TOPICS COVERED ON MPT: assuming things are evenly distributed -> 50/6 ~ 8

ques/topic
● calculus sequence (okay I feel good)
● linear algebra (not as much of an issue)
● probability and statistics
● real analysis
● modern/abstract algebra
● number theory

NUMBER THEORY
● binomial them (a+b)^p = sum from k = 0 to k = p of (p choose k) * a^(p-k)*b^k
● (p choose k ) = p!/(k! * (p-k)!)
● chinese remainder thm: if x= a mod p, x = b mod q, (p,q) = 1
○ then let q1 = q' mod p, p1 = p' mod q, solution is y = aqq1 + bpp1

ABSTRACT ALGEBRA
letting xor = +in a circle, and xand = x in a circle
1. Z is a curcular group!!!!!!
2. Cauchy's thm: let G finite group of order n, let p prime p|n. then G has at least one
subgroup of order p
3. generalization of cauchy's: sylow's first thm: let G be a finite group of order n, and let n =
(p^k)m, where p prime, p not divide m. then G has at least one subgroup of oder p^i for
every int i from 0 to k
4. every abelian group of order n has exactly 1 subgroup of order d, for every d|n
5. every group of order three is isomorphic to Z3
6. let G1 G2 groups with *1 and *2. g1xg2 = {(a,b): a in g1, b in g2}. define * by (a1,
a2)*(b1,b2) = (a1*1b1,a2*2b2). this group is called direct product of g1 g2. if g1 order n,
g2 order m, g1xg2 order nm. if g1 abelian and g2 abelian, g1xg2 is abelian, then
sometimes write g1\xor g2
7. the direct sum of zn zm is cyclic iff gcd(m,n) = 1. then zm xor zn has order mn and is
isomorphic to zmn
8. the direct sum of zm1 xor zm2 … xor zmk iff gcd(mi, mj) = 1 for every distinct pair mi
mj. then this group is isomorphic to zm1m2...mk
9. **overall result** ever finite abelian group is isomorphic to zp1^k1 xor zp2^k2 … xor
zpn^kn; where pi primes are not necessarily distinct, ki pos ints not necessarily distinct.
the collection of prime powers pi^ki are known as the elementary divisors of g
10. also ** every finite abelian grop g is isomorphic to a direct sum of the form zm1 xor zm2
… xor zmk where m1 >= 2; m1 | m2, m2|m3 … mk-1|mk. the integers m1 .. mk not
necessarily distinct, but the list m1 … mk is unique. called invariant factors of g
11. for 1<=k<n, the order of k in Zn = n/(gcd(k,n)) = (lcm(k,n)/k)
12. let (g, .) and (g', *) groups. a function phi: g -> g' with phi(a.b) = phi(a)*phi(b) for all a,b
called homomorphism
13. monomorphism: 1-1 homomorphism
14. epimorphism: onto homomorphism
15. isomorphism: bijective homomorphism
16. endomorphism: homomorphism from group to self
17. automorphism: isomorphism from group to self
18. properties of homomorphisms: phi: G -> G'
1. if e is identity in g, then phi(e) is the identity in g'
2. if g in G has finite order m then phi(g) in G' also has order m
3. if a^-1 is invers of a in g then phi(a^-1) is invers of phi(a) in g'
4. if h is subgroup of g then phi(h) is a subgroup of g' where phi(H) = {phi(h): h in
H}
5. if G is finite then the orger of phi(G) divides the order of G, if G' is finite the oder
of phi(G) also divides the oerder of G'
6. if h' is a subgroup of g' then phi-1(H') is a subgroup of G where phi-1(H') = {h in :
phi(h) in H'}
7. let e' be identity in G'. then {e'} is a subgroup (the trivial subgroup) of G'. then the
inverse image of {e'}is a subgroup of G. called kernel of phi or ker phi. ker phi
={g in G: phi(g) = e'}
8. a homomorphism is a monomorphism ⇔ its kernel is trivial
19. ker phi is always a subgroup in G
20. trivial homomorphism: maps everything to e'
21. normal subgroup: N. xnx^-1 in N for all n in N, all x in G
22. an infinite group can by isomorphic to one of its proper subgroups
23. rings
24. its got two operations. + and *
25. (r,+) is an abelian group, (r,*) is a semigroup (so multiplication is closed and associative,
but doesnt have to have an identity or inverse); distributive law must hold
26. if (r,*) also has an identity its called a ring with unity
27. the identity of (r,+) is usually denoted by 0; the identity of (r,*) is usually denoted by 1;
usually required that 0 != 1
28. if multiplication commutative then its called a commutative ring
29. if s is a subset of r and it satisfies the properties to be a ring with same ops then its called
a subring
30. smallest pos int n st na = 0 for all a in R (if it exists) is called characteristic of R and
denoted by char R = a
31. if no such n exists, like in Z Q R and C, we say the ring has characteristic of 0
32. if char R > 0, then you only have to find the smallest n for which n * 1 = 0
33. examples
33.1. (Z, + , *) this is a ring with unity called ring of ints
33.2. (nZ, +, *) is also a ring but not a ring with unity for n>1
33.3. the ring of ints is a subring of (Q,+,*), (R,+,*) and (C,+,*)
33.4. (Zn, +, *) is the ring of ints mod n
33.5. (Mn(R), +, x) is a ring with unity, noncommutative with n>1; it has subrigns
(Mn(Q), +, x) (Mn(Z), +, x)
33.6. the set R = {a+bsqrt(2): a,b in Z} with normal + and * is a commutative ring with
unity
33.7. the set R = {a+bcubert(2)} is not, since its not closed
33.8. with regular + and * the set Z[i] = {a+bi: a,b in Z, i = sqrt(-1)} is a subring of C,
called ring of gaussian ints
33.9. consider the ring R, the set of polynomials with coefficients in R is also a ring;
called the ring of polynomials in x over R
33.10. the set of all functions f: R -> R is denoted R^R with operations of addition
[(f+g)(x) = f(x)+g(x)] and pointwise multiplication [(fg)(x) = f(x)g(x)] then R^R
is a commutative ring with unity called ring of real valued functions on r. the
additive identity is 0 and the multiplicative identity is 1
34. if r is a ring with unity and s is a subring of r with unity. the unity of r does not have to
equal the unity of s
35. nilpotent - if exists m st a^m = 0 in a ring
35.1. in a commutative ring if a, b nilpotent then a+b nilpotent
35.2. if an element c is invertible in a ring (that is exists c' st c*c' = c'*c = 1) then it
cant be nilpotent
36. ring homomorphisms- (f shoudl be phi) let (r,+,x) and (r', xor, xand) be rings. a fuction f:
r-> r' is a homomorphism if for all a,b in r: f(a+b) = f(a) xor f(b), f(axb) = f(a)xandf(b)
37. properties of ring homomorphisms
37.1. ker f = {a in r: f(a) = 0'} where 0' is additive identity of r'
37.2. ker f is always a subring of r
37.3. the image of r, f(R) = {f(r) for r in R} is a subring of r'
37.4. the image of 0 (additive identity of r) must be 0' (add id of r'). then f(-r) = -f(r) for
all r in R, where -r is additive inverse of r.
37.5. examples
37.5.1. Z -> Zn given by f(x) = x mod n is a homomorphism (epi since onto)
37.5.2. f: R[x] -> R , fa(p(x)) = p(a) is a homomorphism
37.5.3. frobeniums endomorphism; if r commutative ring with unity with char R =
p, then f:R-> R , f (a) = a^p is a ring homomorphism; so (a+b)^p = a^p +
b^p
 37.5.4. ex, let r = r' = Z and f(m) = 2m then addition preserved but not mult
37.5.5. same as previous but for multiplication
37.5.5.1. ex: let r = GL(2,R) and r' = R then f = det(A) is a group
homomorphism bc preserve mult but not addition
37.5.6. all endomorphisms from z to z are: 0 endomorphism and identity one
38. an ideal I in R if xr in I and rx in I for all x in I and all r in R, analogous to normal
subgroup for groups
39. integral domain is a commutative ring with unity where ab = 0 ⇔ a = 0 or b = 0
40. a nonzero element m in Zn is a zero divisor ⇔ (m,n) != 1
41. if n is prime, Zn is an integral domain
42. R^R is not an integral domain: ex f = x - abs(x), g = x + abs(x)
43. unity, the unique multiplicative identity in a ring
44. for a != 0 in a ring R with unity, a is called a unit if a is invertivle, exsits a^-1
45. if every nonzero element in r is a unit, that is if (R*,*) is a group, then R is called a
division ring
46. field: commutative division ring
46.1. examples
46.2. Q R C are fields. Z is not a field (since its only units are 1 and -1)
46.3. the commutitive ring with unity Qsqrt(2) = {a + bsqrt(2): a,b in Q} is a field
46.4. a finite integral domain is a field
46.5. if p is prime then Zp is a field, then every nonzero element is a unit
46.6. K = {Matrix((a,b),(-b,a)): a,b in R} is a field. it is isomorphic to C
46.7. H = {Matrix((a,b),(-b',a')): a,b in C, a' and b' denote complex conjugate}. H is a
division ring, but mult not commutative so not a field. called division ring of real
quaternions
46.8. strictly skew field: noncommutative division ring\
46.9. boolean ring every a in R is idempotent, a^2 = a
47. for phi group homomorphism, if the domain is cyclic, any homomorphism is completely
determined by phi(any generator)
SET THEORTY
● if a set is equivalent to Z+ then it is countably infinite and its cardinality is alef0
● ex of countably infinite sets
○ Z
○ Q
○ Algebraic numbers
○ any infinite subset of the integers, rationals, or algebraic numbers
○ a finite cartesian product of countably infinite sets
○ a finite union of countably infinite sets
○ a countably infinite uniion of countably infinite sets
● if there is an injective function from a to be then card a <= card b
● if there is a surjective function from a to b then card a >= card b
● power sets of A, P(A), set of all subsets of A
● for finite A of card n, P(A), card P(A) has 2^n elements
○ so there is always surjective func from p(A) to a but never injective one
○ card P(A) > A
● uncountable sets
○ R, card R = c = 2^alef0
○ Q^c = R - Q, set of irrationals
○ set of transcendentals
○ any subset of R with a nonempty interval
COMBINATORICS
properties of combinatorics
C(n+1,k) = …
sum( C^2 (n,k)) = (C(2n,k))
sterling numbers, sterlings formula
● sterling partition number: number of ways to partition n objects into k nonempty subsets
{n over k} = S(n,k)
● S(n,k) = \frac{1}{k!} \sum_{i=0}^k (-1)^i \begin{pmatrix} k \\i \end{pmatrix} (k-i)^n
● S(n+1,k) = kS(n,k) + S(n,k-1)
● S(0,0) = 1, S(n,0) = S(0,n) = 0
ways to put balls in boxes with none empty
ways to put things in boxes (all empty, not all empty, min of one)
● if a choices followed by b choices, total options is ab
● P(n,k) no repetitions = n!/(n-k)!
● C(n,k) no repetitions = P(n,k)/k! = n!/[k! (n-k)!]
● P(n,k) with reps = n^k
● C(n,k) with reps = C(n+k-1,k)
● pigeon hole priciple: if n objects are placed into k pigeonholes, and n>k then at least one hole
has more than 1 object
● generalized pigeonhoel principle: n objects, each of which is one of c colors. then for k <=
[(n-1)/c]+1 there must be k objects all of which are the same color
● partition function, memorize first few values
○ total partitions p(n)
○ partitions of n with k parts p(n,k)
○ https://math.berkeley.edu/~mhaiman/math172-spring10/partitions.pdf
PROBABLILITY AND STATISTICS
● P(A) = desired outcomes/possible outcomes
● (boolean) algebra of set S - E subset of P(S) st
○ if A,B sets in E, then AUB and AintersectB are in E (closed under union, intersection)
○ if A set in E then so is A' = S-A
● E always contains null and S
● a function P: E -> [0,1] is called a probability measure on E if:
○ P(null) = 0
○ P(S) = 1
○ P(AUB) = P(A) + P(B) for all disjoint sets A,B in E
● probability space : is a set S, together with an algebra on S, and a probability measure on S
○ S is the sample space
○ elements of S are outcomes
○ and sets in E are called events
○ then P(A) is probability that event A occurs
● P(A U B) = P(A) + P(B) - P(A and B)
● A and B independent if P(A and B) = P(A)P(B)
● if a, b mutually exclusibe then P(A U B) = p(A) + P(B)
● bernouli trials: either success or failure -> binomial distribution P(exactly k successes) =
C(n,k)p^k*q^(n-k)where P(success) = p , P(failure) = 1- p = q ;
○ most likely number of successes for binomial dist is floor(p*(n+1)); if this is an int, then
there are two most likely: p(n+1) and p(n+1)-1
● F(t) = P(X<=t) is the distribution function
○ F(t1 < x <= t2) = F(t2) - F(t1)
○ F(t1 <= x <= t2) = F(t2) - F(t1) + P(x = t1)
● F(x) = integral of f(x) ; probability density function
○ integral of f from -inf to inf = 1
○ must be nonnegative and inegrable
● P(X <= t) = F(t) = integral(-inf, t) of f
● Expectation/mean of X = E(x) = mu(x) = Sum over all ks of k*p(k) = Integral(-inf, inf) of
[t*f(t)]dt
○ E(x) for any binomial variable is n*p
● Var(x) = sum over all ks of [k - mu]^2 * p(k) = integral from -inf to inf of [t-mu]^2 - f(t)
● standard dev = sqrt(variance)
● normal distribution f(t) = [1/(stdev*sqrt(2pi))]*exp[-(t-mu)^2/(2*stdev^2)]
○ to use a table, then z = (t-mu)/stdev, stdev= 1
○ for this variable we have P(t1 < x <= t2) = P((t1 - mu)/s < z <= (t2 - mu)/s)
○ p(z1 < Z <= z2) = PHI(z2) - PHI(z1)
● normal approximation of binomial: for binomial mu = np, s = sqrt(npq), P(a1 <= x <= a2) =
PHI((a2-mu+½)/s) - PHI((a1-mu-½)/s)
● PHI(-z) = 1 - PHI(z)
● PHI(.0) = .5, PHI(.5) = .69, PHI( 1) = .84, PHI(1.5) = .93, PHI(2) = .97, PHI(2.5) = .99,
PHI(3+) ~~ 1
● different distributions, all the most important and when to use them: gamma, lambda,
exponential, poisson
● hypergeometric, poisson, geometric ,
● bayes thm, conditional probability
REAL ANALYSIS
● continuous
● uniformly continuous
● pointwise continuity?/limit
● types of sums: rieman, lebesgue?
● if u is the least upper bound of a set X, then u = lub X, aka supremum or sup X
● if l is greatest lower bound of a set X, then l = glb X, aka infimum or inf X
● least upper bound axiom: if x is a set of reals then E exactly one real number u that is
least upper bound of X
● if x is a set of real numbers thats bounded below, there is exactly one greatest lower
bound l
● cauchy sequence: a sequence of real numbers (xn)_n=1 ^infinity is cauchy if for every
epsilon greater than 0 there exists an integer N such that for every pair of integers m and
n (both greater than or equal to N) we have |xm - xn| <epsilon
○ that is, the terms become aribtrarily close to one another at some point
○ thm: every cauchy sequence of real numbers converges
● any metric space in which every cauchy sequence converges to a point in the space is
called a complete space -> R is complete
● lebesgue (outer) measure: let Rbar denote R +{infinity}. let a subset r, and find a
countable open covering of A by intervals of the form (ai,bi) then Asubset Union_i=1
^infinity (ai,bi). define mu: P(R) -> Rbar by
○ mu*(A) = inf{sum_i=1 ^infinity (bi-ai), for Asubset union_i=1 ^infinity (ai,bi)}
● lebesgue measurable sets M = {M: M subset R, mu*(A) = mu*(A intersect M)+ mu*(Ac
intersect M) for every A in P(R)}
●
○ for mu:M->[0,inity) we have
■ 1) extends length, for any interval, mu(I)= length I
■ 2) monotone, if Asubset b subset R, then 0<=mu(A)<=mu(B) <=infinity
■ 3) translation invariant, mu(A+x0)= mu(A)
■ 4)countably additive, mu(countable union of Ais) = sum(countable #of
mus)
○ every open set in r is measureable
○ every closed set in r is measurable
○ every finite or countably infinite subset of r is measurable
○ the complement of a measurable set is measurable
○ a finite or countably infinite union or intersection of measurable sets is
measurable
● mu(null) = mu({any singleton})= 0
○ therefore, if M is a finite or countably infinite subset of R, then mu(M) = 0
 ○ so mu(Z)=mu(Q) = 0
● if M is an interval, mu(M) = length of interval
● if M is a finite union of dijoint finite intervals then mu(M) = sum of lengths of intervals
● if m is a meausrable set that contains an infinite interval then mu(m) = infinity
● if m is any countable collectino of disjoint sets then mu(union_i=1 ^infinity Mi) =
sum_i=1 ^infinity mu(mi)
● if m1 and m2 in M and m1 subset m2 then mu(m1)<=mu(m2)
● a function f:r->r is (lebesgue) measurable if for every open set in R the inverse image is a
lebesgue measurable set
○ since every open subset of r is measurable, every continuous function is
measurable
● sum dif and prod of measurable functions is also measurable
● if f is measurable then so is |f| = |f(x)|
● for A subset of R, characteristic function chia = 0, x not in A; 1, x in A
○ this function is measurable if and only if a is a measurable set
● if every set Ai is measurable then the step function s = sum_i=1 ^n ai*chiAi is also
measurable
○ this is the most important ex of a measurable noncontinuous function
○ the range of a step funciton is a finite subset of R
● thm: let f be measurable with f>= 0, then E sequence of step functions (s1,s2,...) st
0<=s1<=s2<=... and where limn->infinity sn = f (for all x in R lim of sequence sn(x) =
f(x))
● in general, if function fn is measurable for every positive int n, and fn-> f almost
everywhere (set of x's where fn-\> f has measure 0) then the limit function f is also
measurable
● let s = sum_i=1 ^n ai*chiAi; then s is measurable and s is lebesgue integrable if ai != 0
then mu(Ai) < infinity, that is if mu(Ai) = infinity then ai = 0
● in this case integral of s dmu = sum_i=1 ^n ai*mu(Ai)
● if f>=0, f measurable, with every step fucntion s>=0, s<=f integrable, and if integral s d
mu is finite, then f is lebesgue integrable and integral of f dmu = sup{integral s dmu}
○ taken over all such s
● an arbitrary function is integrable if f+ and f- are inegrable, then integral of f dmu =
integral of positive part - integral of negative part
●
COMPLEX ANALYSIS (let z_ = the complex conjugate of z)
● |z|^2 = z*z_
● theta = arg z = angle with pos real axis; -pi < Arg z <= pi
● z = r(costheta + i sin theta)
● z1z2 = r1r2[cos(theta1+theta2) + i sin(theta1+theta2)] (multiply moduli, add arguments)
● z1/z2 -> divide moduli, subtract arguments
● z^n = r^n[cos(n*theta)+isin(n*theta)]
● z = r*e^(i theta)
● nth roots of unity, solve z^n = 1; z = e^i(2pik/n) for k from 0.. n-1
● nth roots of any w=Re^iTheta -> r = nroot(R), theta = (Theta + 2pik)/n
○ principal nth root of w is with Theta= Arg(w) and k=0 -> nroot(w)=
nroot(R)e^i(Theta/n)
● Log z = log|z| + i Argz
● z^w = e^(w*Logz)
● cosz = (e^iz + e^-iz)/2
● sinz = (e^iz - e^-iz)/2i
● complex sin and cosine are NOT bounded
● cosh x = (e^x + e^-x)/2
● sinh x = (e^x - e^-x)/2
○ cosh(0) = 1; sinh(0) = 0
○ cosh(-x) = cosh(x)
○ sinh(-x) = -sinh(x)
○ cosh^2 (x) - sinh^2(x) = 1
○ cosh (x) .>= 1
○ sinh (x) can be any real number
● cos(iz) = cosh(z)
● cosh(iz) = cos(z)
● sin(iz) = i sinh(z)
● sinh(iz) = i sin(z)
● cos(z) = cos(x+yi) = cosxcoshy - i(sinxsinhy)
● sin(z) = sin(x+yi) = sinxcoshy + i(cosxsinhy)
● |cosz|^2 = cos^2(x) + sinh^2 (y)
● |sinz|^2 = sin^2 x + sinh^2 y
● f(z) = f(x+yi) = u(x,y) + iv(x,y)
● differentiabke if lim h->0 (f(z0+h) -f(z0))/h
● lim z-> 0 of z_/z dne
● cauchy riemann equations: if satisfied, f = u + iv is differentiable
○ du/dx = dv/dy
○ dv/dx = -du/dy
● if f differentiable in some open set O then -> uxx + uyy = 0; vxx + vyy = 0; they are
harmonic in O
● if f is differentiable in some open set around a point then it is analytic at tgat point
● if f analytic everywhere then its called entire
○ all polynomials, exp, sin, cos, rationals anywhere denominator doesnt eqyal 0
● if f is analytic somewhere, then all of its derivatives are analytic at that point
● line integrals: integral_C f(z) dz = integral _a ^b f(z(t))*z'(t)dt; if f has an antiderivative
you can just do F(z(b)) - F(z(a))
● thms about analytic functions
○ cauchy's thm: if f(z) is analytic throughout a simply connected, open set, D, then
for every closed path C in D then integral along closed path C is 0
■ in particular, if f is entire than any closed path in plane has 0 for closed
path integral
○ morera's thm: if f(z) is continuous throughout an open connected set O in C, and
integral along closed curve = 0 for every closed curve c in o then f analytic in O
(converse of cauchy's thm ish)
○ cauchy's integral formulas: if f is analytic at all points within and on a simple
closed path C that surrounds z0 then f(z0) = 1/(2pii) path integral [f(z)/(z-z0)]dz
■ its integrals are also analytic and f^(n) (z0) = (n!/2pii)path integral over c
of [f(z)/(z-z0)^(n+1)]dz
○ cauchy's derivative estimates: let f ber analytic on and within |z-z0| = r, the circle
of radius r centered at z0 -> let M be max value of |f(z)| on circle, then |f^(n) (z0)|
<= (n!M)/r^n
○ liouville's thm: if f is an entire funciton that's bounded then f(z) must be a constant
function
○ the maximum principle/ maximum modulus thm: let O be an oppen connected
subset of the complex plane, and let f be analytic in O. if E z0 in O st
|f(z)|<=|f(z0)| for all z in O, then f is a constant function.
■ that is: if f is analytic and not constant in an open connected subset of the
pane the |f(z)| attains no maximum value in O.
■ if O is bounded then |f| must achieve a maximum value in cl(O). since |f|
cant attain its max in O, it has to be on the boundary of O
○ picards little thm: every entire function whose image omits two or more points is
constant
○ radius of convergence for a complex power series: R = 1/(lim n-> infinity of
nroot|a_n|) if that denominator is nonzero and finite
■ if denominator is infinity then power series converges only for z0
■ and if denom is 0 series converges for every z
 ○ every function thats analytic in an open connected subset O of C can be expanded
in a poiwer series with a disk of convergence in O
○ pwer series is f(z) = sum_n=0 ^inf a_n (z-z0)^n, where a_n= f^(n) (z0)/n!
● if f is not analytic at z0 but is analytic at some point in every punctured disk centered at
z0, then z0 is a singularity
○ kf z0 is a singularity, but f is analytic at every point in some punctured open disk
centered at z0 the nz0 is an isolated singularity
● let z0 be a singularity; if f(z) = g(z)/(z-z0)^n, g analytic in an open disk centered at z0,
g(z0)!= 0 then z0 is a pole of order n
○ simple pole pole of order 1
○ order 2 double pole, etc
○ if f cant be written in this form for any positve int n, then z0 is an essential
singularity
● laurent series: sum_n=1 ^inf a_-n (a-z0)^-n + sum_n=0 ^inf a_n (z-z0)^n, where a_n=
f^(n) (z0)/n!
○ first part is called singular/principal part
○ coefficients a_-n = 1/2pii closed path integral over c of f(z)/(z-z0)^(-n+1) dz,
n>=0
○ if singular part has a_-k (k terms) then z0 is a pole of order k
○ the number a_-1 in the laurent series of f is called the residue of f at z0
■ res(z0,f) = a_-1
○ closed path integral over c of f(z) dz = 2pii * sum_m=1 ^n Res(zm,f), for any
closed c in the annulus where laurent series is valid
○ to find the residue of a pole of order k: Res(z0,f) = 1/(k-1)! * lim z->z0 of
d^k-1/dz^k-1[(z-z0)^k * f(z)]
■ if k =1 then Res(z0,f) = lim z-> z0 [(z-z0)f(z)]
■ if k = 2 then Res(z0,f) = lim z-> z0 d/dz[(z-z0)^2 * f(z)]
●
 TOPOLOGY
● a topology on x is a family of subsets with -> closed under finite intersection and arbitrary
union -> ring with + = intersection, * = union
○ null and x in T
○ if a and b are in t, so is their intersection
○ if ai, i in I is any collection of sets from T then their union is also in t
● any set in T is called an open set
● for if x in X, and N in T, then if there exists O with x in O subset of N, N is called a
neighborhood of x
● if O is open, then its complement Oc is closed
● topological space x with a topology on x, (X,T)
● hausdorff: for every pair of distinct points x,y in X, there are disjoint open sets Ox and Oy
such that x in Ox and y in Oy
● the collection {null, X} is called the indiscrete topology
● the powerset of x P(X) is called the discrete topology, all subsets of x are open; all possible
sets are open
● if we have two topologies t1 and t2, then t1 is finer than t2, or t2 is coarser than t1 if t2 subset
of t1
○ discrete topology is finest, and trivial topology is coarsest
● let (X,T) be a topological space, and S subset X. Then let U subset S be open if it = S
intersect O where O is open in X. this is Ts, the subspace/relative topology
○ if S open in X, and U open in S, then U open in X
○ if S not open in X, then U open in S not necessarily open in X
● Let (X,T) be a topology, and A subset of X. then
○ int(A) = Union of all open sets contained in A
○ ext(A) = union of all open sets that dont intersect A
○ ext(A) = int(Ac)
○ boundary of A bd(A) = {x in X st every open set that contains x intersects A and Ac}
○ x is a limit/accumulation/cluster point of A if every open set containing x also contains at
least one point in A other than x
○ the set of all limit points of A is called derived set of A and is denoted A'
○ closure of A, cl(A) = int(A) union bd(A) = A union A'
● a set is closed iff it contains all of its boundary points and limit points
● basis for a topology - let x be nonempty, let b be a collection of subsets of x that satisfy
○ for every x in X, there is at least one set b in B such that x in b
○ if b1 and b2 are sets in b, and x in b1 intersect b2, then there exists a set b3 such that x in
b3 subset of b1 intersect b2
● the sets of b are called basis elements
● topology generated by b: a subset O of X is open if for all x in O, E a basis element B st x in
B subset of O. Equivalently: T = all possible unions of basis elements
● lower limit topology: sets generated by {[a,b): a,b in R and a<b}
● product topology: if (X,Tx), (Y,Ty) topological spaces, then the topology generated by B =
{Ox x Oy: Ox in Tx and Oy in Ty} is called product topology on XxY
● let (x,t) topological space. if E o1, o1 disjoint open sets in T, and o1 union o2 = x, then x
disconnected
● otherwise, if no pair of disjoint sets with union x exist, then x is connected space
● if s subspace of x, and if E 2 open subsets of x, A and B, with s subset of AUB and A intect
S, B intersect S disjoint and nonempty, then s is disjoint, otherwise connected
● properties of connected spaces
○ if a and b connected and they intersect, then their union is also connected; also holds for
any number of connected sets if their intersection is non empty
○ let a be a connected set and let b st a subset b subset cl(a) then b is connected
○ the cartesian product of connected sets is connected
● let x1, x2 in X. if can be joined by a continuous path, that is if E p:[0,1] -> X with p(0) = x1,
p(1) = x2, p continuous then x is path connected
○ path connected spaces are always connected
○ implies that any convex set in Rn is connected
● let (x,t) be a topological space. a covering of x is a collection of subsets of x whose union is x
○ an open covering is a covering that consists only of open sets
○ if every open covering of x contains a finite subcollection that also covers x, then x is
said to be a compact space
■ R is not compact bc the open covering A = {An:n in z} where An = (n-1, n+1)
contains no finite subcollection that can cover all of R
■ Any space X with trivial topology is compact bc only open sets are null and X and the
finite collection {X} covers x
■ open intervals in R are not compact take An, n>=3, An = (a+1/n, b-1/n)
■ closed intervals in R are compact
● if X compact space and S closed subset of X. then s is closed
○ if hausdorff & s compact, then S closed
● cartesian product of compact sets is compact
● define ||x|| = sqrt(x1^2 +...+xn^2) = norm of x
● a subset A of Rn is bounded if E M>0 st norm of every point in A is less than M (that is, A is
contained in a sphere of radius A centered at the origin)
● Heine-borel thm: a subset of Rn with standard topology is compact if and only if closed and
bounded
● let X be nonempty and d:X x X -> R be a function defined on ordered pairs of X. this is a
metric on x if these properties hold for all x,y ,z
 ○ d(x,y)>= 0 and d(x,y) = 0 if x = y
○ d(x,y) = d(y,x)
○ d(x,z)<= d(x,y) + d(y,z)
● a set x together with a metric is called a metric space
● if epsilon positive number then the sed Bd(x,e) = {x' in X: d(x,x')<e} is called an epsilon ball,
ball of radius e centered at x
● collection of all epsiolon balls is a basis for a topology on x, metric topology (induced by d)
-> O is open if for every x in O, there is some positive ex st Bd(x,ex) subset O
● square metric = max(abs(x1 -x1'),...,abs(xn-xn')) -> gives standard topology on R
● a function is continuous from (x1,t1) -> (x2,t2) if for every open set in x2 the preimage is
open in x1
○ f is contiunuous at x0 if every open set O in x2 containing f(x0) , theres an open set o1 in
x1 with f(o1) subset O
○ if the range space is generated by a basis b, then to check that f is continuous, its enough
to check that the enverse image of every basis element in b is open in x1
○ with the discrete topology on the domain space, every function is continuous
○ if the range space has trivial topology, then every function is continuous
○ for a continuous map f:(x1,t1)->(x2,t2)
■ ⇔ the set finverse(C) is closed in x1 for every closed subset c of x2 (f continuous iff
this)
■ if c is a connected subset of x1 then f(c) is a connected subset of x2
■ if c is compact subset of x1 then f(c) is copact subset of x2
■ for f:x->r continuous, x compact then f must atain an absolute minimum and
maximum (analog of closed and bounded)
○ open map: image of every open set is open
○ if f:x1->x2 is a bijection and f1,finverse continuous (that is f is a continuous open map)
then f is called a homeomorphism
○ homeomorphisms preserve topological properties: cardinality, connected ness,
compactness, etc
○ let f:x1->x2 be a bijective, continuous map. If x1 is compact and x2 is hausdorff, then f is
a homeomorphism