Measure Integral and Real Analysis (2)

Measure Integral and Real Analysis (2)

  1. (A×C)/(B×D)=[A×(C/D)][(A/B)×C]
  2. (AB)×(CD)=(A×C)(B×D)
  3. (AB)×(CD)(A×C)(B×D)

Integration

Integration w.r.t a Measure

Definition 3.1: S-Partition

Suppose S is a σ-algebra on a set X. An S-partition of X is finite collection A1,...,Am of disjoint sets in S s.t A1...Am=X.


Definition 3.1: Lower Lebesgue Sum

Suppose (X,S,μ) is a measure space, f:X[0,] is an S-measurable function, and P is an S-partition A1,...,Am of X. The lower Lebesgue sum L(f,P) is defined by:

L(f,P)=j=1mμ(Aj)infAjf

The definition does not assume X is a subset of R.


Definition 3.3: Integral of a Nonnegative Function

Suppose (X,S,μ) is a measure space and f:X[0,] is an S-measurable function. The integral of f w.r.t μ, denoted by fdμ is defined by:

fdμ=sup{L(f,P): P is an S-partition of X}


Theorem 3.4: Integral of a Characteristic Function χE

Suppose (X,S,μ) is a measure space and ES. Then:

χEdμ=μ(E)

Proof of Theorem 3.4:

If P is the S-partition of E,X/E, then:

L(χE,P)=μ(E)1+0μ(X/E)=μ(E)

Thus,

χEdμμ(E)

Let P be an S-partition A1,...,Am of X, then

μ(Aj)infAjχE={μ(Aj),if AjE0, o. w 

Then:

j=1mμ(Aj)infAjχE={j:AjE}μ(Aj)=μ({j:AjE}Aj)μ(E)


Theorem 3.7: Integral of a Simple Function

Suppose (X,S,μ) is a measure space, E1,E2,..,En are disjoint sets in S, and c1,...,cn[0,]. Then:

(k=1nckχEk)dμ=k=1nckμ(Ek)


Theorem 3.8: Integration is Order Preserving

Suppose (X,S,μ) is a measure space and f,g:X[0,] are S-measurable functions s.t f(x)g(x) for all xX. Then:

fdμgdμ

Proof of Theorem 3.8:

Suppose P is a S-partition A1,...,An of X. Then:

infAifinfAigL(f,P)L(g,P)

Then for any P, we have:

supPL(f,P)supPL(g,P)


Theorem 3.9: Integral Via Simple Functions

Suppose (X,S,μ) is a measure space and f:X[0,] is S-measurable. Then:

fdμ=sup{j=1mcjμ(Aj): A1,...,Am are disjoint sets in Sc1,....,cm[0,)f(x)j=1mcjχAj(x) for every xX}

Here, we have A1,...,AmS as arbitrary disjoint sets, they do not have to be S-partition.


Theorem 3.11: Monotone Convergence Theorem

Suppose (X,S,μ) is a measure space and 0f1f2... is an increasing sequence of S-measurable functions. Define f:X[0,] by:

f(x)=limkfk(x)

Then:

limkfkdμ=fdμ

If:

f(x)=limkn=1kfn(x)

Then:

n=1fndμ=fdμ

Proof of Theorem 3.11:

To show that the equality holds, we need to show that:

  1. limkfkdμfdμ
  2. limkfkdμfdμ

We first prove 1:

Suppose (X,S,μ) is a measure space and 0f1f2... is an increasing sequence of S-measurable functions. Define f:X[0,] by:

f(x)=limkfk(x)

Then by theorem 2.53, we have f to be S-measurable. Since fk(x)f(x),xX,kZ+, by theorem 3.8, we have:

fkdμfdμlimkfkdμfdμ

To prove 2, we need to use theorem 3.9:

Suppose A1,...,Am are disjoint sets in S, and c1,...,cm[0,) are s.t:

f(x)j=1mcjχAj(x)

Then:

fdμ=supall disjoint sets in S{j=1mcjμ(Aj)}

Let t(0,1). For kZ+, let:

Ek={xX:fk(x)tj=1mcjχAj(x)}

Then E1E2,... is an increasing sequence of sets whose union is X (because f(x)j=1mcjχAj(x)) and (E1Aj)(E2Aj)....j{1,...,m} is also an increasing sequence of sets, this implies (by theorem 2.59) that:

μ(limk(EkAj))=μ(XAj)=μ(Aj)

Then for all xX:

fk(x)tj=1mcjχAjEk(x)fkdμtj=1mcjχAjEkdμfkdμtj=1mcjμ(AjEk)

Taking the limit to infinity:

limkfkdμlimktj=1mcjμ(AjEk)=tj=1mcjμ(Aj)

Taking the limit t1:

limkfkdμj=1mcjμ(Aj)

Since right-hand side works for any c1,....,cm[0,), disjoint sets A1,...,Am that satisfies f(x)j=1mcjχAj(x), we have:

limkfk(x)supall disjoint sets in S{j=1mcjμ(Aj)}=fdμ


Theorem 3.13: Integral-type Sums for Simple Functions

Suppose (X,S,μ) is a measurable space. Suppose a1,...,am,b1,...,bn[0,] and A1,...,Am,B1,...,BnS s.t j=1majχAj=k=1nbkχBk. Then:

j=1majμ(Aj)=k=1nbkμ(Bk)


Theorem 3.15: Integral of a Linear Combination of Characteristic Functions

Suppose (X,S,μ) is a measure space, E1,...,EnS, and c1,...,cn[0,). Then:

(k=1nckχEk)dμ=k=1nckμ(Ek)

The difference between this theorem and theorem 3.7 is that, here we do not require E1,...,En to be disjoint


Theorem 3.16: Additivity of Integration

Suppose (X.S,μ) is a measure space and f,g:X[0,] are S-measurable functions. Then:

(f+g)dμ=fdμ+gdμ

Proof of Theorem 3.16:

Suppose f,g are simple functions, then:

(f+g)dμ=(jcjχEj+kckχEk)dμ=jcjμ(Ej)+kckμ(Ek)=fdμ+gdμ

So the integrals of simple functions are additive.

Let f1,f2,..., g1,g2,.... be increasing sequence of simple functions s.t:

limkfk=f

limkgk=g

By theorem 2.89, these sequences exist.

Since fk,gk,f,g are S-measurable functions, f+g is S-measurable function. Then by Monotone Convergence Theorem, we have:

(f+g)dμ=limk(fk+gk)dμ=limkfkμ+limkgkdμ=fdμ+gdμ


Definition 3.17: f+,f

Suppose f:X[,] is a function. Define function f+ and f from X to [0,] by:

f+(x)={f+(x),if f(x)00,if f(x)<0

and

f(x)={0,if f(x)0f(x),if f(x)<0

Then:

f=f+f

and

|f|=f++f

This decomposition allows us to extend our definition of integration to functions that take on negative as well as positive values.


Definition 3.18: Integral of a Real-Valued Function: fdμ

Suppose (X,S,μ) is a measure space and f:X[,] is an S-measurable function s.t at least one of f+dμ and fdμ is finite. The integral of f w.r.t μ, denoted fdμ is defined by:

fdμ=f+dμfdμ

The condition |f|dμ< is equivalent to the condition f+dμ< and fdμ<

Example 3.19

Suppose λ is a Lebesgue measure on R and f:RR is a function defined by:

f(x)={1,if x00,if x<0 Then f+dλ= and fdλ=, so fdλ is not defined.


Theorem 3.20: Integration is Homogeneous

Suppose (X,S,μ) is a measure space and f:X[,] is a function such that fdμ is defined. If cR, then:

cfdμ=cfdμ


Theorem 3.21: Additivity of Integration

Suppose (X,S,μ) is a measure space and f,g:XR are S-measurable functions s.t |f|dμ< and |g|dμ<. Then:

(f+g)dμ=fdμ+gdμ


Theorem 3.22: Integration is Order Preserving

Suppose (X,S,μ) is a measure space and f,g:XR are S-measurable functions s.t fdμ and gdμ are defined. Suppose also that f(x)g(x) for all xX. Then fdμgdμ.


Theorem 3.23: Absolute Value of Integral Integral of Absolute Value

Suppose (X,S,μ) is a measure space and f:X[,] is a function s.t fdμ is defined. Then:

|fdμ||f|dμ


Limits of Integrals and Integrals of Limits

Definition 3.24: Integration on a Subset; Efdμ

Suppose (X,S,μ) is a measure space and ES. If f:X[,] is an S-measurable function, then Efdμ is defined by:

Efdμ=χEfdμ

If the right side of the equation above is defined, o.w Efdμ is undefined.

We can also think of Efdμ as fEdμE where μE is the measure botained by restricting μ to the elements of S that are contained in E.

Notice that, Xfdμ means the same as fdμ.


Theorem 3.25: Bounded an Integral

Suppose (X,S,μ) is a measure space, ES and f:X[,] is a function s.t Efdμ is defined. Then:

|Efdμ|μ(E)supE|f|

Proof Theorem 3.25:

Let c=supE|f|:

|Efdμ|=|χEfdμ|χE|f|dμχEcdμ=cμ(E)


Theorem 3.26: Bounded Convergence Theorem

Suppose (X,S,μ) is a measure space and μ(X)<. Suppose f1,f2,... is a sequence of S-measurable functions from XR. If there exists c(0,) s.t:

|fk(x)c|

for all kZ+ and all xX, then:

limkfkdμ=fdμ

Proof of Theorem 3.26:

The function f is S-measurable by theorem 2.48. Suppose c satisfies the assumption above. Let ϵ>0, by Theorem 2.85, there exists ES, s.t μ(S/E)<ϵ4c and f1,f2,... converges uniformly to f on E. Now:

|fkdμfdμ|=|X/E(fkf)dμ+E(fkf)dμ|=|X/EfkdμX/Efdμ+E(fkf)dμ|

Then by theorem 3.23:

|X/EfkdμX/Efdμ+E(fkf)dμ|X/E|fk|dμ+X/E|f|dμ+E|fkf|dμ|

Since |fk(x)|cxX and by theorem 3.25:

X/E|fk|dμ+X/E|f|dμ+E|fkf|dμ|cχX/Edμ+cχX/Edμ+E|fkf|dμ|ϵ2+μ(E)supE|fkf|

Since fk converges uniformly to f on E, for k sufficient large, we have |fk(x)f(x)|<ϵ,xX, this implies that:

limk|fkdμfdμ|=0


Definition 3.27: Almost Every

Suppose (X,S,μ) is a measure space. A set ES is said to contain μalmost every element of X if μ(X/E)=0. If the measure μ is clear from the context, then the phrase almost every can be used.

For example, almost every real number is irrational w.r.t the usual Lebesgue measure because |Q|=0.

Theorems about integrals can almost always be relaxed so that the hypotheses apply only almost everywhere instead of everywhere.


Theorem 3.28: Integral on Small Sets are Small

Suppose (X,S,μ) is a measure space, g:X[0,] is S-measurable, and gdμ<. Then for every ϵ>0, there exists δ>0 s.t:

Bgdμ<ϵ

For every set BS s.t μ(B)<δ.


Theorem 3.29: Integrable Functions Live Mostly on Sets of Finite Measure

Suppose (X,S,μ) is a measure space, g:X[0,] is S-measurable, and gdμ<. Then for every ϵ>0, there exists ES s.t μ(E)< and

X/Egdμ<ϵ


Theorem 3.31: Dominated Convergence Theorem

Suppose (X,S,μ) is a measure space, f:X[,] is S-measurable, and f1,f2,... are S-measurable functions from X to [,] s.t:

limkfk(x)=f(x)

for almost every xX. If there exists an S-measurable function g:X[0,] s.t:

gdμ<

and

|fk(x)|g(x)

for every kZ+ and almost every xX, then:

limkfkdμ=fdμ


Theorem 3.34: Riemann Integrable Continuous Almost Everywhere

Suppose a<b and f:[a,b]R is a bounded function. Then f is Riemann integrable IFF:

|{x[a,b]: f is not continuous at x}|=0

This implies that the outer measure of the set of discontinuities equals to 0 (The set of discontinuities is countable).

Furthermore, if f is Riemann integrable and λ denotes Lebesgue measure on R, then f is Lebesgue measurable and:

abf=[a,b]fdλ


Definition 3.39: abf

We previously defined the notation abf to mean the Riemann integral of f, now we redefine abf to denote the Lebesgue integral.

Suppose a<b and f:(a,b)R is Lebesgue measurable. Then:

  • abf and abf(x)dx mean (a,b)fdλ, where λ is Lebesgue measure on R.
  • abf is defined to be abf.


Definition 3.40: f;L1(μ)

Suppose (X,S,μ) is a measure space. If f:X[,] is S-measurable, then the L1-norm of f is denoted by f1 and is defined by:

f1=|f|dμ

The Lebesgue space L1(μ) is defined by:

L1(μ)={f:f f is an S-measurable function from XR and f1<}

Example 3.41:

Suppose (X,S,μ) is a measure space and E1,...,En are disjoint subsets of X. Suppose a1,...,an are distinct nonzero real numbers. Then:

a1χE1+...+anχEnL1(μ)

IFF EkS and μ(Ek)<.


Theorem 3.43: Properties of the L1norm

Suppose (X,S,μ) is a measure space and f,gL1(μ), then:

  • f10
  • f1=0 IFF f(x)=0 for almost every xX
  • cf1=|c|f1,cR
  • f+g1f1+g1


Theorem 3.44: Approximation by Simple Functions

Suppose μ is a measure and fL1(μ). Then for every ϵ>0, there exists a simple function gL1(μ) s.t:

fg1<ϵ


Definition 3.45: L1(R);f1

  • The notation L1(R) denotes L1(λ), where λ is Lebesgue measure on either the Borel subsets of R or the Lebesgue measurable subsets of R.
  • When working with L1(R), the notation f1 denotes the integral of the absolute value of f w.r.t Lebesgue measure on R.


Definition 3.46: Step Function

A step function is a function g:RR of the form:

g=a1χI1+....+anχIn

Where I1,....,In are intervals of R and a1,...,an are nonzero real numbers. The intervals can be open, closed or half-open intervals.

Compare with simple function, the sets I1,...,In are required to be intervals in step functions.


Theorem 3.47: Approximation by Step Functions

Suppose fL1(R). Then for every ϵ>0, there exists a step function gL1(R) s.t:

||fg1<ϵ


Theorem 3.48: Approximation by Continuous Functions

Suppose fL1(R). Then for every ϵ>0, there exists a continuous function g:RR s.t:

fg1<ϵ

and {xR:g(x)0} is a bounded set.


Differentiation

Hardy-Littlewood Maximal Function

Theorem 4.1: Markov's Inequality

Suppose (X,S,μ) is a measure space and hL1(μ). Then:

μ({xX:|h(x)|c})1ch1

for every c>0.

Proof of Theorem 4.1:

Suppose c>0. Then:

μ({xX:|h(x)|c})={xX:|h(x)|c}1dμ=1c{xX:|h(x)|c}cdμ1c{xX:|h(x)|c}|h|dμ1c{xX:|h(x)|c}|h|dμ+1c{xX:|h(x)|<c}|h|dμ=1ch1


Definition 4.2: 3 times a Bounded Nonempty Open Interval

Suppose I is a bounded nonempty open interval of R. Then 3I denotes the open interval with the same center as I and three times the length of I.

If I=(0,10), 3I=(10,20)


Theorem 4.4: Vitali Covering Lemma

Suppose I1,...,In is a list of bounded nonempty open intervals of R. Then there exists a disjoint sublist Ik1,...,Ikm s.t:

I1...In(3Ik1)....(3Ikn)

Example 4.5:

Suppose n=4 and

I1=(0,10),I2=(9,15),I3=(14,22),I4=(21,31)

Then:

3I1=(10,20),3I2=(3,21),3I3=(6,30),3I4=(11,41)

Thus, I1I2I3I4(3I1)(3I4)

In this example, I1,I4 is the only sublist of I1,...,I4 that produces the conclusion of Vitali Covering Lemma.


Definition 4.6: Hardy-Littlewood Maximal Function; h

Suppose h:RR is a Lebesgue measurable function. Then the Hardy-Littlewood maximal function of h is the function h:R[0,] defined by:

h(b)=supt>012tbtb+t|h|=supt>012t(b+t,bt)|h|dλ

Where h is a Borel measurable function.

In other words, h(b) is the supremum over all bounded intervals centered at b of average of |b| on those intervals.


Theorem 4.8: Hardy-Littlewood Maximal Inequality

Suppose hL1(R). Then:

|{bR:h(b)>c}|3ch1


Derivatives of Integrals

Theorem 4.10: Lebesgue Differentiation Theorem, First Version

Suppose fL1(R). Then:

limt012tbtb+t|ff(b)|=0

for almost every bR

In other words, the average amount by which a function in L1(R) differs from its values is small almost everywhere on small intervals.


Definition 4.16: Derivative; g; Differentiable

Suppose g:IR is a function defined on an open interval I of R and bI. Then derivative of g at b, denoted g(b) is defined by:

g(b)=limt0g(b+t)g(b)t

If the limit above exists, in which case g is called differentiable at b.


Theorem 4.17: Fundamental Theorem of Calculus

Suppose fL1(R). Define g:RR by:

g(x)=xf

Suppose bR and f is continuous at b. Then g is differentiable at b and:

g(b)=f(b)

Proof of Theorem 4.17:

If t0, then:

|g(b+t)g(b)tf(b)|=|b+tfbftf(b)|=|bb+tftf(b)|=|bb+tftbb+tf(b)t|=|bb+tff(b)t|sup{xR:|xb|<|t|}|f(x)f(b)|

Since f is continuous at b, for all ϵ>0, there exists δ>0, s.t |xb|<δ|f(x)f(b)|<ϵ, if we let |t|0, s.t |t|<δ,δ>0, then:

limt0sup{xR:|xb|<|t|}|f(x)f(b)|<ϵ

Thus, g(b)=f(b).


Theorem 4.19: Lebesgue Differentiation Theorem, Second Version

Suppose fL1(R). Define g:RR by:

g(x)=xf

Then g(b)=f(b) for almost every bR


Theorem 4.21: L1(R) Function Equals Its Local Average Almost Everywhere

Suppose fL1(R). Then:

f(b)=limt012tbtb+tf

for almost every bR.

The results holds for at every number b at which f is continuous. Even if f is discontinuous everywhere, the conclusion holds for almost every real number b.


Definition 4.22: Density

Suppose ER. The density of E at a number bR is:

limt0|E(bt,b+t)|2t

If this limit exists (o.w the density of E at b is undefined).

Example 4.23

The density of [0,1] at b ={1,If b(0,1)12,If b=0 or b=10,o.w


Theorem 4.24: Lebesgue Density Theorem

Suppose ER is a Lebesgue measurable set. Then the density of E is 1 at almost every elment of E and is 0 at almost every element of R/E.


Theorem 4.25: Bad Borel Set

There exists a Borel set ER s.t

0<|EI|<|I|

For every nonempty bounded open interval I.


Product Measures

Products of Measures

Definition 5.1: Rectangle

Suppose X,Y are sets. A rectangle in X×Y is a set of the form A×B, where AX,BY.


Definition 5.2: Product of Two σ-algebras; ST; Measurable Rectangle

Suppose (X,S) and (Y,T) are measurable spaces. Then:

  • The product ST is defined to be the smallest σ-algebra on X×Y that contains: {A×B:AS,BT}
  • A measurable rectangle in ST is a set of the form A×B, where AS and BT.

Thus, ST is the smallest σ-algebra that contains all the measurable rectangles in ST.


Definition 5.3: Cross Sections of Sets; Ea and Eb

Suppose X,Y are sets and EX×Y. Then for aX and bY, the cross sections Ea and Eb are defined by:

Ea={yY:(a,y)E},Eb={xX:(x,b)E}

Example 5.5:

Suppose X,Y are sets and AX,BY, then:

[A×B]a={B,if aA,if aA

and

[A×B]b={A,if bB,if bB


Theorem 5.6: Cross Sections of Measurable sets are Measurable

Suppose S is a σ-algebra on X and T is σ-algebra on Y. If EST, then

EaT,aX,EbS,bY

In other words, Let ε denotes the collection of subsets E of X×Y for which the conclusion of this result holds. Then ε is σ-algebra that contains ST.


Definition 5.7: Cross Sections of Functions; fa,fb

Suppose X,Y are sets and f:X×YR is a function. Then for aX and bY, the cross section functions fa:YR and fb:XR are defined by:

fa(y)=f(a,y),yY

and

fb(x)=f(x,b),xX


Theorem 5.9: Cross Sections of Measurable Functions are Measurable

Suppose S is a σ-algebra on X and T is a σ-algebra on X and T is a σ-algebra on Y. Suppose f:X×YR is an ST-measurable function. Then:

fa is a T-measurable function on Y for every aX

and

fb is a S-measurable function on X for every bY

Proof of Theorem 5.9:

Suppose D is a Borel set of R and aX. For all yY, we want to show that:

y(fa)1(D)y[f1(D)]a

Then,

y(fa)1(D)f(a,y)D(a,y)f1(D)y[f1(D)]a

Thus, (fa)1(D)=[f1(D)]a

If f is ST-measurable, then:

f1(D)ST[f1(D)]aT(fa)1(D)T(fa)1 is T-measurable


Monotone Class Theorem

The following standard two-step technique often works to prove that every set in a σ-algebra has a certain property:

  1. Show that Every set in a collection of sets that generates the σ-algebra has the property.
  2. Show that the collection of sets that has the property is a σ-algebra.

Definition 5.10: Algebra

Suppose W is set and A is a set of subsets of W. Then A is called an algebra on W if the following three conditions are satisfied:

  1. A.
  2. If EA, then W/EA.
  3. If E,FA, then EFA.

Thus, an algebra is closed under complementation and under finite unions. A σ-algebra is closed under complementation and countable unions.


Theorem 5.13: The Set of Finite Unions of Measurable Rectangles is an Algebra

Suppose (X,S),(Y,T) are measurable spaces. Then:

  1. The set of finite unions of measurable rectangles in ST is an algebra on X×Y.

  2. Every finite union of measurable rectangles in ST can be written as a finite union of disjoint measurable rectangles in ST.

    (A×B)(C×D)=(A×B)(C×(D/B))((C/A)×(BD))


Definition 5.15: Monotone Class

Suppose W is a set and M is a set of subsets of W. Then M is called a Monotone class on W if the following two conditions are satisfied:

  • If E1E2... is an increasing sequence of sets in M, then: k=1EkM
  • If E1E2... is an decreasing sequence of sets in M, then: k=1EkM

Every σ-algebra is a monotone class. However, some monotone classes are not closed under even finite unions.


Theorem 5.17: Monotone Class Theorem

Suppose A is an algebra on a set W. Then the smallest σ-algebra containing A is the smallest monotone class containing A. (The smallest monotone class is the intersection of all monotone classes on W that contain A.)


Definition 5.18: Finite Measure; σ-Finite Measure

  • A measure μ on a measurable space (X,S) is called finite if μ(X)<.
  • A measure is called σ-finite if the whole space can be written as the countable union of sets with finite measure.
  • More precisely, a measure μ on a measurable space (X,S) is called -finite if there exists a sequence X1,X2,... of sets in S s.t: X=k=1Xk,μ(Xk)<,kZ+


Theorem 5.20: Measure of Cross Section is a Measurable Function

Suppose (X,S,μ) and (Y,T,v) are σ-finite measure spaces. If EST, then:

  1. xv(Ex) is an S-measurable function on X (Think of the function as f(x)=v(Ex)).
  2. yμ(Ey) is a T-measurable function on Y (Think of the function as g(y)=μ(Ey)).


Definition 5.21: Integration Notation

Suppose (X,S,μ) is a measure space and g:X[,] is a function. The notation:

g(x)dμ(x):=gdμ

Where dμ(x) indicates that variables other than x should be treated as constants.


Definition 5.23: Iterated Integrals

Suppose (X,S,μ) and (Y,T,v) are measure spaces and f:X×YR is a function. Then:

XYf(x,y)dv(y)dμ(x)

means:

x(Yf(x,y)dv(y))dμ(x)

In other words, to compute the integral above, first fix xX and compute Yf(x,y)dv(y) if this integral makes sense. Then compute the integral w.r.t μ of the function:

xYf(x,y)dv(y)

If this integral makes sense.


Definition 5.25: Product of Two Measures; μ×v

Suppose (X,S,μ) and (Y,T,v) are σ-finite measure spaces. For EST, define (μ×v)(E) by:

(μ×v)(E)=XYχE(x,y)dv(y)dμ(x)

Example 5.26

Suppose (X,S,μ),(Y,T,v) are σ-finite measure spaces. If AS and BT, then: (μ×v)(A×B)=XYχA×B(x,y)dv(y)dμ(x)=XYχB(y)χA(x)dv(y)dμ(x)=XχA(x)v(B)dμ(x)=μ(A)v(B)

Product measure of a measurable rectangle is the product of the measures of the corresponding sets.


Theorem 5.27: Product of Two Measures is a Measure

Suppose (X,S,μ) and (Y,T,v) are σ-finite measure spaces. Then μ×v is a measure on (X×Y,ST).


Theorem 5.28: Tonelli's Theorem

Suppose (X,S,μ) and (Y,T,v) are σ-finite measure spaces. Suppose f:X×Y[0,] is ST measurable. Then:

xYf(x,y)dv(y) is an S-measurable function on X

yXf(x,y)dμ(x) is an T-measurable function on Y

and

X×Yfd(μ×v)=XYf(x,y)dv(y)dμ(x)=YXf(x,y)dμ(x)dv(y)


Theorem 5.31: Double Sums of Nonnegative Numbers

If xj,k:j,kZ+ is a doubly indexed collection of nonnegative numbers, then:

j=1k=1xj,k=k=1j=1xj,k


Theorem 5.32: Fubini's Theorem

Suppose (X,S,μ) and (Y,T,v) are σ-finite measure spaces. Suppose f:X×Y[,] is ST-measurable and X×Y|f|d(μ×v)<, then:

Y|f(x,y)|dv(y)< for almost every xX

and

Y|f(x,y)|dμ(x)< for almost every yY

Furthermore:

xYf(x,y)dv(y) is an S-measurable function on X

yXf(x,y)dμ(x) is an T-measurable function on Y

And:

X×Yfd(μ×v)=XYf(x,y)dv(y)dμ(x)=YXf(x,y)dμ(x)dv(y)


Definition 5.34: Region Under the Graph; Uf

Suppose X is a set and f:X[0,] is a function. Then the region under the graph of f, denoted Uf is defined by:

Uf={(x,t)X×(0,):0<t<f(x)}


Theorem 5.35: Area Under the Graph of a Function Equals the Integral

Suppose (X,S,μ) is a σ-finite measure space and f:X[0,] is an S-measurable function. Let B denotes the σ-algebra subsets of (0,), and let λ denotes the Lebesgue measure on ((0,),B). Then UfSB and:

(μ×λ)(Uf)=Xfdμ=(0,)μ({xX:t<f(x)})dλ(t)


Lebesgue Integration on Rn

Assume throughout this section, m,n are positive integers.

  1. An open cube B(x,δ) with side length 2δ is defined by: B(x,δ)={yRn:yx<δ}
  2. A subset GRn is called open if for every xG, there exists δ>0 s.t B(x,δ)G.
  3. The union of every collection (finite or infinite) of open subsets of Rn is an open subset of R.
  4. The intersection of every finite collection of open subsets of Rn is an open subset of Rn.
  5. A subset of Rn is called closed if its complement in Rn is open.
  6. A set ARn is called bounded if sup{a:aA}<.
  7. Rm×Rn is identified with Rm+n, note that R2×R1R3 however we can identify ((x1,x2),x3) with (x1,x2,x3), so we say that they are equal.
  8. B(x,δ)×B(y,δ)=B((x,y),δ)Rm+n,xRm,yRn

Theorem 5.36: Product of Open Set is Open

Suppose G1 is an open subset of Rm and G2 is an open subset of Rn. Then G1×G2 is an open subset of Rm+n.


Definition 5.37: Borel set; Bn

  • A Borel set of Rn is an element of the smallest σ-algebra on Rn that containing all open subsets of Rn.
  • The σ-algebra of Borel subsets of Rn is denoted by Bn.


Theorem 5.38: Open Sets are Countable Unions of Open Cubes

  1. A subset of Rn is open in Rn if and only if it is a countable union of open cubes in Rn.
  2. Bn is the smallest σ-algebra on Rn containing all open cubes in Rn.


Theorem 5.39: Product of the Borel Subsets of Rm and the Borel Subsets of Rn

BmBn=Bm+n

In other words, Bm+n is the smallest σ-algebra on Rm+n that containing all measurable rectangles {A×B:ABm,BBn}.

And we can define BmBnBp directly as the smallest σ-algebra on Rm+n+p that containing {A×B×C:ABm,BBn,CBp}


Definition 5.40: Lebesgue Measure; λn

Lebesgue measure on Rn is denoted by λn and is defined indutively by:

λn=λn1×λ1

where λ1 is the Lebesgue measure on (R,B1)

or identically, we can write:

λn=λj×λk

where j,kZ+,j+k=n.


Theorem 5.41: Measure of a Dilation

Suppose t>0. If EBn, then tEBn and λn(tE)=tnλn(E)

where tE={tx:xE}


Definition 5.43: Open Unit Ball in Rn; Bn

The open unit ball in Rn is denoted by Bn and is defined by:

Bn={(x1,...,xn)Rn:x12+....+xn2<1}


Definition 5.46: Partial Derivatives; D1f and D2f

Suppose G is an open subset of R2 and f:GR is a function. For (x,y)G, the partial derivatives (D1f)(x,y) and D2f(x,y) are defined by:

(D1f)(x,y)=limt0f(x+t,y)f(x,y)t

(D2f)(x,y)=limt0f(x,y+t)f(x,y)t

If these limit exists.


Theorem 5.48: Equality of Mixed Partial Derivatives

Suppose G is an open subset of R2 and f:GR is a funciton s.t D1f, D2f, D1(D2f), D2(D1f) all exists and are continuous functions on G. Then:

D1(D2f)=D2(D1f)