101 views

Uploaded by Joseph Heavner

These are my notes for the graduate course Fundamental Concepts of Topology at University of Maryland (College Park), Fall 2016, taught by Jonathan Rosenberg. The course webpage is https://umd.instructure.com/courses/1201901, and that is where the homework can be found. I hope to update these notes every few weeks as the course progresses.
Any comments or corrections should be sent to jheavner (at) umd (dot) edu.
Note: I have dropped this class for various reasons, and I have largely stopped attending lecture. I would not expect more notes to be posted, but there is a small possibility that they might be.

- Electromagnetism MIT 8.022
- Math 600 (Graduate Algebra, UMD) Notes (Sept. 2016)
- Math 630 (Graduate Real Analysis, UMD) Notes (September 2016)
- Elliptic Curves - Physics Research Interaction Team, UMD, 2016 - Notes
- MAT-208-H Chapter 1.10 Problem Set Solutions
- Leontief Input-Output Model Example Problem
- [T. W. Korner] a Companion to Analysis a Second F(BookFi.org)
- Willard - General Topology (Solutions)
- Companion to Lang's Algebra (G. Bergman)
- Zill Complex Analysis Solutions 3 (Misc, Ch. 2.5-3.2 without 2.7)
- Topology Without Tears
- Cw Complexes
- Zill Complex Analysis (ed 2) Solutions 4 (Chapter 4.1-4.4)
- Notes - Demystifying the Higgs Boson With Leonard Susskind
- 00 Metric Spaces
- 9780387244136-c2
- Isomorphisms in Linear Algebra: A Formal Approach
- The Special and General Linear Groups
- Hw7Sol_171.pdf
- Universitext Karen E. Smith Pekka Kekäläinen Lauri Kahanpää William Traves an Invitation to Algebraic Geometry Springer 2000

You are on page 1of 10

Lecture Notes

Fall 2016, Prof. Rosenberg

Notes by Joseph Heavner

September 5, 2016

Abstract

These are simply a retying of my in class notes, perhaps with a few

gaps filled in. They are not sufficiently detailed to replace our course

text (Hatcher and some outside references such as Taos rapid overview

of point-set topology) or Prof. Rosenberg lectures, but these notes may

serve as a good supplement to such resources, particularly for review purposes. Any special material should be marked in some way or another

as supplementary. The course webpage is via Canvas/ELMS and can be

found at https://umd.instructure.com/courses/1201901.

Send any corrections or suggestions to jheavner (at) umd (dot) edu.

Contents

1 August 29: Review of Point-Set (General) Topology

1.1 Logistics . . . . . . . . . . . . . . . . . . . . . . . . . .

1.2 Main Lecture . . . . . . . . . . . . . . . . . . . . . . .

1.3 Preview . . . . . . . . . . . . . . . . . . . . . . . . . .

I

. . . . . .

. . . . . .

. . . . . .

1

1

2

4

2.1 Main Lecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.2 Preview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5

5

7

3 September 2: Review

3.1 Last Time . . . . .

3.2 Main Lecture . . .

3.3 Preview . . . . . .

8

8

8

9

1

1.1

of Point-Set (General)

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

Topology III

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

Logistics

The syllabus is on Canvas. We will have three exams, as required by the university, including a midterm on September 30, a midterm on November 7, and

a final on December 16 (with the finas date being set by the university). Each

midterm is worth 100 points, and the final is worth 200 points. Homework totals to 200 points, so that the grade breakdown is 1/3, 1/3, 1/3. We will cover

1

chapters 0, 1, and 2 of Hatcher. Roughly, the plan is to cover point-set topology for 1 week (outside resources such as Tao and Goodwillie), basic geometric

topology for about 3 weeks (Chapter 0), the fundamental group and covering

spaces for about 3 weeks (Chapter 1), the classification of surfaces for about 1

week (Chapter 1 and outside sources such as Zeeman and Koch), and finally we

will cover simplical, singular, and cellular homology for the last 7 weeks of the

course (Chapter 2).

1.2

Main Lecture

angles and all the specifics, but we keep adjacency and still cover shapes like in

geometry, just less quantitatively.

Definition. A topological space is a set X with the family T of subsets of

X such that: (1) , X T , (2) T is closed with respect to arbitrary unions

(Pi T = i Pi T ), and (3) T is closed with respect to finite intersections

(Pi T = ni Pi T ). The elements of T are called the open sets of X

or of the topology (X, T ). (We will frequently use the abuse of notation that

(X, T ) = X when the topology is clear.)

If O T , then X O is called a closed set.

Examples. For all X the discrete topology is where all sets are open, and the

indiscrete topology is where only the empty set and the whole set are open. We

have better examples, for instance R with the standard topology so that a set

is open iff it is a union of open intervals Ii = (a, b) a, b R. (The discrete

topology is uninteresting here because it is uncountably infinite.) This should

serve as some intuition. We now look at a special case of a topological space.

Definition. A metric space (X, ) is a set X with a distance function (AKA

metric) such that (1) (x, y) = (y, x), (2) (x, y) 0 with equality only when

x = y, and (3) the triangle inequality (x, z) (x, y) + (y, z) holds.

Examples. The real line with the standard metric d(x, y) = |x y| is a metric

n

space,

pP as is the more general n-space R with the standard metric d(x, y) =

2

(xi yi ) .

Topological spaces extend, abstract, and generalize metric spaces, which are

a more intuitive starting point. If X is a metric space, then X inherits a topology

associated to where open sets are unions of open balls

Br (x) = {y X : d(x, y) < r}

For instance, open balls on the real line and open intervals. It is important

to note that a topology may carry many different metrics, because to be a

topological space is a strictly weaker condition. Some may not carry any metric

at all.

Definition. A topological space is called metrizable iff a topology is induced

by some metric.

So, for instance, the indiscrete topology is non-metrizable if there is more

than one point, because you cannot define a distance function if you cannot

even tell points apart! (1 point is trivial)

2

Definition. The separation axioms (the first three, main ones, at least) are a

way to tell points apart and help with this problem. They are as follows:

1. T0 x y means there exists an open set containing one but not the other.

(A small example is the topology on two points where the open sets are

the empty set, the whole set, the first point, and both points.)

2. T1 x X means {x} is closed (so that points are closed). (A good

example here is the topology on two points where the open sets are the

left point, the right point, and both points, in addition to the trivial open

sets.)

3. T2 (Hausdorff ) x 6= y means x, y are contained in disjoint open sets.

Actually, most spaces we (and most other people) care about are Hausdorff,

and we will often assume this condition. An alternative terminology to T n is

nth countable.

Fact. All metric spaces are Hausdorff, because if P1 , P2 are r apart, then take

open balls of radius less than or equal to r/2, and we have that they do not

intersect and that P1 , P2 are contained in disjoint open sets, i.e., are Hausdorff.

A question was asked about inner product spaces in this arena. If you take

an n-dimensional vector space and you take d(x, y) := ||hx y, x yi||1/2 , then

you get and two products on the same vector space give some topology, because

open balls can be connected between unions.

We often just list open sets, which generate the pothers, e.g., open balls in

(X, d). So, we make a definition.

Definition. A basis for a topology is a set of open sets such that every open

set in the topology is equal to some union of open sets in the basis.

Definition. A neighborhood of a point x is a set containing an open set containing x.

Definition. An open neighborhood requires that the set be open. We will

sometimes just call open neighborhoods neighborhoods. (Compare with the

idea of an -neighborhood in Rn or Cn .)

Definition. A neighborhood basis of x is a set of open neighborhoods of x,

which generate all neighborhoods.

Fact. In (X, d) all points have a countable neighborhood basis, namely B1/n (x),

because A X containing an open set in X iff B1/n A.

Definition. The first countability axiom says that there exists a countable

neighborhood basis for all points. (Such spaces are called first-countable.)

Fact. Hausdorff, first countable, and some technical conditions only a topologist

could love give metrizability.

Definition. The second countability axiom says that there exists a countable

basis for all neighborhoods.

For instance, R and {(r, s) : r < s & r, s Q}. The discrete topology on the

real line is not second countable, but it is metrizable (discrete metric).

3

1.3

Preview

2.1

Main Lecture

We forgot to define some basic notions last time, so we will do that now.

1. If X is a topological space and A X, then A inherits the subspace

topology A TX .

2. If X and Y are topological spaces, then f : X Y is continuous iff the

pullback f 1 (O) of an open set is open in X for all open sets in Y . The

opposite direction is just given by being open; we will see later why the

inverse part is really important.

Calculus defines continuity as if {xn } is convergent in X, then {f (xn )} is

convergent in Y ; keep this in mind.

Definition. A directed set (I, ) is a partially ordered set (poset, which, if

you have forgotten, is binary relation over a set which is reflexive (a a),

antisymmetric (a b and b a means a = b), and transitive (a b and b c

means a c)) with i1 , i2 I = i I with i i1 and i i2 (common upper

bound). i1 , i2 may themselves be incomparable. In particular, we do not require

the order be linear or total.

Definition. A net in a topological space X is a sequence {xi }iI with xi X

where the index set I is directed.

Definition. A net converges to x X if all neighborhoods N of x we have xi

is eventually in N . (You can always go arbitrarily far out in a partial order

to get to N and back.)

Proposition. f : X Y is continuous iff for all nets xi converging to some x

in X, we have f (xi ) f (x) Y .

Proof. We prove right implication and leave left implication as an exercise. If

f is continuous and xi x X, then choose open O Y containing f (x).

f 1 (O) is open by continuity and it contains x because x f 1 (f (x)), so xi

eventually lies in f 1 (O) by definition of convergence. Apply f and so f (xi )

eventually lies in O.

is the

Definition. X a topological space and A X, then the closure of A (A)

smallest closed set containing A.

(closed sets containing A) is closed an minimal, so A is well-defined. We

note that there is a sort of one-to-one correspondence between topology and

closure relations. Indeed, we can define a topology in terms of closure instead

of open sets. (Exercise.)

Theorem 1. X a topological space with A X, then x A iff there is a net

xi A so that xi x.

because each neighborhood of X meets A.

the largest open set disjoint from A and the other way around.

For the other way, we have each open neighborhood U of X meets A. Choose

XU U A, then xU is a net indexed by open neighborhoods of x, which are

convergent to x.

By the way, we always assume choice in this class. Without the axiom of

choice, you are not going to get very far in topology.

Theorem 2. A topological space is Hausdorff iff any net in X converges to at

most one limit.

Before proving this, let us demonstrate that this is special to Hausdorff

spaces; this also shows a big reason why we want our spaces to be Hausdorff.

Let X = {a, b} with non-trivial open sets {a}, {a, b}, so {b} is closed but {a} is

and the net {a} converges to both a and b, so the net has two

not, so b {a}

limit points.

Proof. Hausdorff means distinct points have distinct neighborhoods, so convergence is only possible to one point; this is really easy. The converse is proven

as follows. Assume nets have at most one limit point and X is not Hausdorff.

Choose s 6= t such that s meets each of t. Let Us , Ut be families of neighborhoods of s, t and order Us Ut by (T, U ) (V, W ) T V & U W , so then

Us Ut is clearly directed. Choose x (T, U ) U for all T Us with U Ut ,

then the net converges to t and to s.

Definition. X is compact if for all open covers (families of open sets such that

all points in X are in at least one open set) of X there is a finite sub-cover. (This

is known as the Heine-Borel property after the theorem of the same namesake,

which proved that the unit interval is compact.)

One can phrase this in terms of closed sets and convergence, should he or

she desire. From this, you get the equivalent definition that X is compact if

all families F of closed sets in X with the finite intersection property (FIP),

meaning if f1 , . . . , fn F, then f1 fn 6= ., then

\

F 6=

F

Theorem 3. X is compact iff each next in X has a convergent subnet (throw

away, reindex new order compatible).

For instance, R is not compact; {(n, n + 2) : n Z} is an open cover with

no finite sub-cover.

Proof. We start with right implication. X is compact and {xi } a net. For i T ,

let Ai := {xj : j i} so {Ai } (or its closure) has FIP, since there is no point

further than all which is in the intersection, so there is an x Ai . So, a subset

of {xi } goes to x. (Prove the last assertion.)

To go the other way, we say every net has a convergent subnet. Assume

F is a family of closed sets with FIP. Finite intersections of elements of F are

6

indexed by inclusion. Choose a point in each. This gives a net, and if a subnet

goes to x, then it is easy to see that

\

F

x

F F

Before we go to preview, I will throw in some side notes I made that I was

too lazy to better integrate. First, a sequence can naturally be considered a

net, and you can basically think of nets as sequences, because most interesting

things transfer over to the general case of topology from your experience with

analysis. Second, nets can have different subnet limit points but a convergent

net in a Hausdorff space has just one (see this as it relates to the proof above).

2.2

Preview

Next time, we look at normality (a stronger condition than T2) and interesting

questions like the lifting property.

III

3.1

Last Time

1. A X and x A means there is a net in A which converges to x (and

conversely)

2. X Hausdorff iff a net in X has at most one limit point

3. X compact iff there exist nets in x with convergent subnets

3.2

Main Lecture

Definition. X is normal if disjoint closed sets have disjoint open neighborhoods. (Draw!)

So, we generalized points to closed sets (comparing to T2). Let us get on

with trying t prove that T2 and compact is enough for normality.

Theorem 4. X Hausdorff with A X compact (induced topology) means A

is closed in X and x X A means x and A have disjoint open neighborhoods.

Proof. Fix x X A, then for each y A there is an open neighborhood

Uy of y X so that x

/ Uy because X is T2. The collection of Uy covers

A. By compactness of A, there are finitely many points y1 , . . . , yn A with

,

Uy1 Uyn = U A (finite subcover). x

/ Uy by assumption and x

/U

so A is closed.

Theorem 5. If X is T2 and A, B disjoint compact subsets, then A, B have

disjoint open neighborhoods. (Compact + T2 = normal)

Proof. x A, then Theorem 4 gives that there is Ux such that B is disjoint from

Ux . So, {Ux } covers A. There exist x1 , . . . , xn A with Ux1 Uxn = U A

B = . The complement of U

then gives the result.

and with U

Some books even go so far as to assume Hausdorff in the definition of compactness, and call compact topological spaces, which are not Hausdorff quasicompact, but we will not do this. (This is the Bourbaki convention.)

Definition. X, Y are homemorphic if there exist f : X Y and g : Y X

continuous with f g = 1Y and g f = 1X . (We write X ' Y here, but others

might use

= or .)

Homemorphic spaces are indistinguishable from the perspective of topology,

just like isomorphic groups are in algebra. Clearly, ' gives an equivalence

relation. An even weaker condition called homotopy equivalence, which we will

see later, can be thought of as equivalence, too.

Definition. X is locally compact if every x X has a compact neighborhood.

Examples.

8

2. (0, 1) ' R given by the bicontinuous map (equiv. homeomorphism) x 7

(tanh(x) + 1)/2)

Definition. X, Y topological spaces, then X Y is a topological space with

open sets generated by {U V } where U open in X and V open in Y . This is

called the product topology.

A net (xi , yi ) converges to (x, y) in X Y iff (xi ) x and (yi ) y, as one

would expect. It turns out, this is actually more fundamental then the above

construction, because matters get tricky with infinite products.

Definition. Suppose (X

Qi ) is a collection of topological spaces, then there is

a natural topology on i Xi called the product topology, for which the net

convergence property above holds. A topology is given by sets of the form

Y

Ui Ui open in Xi

i

where all but finitely many Ui (called cylinder sets) are equal to Xi .

Theorem. Tychonoffs theorem states that an arbitrary prduct of compact

(Hausdorff) spaces is compact (Hausdorff).

Proof. Axiom of choice :-)

Theorem. Every locally compact Hausdorff space X has a unique one point

compactification X + with X + = X if X is compact, X + = X {} for X

non-compact and X + is compact Hausdorff.

Proof. Exercise (no time, sorry)

Examples.

If X = Rn , then X + = S n , the n-sphere (using the topological notation so

that a circle is S 1 ). Some other locally compact spaces include the vector space

Rn (not compact), the n-sphere S n := {x Rn+1 : |x| = 1}, the closed n-disk

n

or ball B n or Dn defined by {x Rn : |x| 1}, and the open n-disk DO

(' Rn

given by radial stretching, e.g., x 7 x/(1 |x|)).

Definition. X is connected if it cannot be written as a disjoint union of 2 open

subspaces (so, it is not disconnected).

Clearly, R is connected

Definition. X is path-connected if for all x, y X there exists a continuous

map f : [0, 1] X with f (0) = x and f (1) = y.

Path-connectedness is where we get our intuition from, but it is a stronger

condition that connectedness. Still, if a space is path-connected, it is clearly

connected.

3.3

Preview

Next time we will look at a few loose ends. We should prove the following

proposition: f : X Y continuous and X connected means f (X) connected.

We will also look at Urysohns lemma, which says that if X is normal and A, B

disjoint, normal subspaces, then there is a continuous f : X [0, 1] with f = 0

on A and f = 1 on B. We also investigate the corollary called the Tietze

extension theorem, which says that X normal and A X closed gives than any

For now we are done, and we leave for Labor Day break.

10

- Math 600 (Graduate Algebra, UMD) Notes (Sept. 2016)Uploaded byJoseph Heavner
- Math 630 (Graduate Real Analysis, UMD) Notes (September 2016)Uploaded byJoseph Heavner
- Elliptic Curves - Physics Research Interaction Team, UMD, 2016 - NotesUploaded byJoseph Heavner
- [T. W. Korner] a Companion to Analysis a Second F(BookFi.org)Uploaded bysaradha_ramachandran
- Willard - General Topology (Solutions)Uploaded byNg Ming
- Topology Without TearsUploaded byFunmath
- Cw ComplexesUploaded byBa Nguyen
- 00 Metric SpacesUploaded byEhab Alex
- 9780387244136-c2Uploaded byLejzerCastro
- Hw7Sol_171.pdfUploaded bypratheesh mathew
- Universitext Karen E. Smith Pekka Kekäläinen Lauri Kahanpää William Traves an Invitation to Algebraic Geometry Springer 2000Uploaded byNikolai

- Electromagnetism MIT 8.022Uploaded byRossoVerdi
- Notes - Demystifying the Higgs Boson With Leonard SusskindUploaded byJoseph Heavner
- MAT-208-H Chapter 1.10 Problem Set SolutionsUploaded byJoseph Heavner
- Leontief Input-Output Model Example ProblemUploaded byJoseph Heavner
- Companion to Lang's Algebra (G. Bergman)Uploaded byJoseph Heavner
- Zill Complex Analysis Solutions 3 (Misc, Ch. 2.5-3.2 without 2.7)Uploaded byJoseph Heavner
- Zill Complex Analysis (ed 2) Solutions 4 (Chapter 4.1-4.4)Uploaded byJoseph Heavner
- The Special and General Linear GroupsUploaded byJoseph Heavner
- Isomorphisms in Linear Algebra: A Formal ApproachUploaded byJoseph Heavner
- A Non-Standard Proof of Clairaut's Theorem for the Symmetry of Partial DerivativesUploaded byJoseph Heavner
- Markov Chain Example ProblemUploaded byJoseph Heavner
- Proof of Euler's Formula (problem posed in Zill's Complex Analysis)Uploaded byJoseph Heavner

- nbhmra12Uploaded bysd1802
- G02107476.pdfUploaded byAJER JOURNAL
- Ito,_Kiyoso_-_Encyclopedic_Dictionary_Of_Math_Volume_2_(PDF).pdfUploaded byhascribd
- 04 - Separation axioms.pdfUploaded byBandile Mbele
- The Convex Feasibility Problem in Image RecoveryUploaded byFrabato El Mago
- SOME SEMI-REGULAR WEAKLY CONTINUOUS FUNCTIONS IN TOPOLOGICAL SPACES.Uploaded byIJAR Journal
- pgmq_an17aUploaded byJothi Kumar
- 03 - Continuous FunctionsUploaded byScreaming Silence
- UCB104 real analys.pdfUploaded byLucian Lazar
- TopologyUploaded byStephie Galindo
- On α-ρ-Continuity Where ρ ϵ {L, M, R, S}Uploaded byIOSRjournal
- Cw ComplexesUploaded byBa Nguyen
- M. Spivak - Calculus on manifolds.pdfUploaded byEstudante de Ciências
- ZLSJ2005Uploaded bytplcgn
- Notes on AnalysisUploaded byspaul4u
- Algebraic Geometry Class Notes - GathmannUploaded byx x x
- Solutions 2Uploaded byPaulo Marcos
- cg-2013Uploaded byManuel Zanin
- Introduction to Topology - Ryan Lok-Wing PangUploaded byPang Lok Wing
- 137664639-Hutchinson-Fractals-Self-Similarity.pdfUploaded byDiana Nuică
- Topo Intro102Uploaded byTATHAGATA MACHERLA
- khalafIJCMS1-4-2015Uploaded byKarthik Ramya
- On Some New Notions and Functions in Neutrosophic Topological SpacesUploaded byMia Amalia
- General Topology -Part 1Uploaded byAlayou Yirga
- Topology.PDF.pdfUploaded byKainat Khowaja
- MA2223_ch1.pdfUploaded byRAJESH KUMAR
- Neutrosophic Pre-open Sets and Pre-closed Sets in Neutrosophic TopologyUploaded byMia Amalia
- Fuzzy Soft B-Open Sets in Fuzzy Soft Topological SpaceUploaded byIJSTE
- Generalized b Compactness and Generalized b Connectedness in Topological SpacesUploaded byEditor IJTSRD
- Some Topology Problems and SolutionsUploaded byM2C7r6