hc  =(2) H n k (Sn A. expressing the true answer as an integral over a continuum provides a clue as to how to approximate the result over a discretely sampled domain. a com- pact locally contractible (nonempty proper) subset. Cohomology 6. Among the most useful is Alexander duality.5 Alexander duality Poincaré duality can be adapted to several related settings involving manifolds. 6.7 that certain problems in data aggregation over a network are expressible as an integral with respect to d over a tame space X  Rn . U )  ~ n k 1 (U ) =(4) H  ~ n k 1 (A ).

7.6. 42 h d = 1 ∑ s =0 (.2) that ruin an Euler characteristic approximation. For ∫ h : R2 ! N constructible and upper semi-continuous. Convex geometry fake holes 105  higher-dimensional spheres (cf.11. Theorem 6. Ÿ2.

0 fh > s g .

0 fh  s g + 1) . (6.2) where .

so long as the connectivity of the excursion sets is properly sampled.12. Since A  R2 . . that The degree of sampling required to ensure exact approximation of ∫ G h d over a planar network G is h. in practice. the number of connected components of the set. Let A be a compact nonempty subset of R2 . (A) = 1 ∑ s =0 ( 1)s dim Hs (A). A) = dim H Since h is upper semi-continuous.0 = dim H0 . 6. By Alexander duality. Hs (A) = 0 for all s > 1. From the homological denition of the Euler characteristic and compactness of A. one has: ∫ A = fh > s g is compact. and it suces to compute (A) = dim H0 (A) dim H1 (A). which. are not dense enough to ll regions without any holes. dim H1 (A) = dim H 0 (R2 A. Using (co)homology yields a transparent proof. the set R2 A = fh  s g. these can be mitigated. Thanks to Alexander duality. correctly samples the connectivity of all the upper and lower excursion sets of This formulation is extremely important to numerical implementation of this integration theory to planar sensor networks. Proof. where H denotes singular homology in R coecients.7 Convex geometry The following theorem is a classic result in convex geometry and geometric combinatorics. Noting that h d = 1 ∑ s =0 fh > s g = 1 ∑ s =0 dim H0 (fh > s g) (dim H0 (fh  s g) 1) : Corollary 6. ~0 (R2 A) = dim H0 (R2 A) 1.

cohomology ows naturally from multivariable calculus [?. However. One benet of the topological approach is that extensions to the non-compact and non-convex world are natural and easily discerned. The algebra (V ) is graded: p=0 where p (V ) is the vector space of p -forms. N ' Sn has the homology of a subset of Rn . If the common intersection is empty. Thus. In other words. compact convex subsets of U Rn Let U = such that every (n + 1)-tuple U fU g be a collection of M > n+1 of distinct elements of has a point in common. Consider the nerve N of U. It is a subcomplex of the (n + 1)-simplex which. The induction step is a simple modication. Given a basis fxi gn 1 for V .3. 6. nonempty intersections have H n 1 (Rn A)  = 0. by hypothesis. in the sense that all sets and their ~ = 0. contains all faces. that dxi ^ dxj = dxj ^ dxi for all i and j . and the cover U must where. Scalar multiplication in (V ) is over R and the sum is induced by that on V _ . Proof. Given a R-vector space V . note. The critical ingredient is that the resulting cover U is acyclic. it is a good cover (all nonempty intersections are acyclic) and. Note that the hypotheses for Theorem 6. beginning at M = n + 2. meaning. it is impossible for a subset A  Rn to have the homology type of Sn via Alexander duality. These 1-forms combine to yield p-forms which acts on p-tuples of elements of V . A p-form takes as its argument an ordered p-tuple of (V ) = 1 ⊕ p (V ).8 Forms and Calculus In the setting of dierential manifolds. thanks to Theorem 2. in particular. N 6' Sn . It is alternating. with basis dxi1 ^    ^ dxip for 1  i1 <    < ip  n = dim V . Cohomology Theorem 6. . H have a common intersection point. Then all elements of have a point in common.13 (Helly's Theorem). The product in the algebra (V ) is called the wedge product and denoted ^.11 are satised.106 Chapter 6. and thus. let (V ) denote the algebra of forms on V  alternating multilinear maps from products of V to R. ?]. explicit generators for (V ) are given by the _ dual 1-forms dxi 2 V . then N = @ n+1 ' Sn . As the cover U is by convex sets. One begins with multilinear algebra. where dxi : V ! R returns the xith coordinate of a vector in V . Induct on M .  (U)  H = H (N)  = H ([ U ) : ~n Hn (A)  =H ~ 1 (X ) = 0 for all X nonempty. The reader may have seen proofs of Helly's Theorem based on convex geometry or functional analysis.

Recall from Ÿ1. On all manifolds. Likewise. d satises a Leibniz rule: d ( ^ . d is linear with respect to addition of forms and scalar multiplication. p (Tx M ). 2. A (smooth) vector eld V is a choice of elements of Tx M varying (smoothly) in x . parameterized by points in M . or. the vector spaces of p -forms. For example.6. Algebraic operations on  pass to operations on operating pointwise. As with the gradient 1-forms of Example 6. determinants.8. The space of all such sections  the p-forms on M  is denoted p = p (M ). form elds change from point-to-point: such changes are measured by a derivative. These constructions pass to manifolds. the p alternating property of ^ implies that  = 0 for all p > n = dim V . The appropriate dierential operator for forms is d : p ! p+1 . In like manner. p (M ) = 0 for all k > dim M . a p -form eld (shortened to p -form in practice) is a section : M ! p (T M ) giving x 2 p (Tx M ) varying smoothly in x . since dim 0 = 1. the wedge product extends to ^ : p  q ! p+q . more precisely.1. 0 (M ) = C 1 (M . the tangent bundle T M is a collection of n-dimensional vector spaces. The uniqueness of the determinant implies that dim  = 1. of which the cotangent bundle T  M is the case p = 1. dened implicitly via: 1. form a bundle  a family parameterized by points of M  of vector spaces. a section taking x 7! V (x ) 2 Tx M .3 that for an n-manifold M . R). Forms and Calculus 107 vectors in V and returns a real number in a manner that is multilinear and alternating. n cf. In the passage from  to .

) = d ^ .

+ d.

Example 6. Some of these forms can be written as dierentials = dg for some g 2 0 . 3. The 2-forms 2 are represented as . On 0-forms. where the coecient functions are smooth. ^ . d is the dierential d : f 7! df . the 1-forms 1 are repre- sented as = fx dx + fy dy + fz dz .14 (Vector calculus on R3 ) On Euclidean R3 .

and div tie together functions C = C 1 (R3 ) and vector elds X.  = C  r / r / X X   = = d  / 1 0 d  / 2 r / : C  = d  / 3 (6. d acts as the divergence operator. Every 3-form on R3 is of the form h dx ^ dy ^ dz for some h. There is a commutative diagram. and identify vector elds with 1. d is the curl operator. the vertical arrows identify 0. Recall from Equation (5.and 2-forms in the obvious . taking a function f not to its gradient vector eld. = gx dy ^ dz + gy dz ^ dx + gz dx ^ dy .3) On 0 (R3 ). but to the more natural gradient 1-form df . The dierential d is familiar to students of vector calculus. Here. curl.3) how grad.and 3-forms with functions. d is the gradient operator. on 2 (R3 ). On 2 (R3 ).

Cohomology ~ = Fx^ ^ ! Fx dx + Fy dy + Fz dz = F F i + Fy ^ j + Fz k ~ ~ = Fx^ ^ ! Fx dy ^ dz + Fy dz ^ dx + Fz dx ^ dy = .108 ways: Chapter 6.

~ 4 ~ rE c @t = r  B c J ~. tous in mathematical physics. and c are the electric eld. The calculus version of Maxwell's equations (on Euclidean R3 . Using the Euclidean structure to convert a eld F into a 1-form F or a 2-form . charge density. . B ~. J ~.F F i + Fy ^ j + Fz k ~ The reader should return to simplicial examples of cochains and coboundaries and be convinced that what is measured on the algebraic level is indeed a discrete analogue of gradients. magnetic eld. current. respectively. where E ~ and speed of light. and divergences. in a vacuum) are as follows: Example 6.15 (Maxwell's equations) The language of dierential forms is ubiqui- ~ ~ =0 rB c @t = r  E ~ 1 @E ~ = 4. Maxwell's equations admit a particularly simple interpretation. curls.

one obtains: ~ ~ F d (c E ~ ^ dt + . as in Example 6.15.

B ~) = 0 d (c B ~ ^ dt .

E ~ ) = 4.

the integral of d over an oriented piecewise-smooth closed curve is.4). The language of dierential forms is designed for integration: a p -form is perhaps best thought of as an object that can be integrated over a p -dimensional domain. precisely. the denition of the winding number of about 0: cf. Specically. In Example 6. the Chain Rule implies an invariance of the integral with respect to local orientation-preserving coordinate representations.J ~ ^ dt  dx ^ dy ^ dz.18.11. given an oriented p -dimensional ∫submanifold with corners. If one ignores the dextrous handwaiving details about induced orientations on boundaries. the fundamental theorem of vector calculus is clear: ~ 1 @B . S . Equation (3. These equations can be made more compact still and extended to arbitrary geometric manifolds using some of the constructions in Ÿ6. As is the case in the more familiar setting of line integrals in vector calculus. there is an integral operator S : p ! R dened by evaluating the p -form pointwise on oriented p tuples of tangent vectors to S and integrating on coordinate charts using the standard Lebesgue integral.

thanks to the Fundamental Theorem of Integral Calculus. This prompts the interpretation of as a complex: the de Rham complex of a manifold M is the cochain complex dR C = (  . d denes a non-trivial cohomology class [d] 2 dR H 1 (P ). d ) of forms with coboundary the exterior derivative d . scends to a product on cohomology. A de Rham cohomology class is an equivalence class of closed forms modulo exact forms. On the punctured plane P = R2 f0g.6. Despite being denoted d. dimensional submanifold with corners in M.17 (de Rham cohomology) The reduced de Rham cohomology of Rn is trivial for all n. dR H  (M ) is the cohomology of  .9 De Rham cohomology The antisymmetry property of forms is reminiscent of the cancelations via judicious 0 choice of ( 1) s at the heart of all homology and cohomology. De Rham cohomology 109 For Theorem 6.16 (Stokes' Theorem). it is traditional to denote the cocycles as closed forms and the coboundaries as exact forms. ∫ a p-form and S an oriented p + 1- ∫ S d = @S : 6. By dening [ ] ^ [. In this theory. thus. there is no single-valued 0-form  whose gradient 1-form is as above. The de Rham cohomology of M . Thanks to this antisymmetry and the commutativity of mixed partial derivatives. d 2 = 0.9. Example 6. the closed 1-form d = (x dy y dx )=(x 2 + y 2 ) is not exact.

] := [ ^ .

]. one notes that since and .

are closed. d ( ^ .

) = d ^ .

+ d.

^ = 0. furthermore. if or .

is exact. then so is ^ .

the wedge product inherits a volumetric interpretation. locally. complete with dierential d . the generator for dR H n (Tn )  = R is the volume form d1 ^  ^ dn . a basis Euclidean k -form measures oriented projected k -dimensional volumes. For example. Let  c (M ) denote the complex of compactly supported forms on M . A little more eort yields a calculus-based version of Poincaré duality. dR H (M )  = H  (M . It is no coincidence that the de Rham cohomology of Tn has the same dimension as in the singular theory. the 1-forms di . Since. yields a well-dened cohomology dR H c with compact supports. which. as an extension of Theorem 6. .18 (Wedge) In de Rham cohomology.20. . of course. Example 6. For M a manifold. This. ^ : dR H q  dR H p ! dR H p+q turns dR H  into a ring. Thus. is isomorphic to Hc . i = 1 : : : n generate the cohomology ring dR H  (Tn ). On the torus Tn with angular coordinates i .19 (de Rham). the wedge product for forms de- Theorem 6. R).

de Rham version). Cohomology For M an oriented manifold n. For [ ] 2 dR H p . wedge and integration of forms yields isomorphisms Theorem 6. R)  ! dR Hp (M )_ It is an instructive exercise to show that integration descends to homology and cohomology. of dimension ∫ M = p _  ^  : dR Hp (M )  ! dR Hn c (M ) ∫  =  : Hp (M .20 (Poincaré Duality.110 Chapter 6. .

then ∫ ∫ ∫ ∫ ∫ ∫ S+@T + d. 2 p 1 . S a boundaryless (p 1)-dimensional submanifold and T a p -dimensional submanifold with boundary.

= S + @S .

+ T d + @S d.

a piecewise-smooth vector eld V is likewise a 1-current. R) ! n p _ dR H c (M ) of Ÿ6. which may then be integrated over M (to a nite value. any (piecewise-smooth) oriented knot or link is a 1-current. @ ) with resulting de Rham homology dR H  (M ). the homology class of a vector eld. and there is a resulting complex (  (M ). R). let the reader assume (via restriction to the o-minimal structure of globally subanalytic sets) sucient (piecewise) smoothness where needed. To avoid the numerous technicalities involving regularity and rectiability [?]. Let p (M ) = ( p c (M )) be the space of p -currents  real-valued functionals on compactly supported p -forms. This section touches on analytic tools based on geometric measure theory [?. ?.4. Thus. = S . thanks to compact support of ). calculus provides the convenient language of forms for cohomology. since one can integrate a 1-form over oriented curves. is dened as the adjoint to the exterior derivative: @T ( ) = T (d ). only [S ] and [ ] matter.20 holds: dR H  (M )  = H (M . Clearly.10 Currents On smooth manifolds. The analogue of Theorem 6. Upon xing a volume form on M . since one can integrate a p -form over it: its de Rham homology class coincides with its singular homology class. A cycle is a current T with @T = 0. The isomorphisms of Theorem 6. since any 1-form pairs with V pointwise as (V ). Currents have a homological nature. . ?. The 2-currents on a manifold with volume form can range in shape from oriented surfaces to pairs of vector elds to a pair of tangent vectors at a single point. ?]. @ 2 = 0. inter alia. The chief advantage in using currents is their visualizability. An oriented p dimensional submanifold in M is a p -current. the boundary of T . via liberal use of Stokes' Theorem. Duals of forms provide an extremely exible interpolation between smooth and discrete homological structures on manifolds that allow one to talk about. For example. Given any p -current T 2 p . _ Fix M an oriented manifold of dimension n.21 eect the Poincaré Duality isomorphism PD : Hp (M . 6. @T 2 p 1 .

(6.25] that any volume-preserving vector eld on an oriented manifold can be realized as a limit of 1-currents supported on a compact 1-dimensional submanifold: an oriented link. . Arnol'd [?] (following Moat [?] ( following Cal ugarean u [?])) showed that the helicity is the asymptotic linking number: Theorem 6. and the energy of the uid . This implies that any volume-preserving ow on S3 is the limit (in the sense of 1-currents) of a sequence of ever-lengthening.10. This is of great signicance in uid dynamics.22 (Helicity) It has been known for a long time what is the appropriate asymptotic analogue of linking number for volume-preserving vector elds on S3 [?. since the velocity eld of an ideal uid evolves in time according to the Euler equations (Example ??). ∫ ∫ `k (V ) = 53 53 `k (x. yields the asymptotic linking number of the ow. Currents 111 Example 6. A volume-preserving vector eld V on S3 is closed and nullhomologous as a 1-current. y 2 S3 . II. normalized by st . this implies that the vector eld contracted into the volume form  yields an exact 2-form (V. Close these curves with short paths and compute the linking number (welldened for almost all x and y ). it follows from a result of D. until the owlines come close to their starting points (that this happens for almost-every x and y innitely often follows from the Poincaré Recurrence Theorem for volume-preserving ows [?]). since linking numbers are unchanged by such. ?. t ! +1 to a function `k (x. Prop. the helicity is an invariant of V under the action of volumepreserving dieomorphisms of S3 . For example. The limit of this linking number. The construction is as follows: given any two points x. ever-coiling links. y )dx dy: The techniques of forms and currents makes the computation of this seeminglyintractable quantity elementary. which.6.21 (Volume preserving links) One beauty of the language of currents is that it allows one to compare both knots and vector elds on a manifold. This suggests a reformulation of knot/link theory in terms of volume-preserving vector elds on S3 . y ). ) = d for some .4) 53 One shows well-denedness with respect to choice of via Stokes' Theorem. ?]. Sullivan [?. evolve them forward under the ow of the vector eld for times s and t respectively. What is the appropriate vector eld analogue of the unlink? Example 6. H(V ) = `k (V ) As a corollary.23 (Helicity Theorem [?]). converges as s. The helicity of V is the integral of the wedge product of with its dierential: ∫ H(V ) := ^ d . when integrated over S3  S3 with the conserved volume form.

Recall that for a vector space V of dimension n. Dene the Hodge star ? : p (V ) ! n p (V ) on basis elements as follows: ?dxi1 ^ dxi2 ^    ^ dxip := dxip+1 ^    ^ dxin 1 ^ dxin . It is. for lack of a better explanation in this text. i and choose an orthonormal basis fxi gn 1 . there is another manifestation of Poincaré duality in cohomology for manifolds via partial dierential equations.112 Chapter 6.g. reected in the sign change. The conormal cycle of a tame set A  Rn is a particular n-current CA 2 n (T  Rn ) on the cotangent bundle. For n odd. ?].. Each of the intrinsic volumes k of Ÿ3. When subtracting the conormal cycle C@Dn from  (D n D C . 6. The normal cycle of a tame set A  Rn is a special (n 1) Rn  current NA on the unit cotangent bundle T1 = Sn 1  n n R . The proof of the Helicity Theorem in [?] uses currents on S3  S3 to capture linking behavior.11 Laplacians With the addition of a geometric structure. the normal cycle is best visualized as having support on the set of points a `unit' distance from A. e. including Euler charac1  n teristic  = 0 . displays a combinatorial (n ) duality: dim p (V ) = p = dim n p (V ). Fix a geometry on V in the form of an inner product h. the algebra of forms.5) The valuation property of the intrinsic volumes is expressed in terms of addition of currents: e. For A  R of positive codimension. This means.24 (Normal and conormal cycles) Many of the constructs of Chapter 3 k (A) = ∫ NA k = ∫ CA !k : (6. that one can see the dierence between the Dn Euler characteristic of a compact disc (Dn ) = 1 and of its interior (  ) = ( 1)n = n @Dn ) as being a reection. . can be dened as the integral of a canonical form ( k 2 n c (T1 R ) n  n or !k 2 c (T R )) over the appropriate cycle: Example 6. the cone over the normal cycle..10. Cohomology (the integral of the norm of the velocity eld) is bounded below by helicity  H is a topological measure of a uid's inability to relax [?. NA[B = NA + NB NA\B and likewise for C. (V ). this results in an orientation-reversal. concerning Euler characteristic and intrinsic volumes have a representation in terms of currents.g. but the orientation in each axis is reversed. the support in Rn is the same. Fix also an orientation on V in the form of an equivalence class of orderings [xi ]n 1 of basis elements up to even permutations.

? F ~ = . If an oriented manifold M is Riemannian.11.6. then the Hodge star extends to ? : p ! n p . ) on tangent spaces of M . in Euclidean R3 with the standard basis as per Example 6. It satises a signed duality ?? = ( 1)p(n p) Id. not on the basis itself. meaning that there is a smoothlyvarying inner product g (. For example. Extend to all of (V ) via linearity. The Hodge star depends only on the inner product and the orientation. Laplacians 113 where the ordering [xij ] respects orientation.15.

The Hodge star yields an inner product on each p via integration: ∫ h . Every Riemannian manifold has a well-dened volume form  2 n which. is dx1 ^    ^ dxn and which is given by  = ?M .F ~ . in local orthonormal coordinates. .

i := M ^ ?.

d.: With this geometry in place. one may dene a codierential  : p ! p 1 given by the adjoint: h .

i = h . .

The harmonic forms are dened as H  (M ) := ker . Let 2 Hp be a harmonic form. This implies that ? is also harmonic.26.25 (Hodge Theorem). i. H (M )  = dR H  (M ).26 implies that d =  = 0. cf. p = d p 1  H p   p+1 : Corollary 6. For M a compact oriented Riemannian manifold. Theorem 6. ? is an incarnation of Poincaré duality. Theorem 6. and topological features. The Laplacian carefully entwines analytic. One of the benets of using dierential-topological constructs is the ability to import and export ideas between smooth and discrete frameworks. There is a simple simplicial analogue of the Hodge theorem which has the advantage of requiring no . The Laplacian is the operator  :  !  given by:  := (d +  )2 = d + d: Note that the Laplacian is degree zero. geometric. since (? ) = ( 1)p(n p) (d?d? + ?d?d )? = (d?d + ?d ) = 0: The trouble of working with geometry has the following payo: the Hodge star One may therefore realize the Poincaré dual as ? : H k (M ) ! H n k (M ).  := ( 1)p(n p) ?d?. more explicitly.21. the kernel of the Laplacian. For M a compact oriented Riemannian manifold. p (M ) has an orthogonal decomposition: Theorem 6. and for p = 0 is the familiar second-order dierential operator.

. Thus. has 0 as an asymptotically stable solution if and only if ker  = 0. by solving a heat equation with random initial conditions. Cohomology forms.  i = w . or manifold structures. A cohomological approach becomes the appropriate tool for addressing a related problem of coordinatizing a point cloud.27 (Distributed homology) The Laplacian is a local operator and. Consider a cell complex X with cellular cochain complex C = (C  . and VejdemoJohansson [?] is a slick application of algebraic topology that highlights the particular benets of cohomology and the role of coecients. d dt =  . but merely an implicit geometry.114 Chapter 6. The outline of their work [?] is as follows. S1 ] of maps X ! S1 is isomorphic to H 1 (X . as such.12 Circular coordinates in data sets In Ÿ5. One begins with the following result from homotopy theory: for any space X . As in the smooth theory. Morozov. i so that indicator functions over the cells of X are orthogonal. dierentiability. The solution of de Silva. Choose an inner product h. The work of Tahbaz-Salehi and Jadbabaie [?] details the use of the simplicial Laplacian to distributed computation of the homological coverage criterion of Ÿ5. the harmonic cochains are H := ker . Many of the existing algorithms for assigning circular coordinates to a point cloud [?] have some convexity assumptions: X must be either convex [?] or the isometric embedding of a convex set [?]. By Theorems 5. use the adjoint  of d to dene the combinatorial Laplacian:  = d + d . The coordinatization function  : X ! S1 therefore is naturally approachable via cohomology. d ) in R coecients. is well-suited to distributed computation.13 the problem of determining the topology of a point cloud was addressed by means of persistent homology. Example 6.15 and ??. the group of homotopy classes [X. It is easy to see that the heat equation. one can safely (to the degree one trusts in random initial conditions) conclude coverage if the solution converges to zero. The implicit choice is in the weight of each simplex h . Z). The following combinatorial Hodge theorem is a simple exercise. 1. With this inner product structure. 6. The paper [?] details the use of gossip-style algorithms for computing ker  in a distributed manner. Assume for the present that Q  Rn is a pointcloud whose topology is known or suspected to be sufciently circular so as to merit outtting circular coordinates  : Q ! S1 in a manner that respects the underlying topology of the space X  Rn (homotopic to S1 ) that Q is presumed to sample.11. veried coverage in a sensor network in R2 follows from showing that ker  = 0 on combinatorial 1forms of the ag complex F of the network. That this equation can be solved locally and in a distributed manner should come as no surprise to the reader who has spent time with the heat equation.

and (persistent) classes in Fp coecients therefore lift to integral classes. lift to R coecients and nd a cohomologous harmonic cocycle 2 H 1 (X ). Thanks to the local-averaging properties of the Laplacian. Z) p / H 2 (X . let C k be the set of functions (not k +1 necessarily maps!) f : X . The short exact sequence of p coecients 0 ! Z ! Z ! Fp ! 0 yields a short exact sequence of chain complexes on X and. To nd a cohomology class for X based on a sampling of nodes Q. this 1-cocycle integrates to a well-regulated coordinate function  : X ! S1 .12.5.17). 4. see [?]. Z) ! H 2 (X . a long exact sequence on cohomology: / H 1 (X . ker (p ) = coker  . for numerical reasons (to avoid roundo errors). Z) The kernel of p : H 2 (X . Z) consists of p -torsional cohomology classes: for p > 2 these would seem to be rare occurrences in organic spaces X living behind data sets. spaces. The resulting integer cocycle is perhaps a poor S1 -coordinatization  all the circular motion may be concentrated over a small subset of X . The conation of objects with duals is ubiquitous and insidious. The idea of the impossible tribar as a 1-cocycle was suggested by Penrose [ ] (and was noted to the author by Vin de Silva). : : : . The Alexander-Spanier cohomology of X is H  (C=C0 ).6. A (persistent) class [ p ] 2 H 1 (X . while highlighting the delicate interplay between real. Z) / / H 1 (X . 4. even in the (barely permissible) nite-dimensional case. This work illustrates well the utility of cohomology. above all in this text. 3. : : : . This chapter. [ ]. move on with all haste to. xi 1 . it implies that H 1 (X . x ) forms a subcomplex C0 . By the Structure Theorem (Theorem 5. Fp ) is surjective. when this is zero.) 3. compute the persistent cohomology of a sequence of Vietoris-Rips complexes. : : : . Given . Fp )  / H 2 (X . Fp ) can be converted to a integral class [ ] 2 H1 (X .g. Z) by means of the following process. By exactness. Circular coordinates in data sets 115 2. integral. coecients in a nite eld Fq are preferred. for reasonable  spaces. xk ) to the sum i ( 1)i f (x0 . isomorphic to the singular H (X ). For details on computational aspects and implementation. One of the many cohomology theories not covered in this text is related to conguration ? X a topological space and 4 a ring. Notes 1. motivational text such as this cannot do justice to cohomology theory.13. to learn the theory properly.. Let the reader resolve not to do so.125. xi +1 . It is. e. as in Ÿ5. Examples include ? The interested reader should confusing gradient 1-forms with vector elds and dening simplicial chains as functions from simplices to coecients. this requires eld coecients. Z) ! H 1 (X . xk ). 2. (The author hopes that sermons such as this will help expiate his own sins in this regard. and cyclic coecients. : : : . is woefully incomplete: a short. via Theorem 5. To relax to a smooth circular coordinate system. The set of all functions which vanish in a neighborhood of the grand diagonal (x. This gives a complex C with dierential d taking ∑ !4 f (x0 .

That is has a purely topological proof is a testament to the power of topological methods. (It holds in non-tame cases. Consider a point charge at rest in Rt . and helicity. What is the analogue of Alexander duality for subsets of relationship between 6. 7. Many subtler and deeper invariants come from other PDEs using auxiliary structures (gauge theory.3. x3 ) is a well-dened volume-preserving Compute its asymptotic linking number . Show that any compact boundaryless odd-dimensional manifold has vanishing Euler H (A) and H (Rn A)?. R). There is. The homological proof was known to Helly [ ] and many topological generalizations exist [ ]. @M ) 6. Hodge theory is merely a hint at how partial dierential equations (in this case. a point charge in 2 R3 is a of its exterior. 9. characteristic. as one might suspect. Show that helicity is independent of the choice of 1-form 6. Laplace's equation) on geometric manifolds can lead to topological invariants. the Euler characteristic dened using Hc T2 . 6. not just smooth manifolds. generator of H k (M )  = Hn 3 R and let X be the and 2 dR H (X ).7.) 6. as in the denition of the o-minimal Euler characteristic of Chapter 3. Rn ? Rn Specically. what is the 6. of which Poincaré. and Lefshetz are emanations.116 Chapter 6. Alexander. ? ? ? Verdier duality + for sheaves is perhaps the best encapsulation of the scope and power of duality theorems in (co)homology. making singular cohomology H a ring for any space. c (Rn ) = ( 1)n . and P2 in Z coecients by hand. 8. Hk (M )  = Hn Lefshetz duality: k (M. ∑ 1 Show that. x1 . x4 . @M ). 6.2. K 2 . n with exterior of the point times Assume that the magnetic eld vanishes and that the current and electric elds are radial.4. Helly's Theorem is important in a wide array of combinatorics and optimization problems: see [ ]. It is remarkable how many experts nevertheless consign Helly's Theorem to convex geometry. A more challenging exercise is to show that for any locally compact tame set X .20) uses a double complex (see the notes to Chapter 5).1. Consider  : thus. x2 . Seiberg-Witten equations) and have implications in string theory and algebraic geometry. k k =0 dim Hc (X . How much can you say about the orbits of this eld? . Compute the simplicial cohomology of 6. Exercises 6.5. but requires more delicate hypotheses and analysis. c (X ) := c . one unbounded. the time axis. A very clean proof of the de Rham Theorem (6.8. a deeper form of duality. Prove the Jordan Curve Theorem: any (tame) subset of divides Rn homeomorphic to Sn 1 into exactly two connected components  one bounded. Show that the vector eld vector eld on S3 : V =( it is called the Hopf eld. Give a proof of boundary @M . Use Maxwell's equations to show that the Faraday form is exact but the Maxwell form generates H Thus. for M a compact manifold of dimension k (M. The wedge product algebraic cup product ^: H p  H q ^ in de Rham cohomology is the smooth counterpart to the more ! Hp q . Cohomology 5.6. c (X ) = (X ). 6.9.