Tag Archives: analysis

A semi-continuous function with a dense set of points of discontinuity

Let’s come back to Thomae’s function which is defined as:
f:|RRx0 if xRQpq1q if pq in lowest terms and q>0

We proved here that f right-sided and left-sided limits vanish at all points. Therefore lim supxaf(x)f(a) at every point a which proves that f is upper semi-continuous on R. However f is continuous at all aRQ and discontinuous at all aQ.

Converse of fundamental theorem of calculus

The fundamental theorem of calculus asserts that for a continuous real-valued function f defined on a closed interval [a,b], the function F defined for all x[a,b] by
F(x)=xaf(t)dt is uniformly continuous on [a,b], differentiable on the open interval (a,b) and F(x)=f(x)
for all x(a,b).

The converse of fundamental theorem of calculus is not true as we see below.

Consider the function defined on the interval [0,1] by f(x)={2xsin(1/x)cos(1/x) for x00 for x=0 f is integrable as it is continuous on (0,1] and bounded on [0,1]. Then F(x)={x2sin(1/x) for x00 for x=0 F is differentiable on [0,1]. It is clear for x(0,1]. F is also differentiable at 0 as for x0 we have |F(x)F(0)x0|=|F(x)x||x|. Consequently F(0)=0.

However f is not continuous at 0 as it does not have a right limit at 0.

Counterexamples around series (part 2)

We follow the article counterexamples around series (part 1) providing additional funny series examples.

If un converges and (un) is non-increasing then un=o(1/n)?

This is true. Let’s prove it.
The hypotheses imply that (un) converges to zero. Therefore un0 for all nN. As un converges we have limnnk=n/2uk=0. Hence for ϵ>0, one can find NN such that ϵnk=n/2uk12(nun)0 for all nN. Which concludes the proof.

un convergent is equivalent to u2n and u2n+1 convergent?

Is not true as we can see taking un=(1)nn. un converges according to the alternating series test. However for nN nk=1u2k=nk=112k=1/2nk=11k. Hence u2n diverges as the harmonic series diverges.

un absolutely convergent is equivalent to u2n and u2n+1 absolutely convergent?

This is true and the proof is left to the reader.

un is a positive convergent series then (nun) is bounded?

Is true. If not, there would be a subsequence (uϕ(n)) such that ϕ(n)uϕ(n)2. Which means uϕ(n)2ϕ(n) for all nN and implies that the sequence (un) is unbounded. In contradiction with the convergence of the series un.

If (un) is strictly positive with un=o(1/n) then (1)nun converges?

It does not hold as we can see with un={1nlnnn0[2]12nn1[2] Then for nN 2nk=1(1)kuknk=112kln2k2nk=112knk=112kln2k1. As 12kln2k diverges as can be proven using the integral test with the function x12xln2x, (1)nun also diverges.

A nonzero continuous map orthogonal to all polynomials

Let’s consider the vector space C0([a,b],R) of continuous real functions defined on a compact interval [a,b]. We can define an inner product on pairs of elements f,g of C0([a,b],R) by f,g=baf(x)g(x) dx.

It is known that fC0([a,b],R) is the always vanishing function if we have xn,f=baxnf(x) dx=0 for all integers n0. Let’s recall the proof. According to Stone-Weierstrass theorem, for all ϵ>0 it exists a polynomial P such that fPϵ. Then 0baf2=baf(fP)+bafP=baf(fP)fϵ(ba) As this is true for all ϵ>0, we get baf2=0 and f=0.

We now prove that the result becomes false if we change the interval [a,b] into [0,), i.e. that one can find a continuous function fC0([0,),R) such that 0xnf(x) dx for all integers n0. In that direction, let’s consider the complex integral In=0xne(1i)x dx. In is well defined as for x[0,) we have |xne(1i)x|=xnex and 0xnex dx converges. By integration by parts, one can prove that In=n!(1i)n+1=(1+i)n+12n+1n!=eiπ4(n+1)2n+12n!. Consequently, I4p+3R for all p0 which means 0x4p+3sin(x)ex dx=0 and finally 0upsin(u1/4)eu1/4 dx=0 for all integers p0 using integration by substitution with x=u1/4. The function usin(u1/4)eu1/4 is one we were looking for.

Counterexamples around series (part 1)

The purpose of this article is to provide some basic counterexamples on real series. Counterexamples are provided as answers to questions.

Unless otherwise stated, (un)nN and (vn)nN are two real sequences.

If (un) is non-increasing and converges to zero then un converges?

Is not true. A famous counterexample is the harmonic series 1n which doesn’t converge as 2pk=p+11k2pk=p+112p=1/2, for all pN.

If un=o(1/n) then un converges?

Does not hold as can be seen considering un=1nlnn for n2. Indeed x2dttlnt=ln(lnx)ln(ln2) and therefore 2dttlnt diverges. We conclude that 1nlnn diverges using the integral test. However nun=1lnn converges to zero. Continue reading Counterexamples around series (part 1)

Counterexamples on real sequences (part 3)

This article is a follow-up of Counterexamples on real sequences (part 2).

Let (un) be a sequence of real numbers.

If u2nun1n then (un) converges?

This is wrong. The sequence
un={0for n{2k ; kN}12kfor n=2k
is a counterexample. For n>2 and n{2k ; kN} we also have 2n{2k ; kN}, hence u2nun=0. For n=2k 0u2k+1u2k=2k2k12k=1n and limku2k=1. (un) does not converge as 0 and 1 are limit points.

If limnun+1un=1 then (un) has a finite or infinite limit?

This is not true. Let’s consider the sequence
un=2+sin(lnn) Using the inequality |sinpsinq||pq|
which is a consequence of the mean value theorem, we get |un+1un|=|sin(ln(n+1))sin(lnn)||ln(n+1)ln(n)| Therefore limn(un+1un)=0 as limn(ln(n+1)ln(n))=0. And limnun+1un=1 because un1 for all nN.

I now assert that the interval [1,3] is the set of limit points of (un). For the proof, it is sufficient to prove that [1,1] is the set of limit points of the sequence vn=sin(lnn). For y[1,1], we can pickup xR such that sinx=y. Let ϵ>0 and MN , we can find an integer NM such that 0<ln(n+1)ln(n)<ϵ for nN. Select kN with x+2kπ>lnN and Nϵ with lnNϵ(x+2kπ,x+2kπ+ϵ). This is possible as (lnn)nN is an increasing sequence and the length of the interval (x+2kπ,x+2kπ+ϵ) is equal to ϵ. We finally get |uNϵy|=|sin(lnNϵ)sin(x+2kπ)|(lnNϵ(x+2kπ))ϵ proving that y is a limit point of (un).

A strictly increasing continuous function that is differentiable at no point of a null set

We build in this article a strictly increasing continuous function f that is differentiable at no point of a null set E. The null set E can be chosen arbitrarily. In particular it can have the cardinality of the continuum like the Cantor null set.

A set of strictly increasing continuous functions

For p<q two real numbers, consider the function fp,q(x)=(qp)[π2+arctan(2xpqqp)] fp,q is positive and its derivative is fp,q(x)=21+(2xpqqp)2 which is always strictly positive. Hence fp,q is strictly increasing. We also have limxfp,q(x)=0 and limxfp,q(x)=π(qp). One can notice that for x(p,q), fp,q(x)>1. Therefore for x,y(p,q) distinct we have according to the mean value theorem fp,q(y)fp,q(x)yx1.

Covering E with an appropriate set of open intervals

As E is a null set, for each nN one can find an open set On containing E and measuring less than 2n. On can be written as a countable union of disjoint open intervals as any open subset of the reals. Then I=mNOm is also a countable union of open intervals In with nN. The sum of the lengths of the In is less than 1. Continue reading A strictly increasing continuous function that is differentiable at no point of a null set

A monotonic function whose points of discontinuity form a dense set

Consider a compact interval [a,b]R with a<b. Let’s build an increasing function f:[a,b]R whose points of discontinuity is an arbitrary dense subset D={dn ; nN} of [a,b], for example D=Q[a,b].

Let pn be a convergent series of positive numbers whose sum is equal to p and define f(x)=dnxpn.

f is strictly increasing

For ax<yb we have f(y)f(x)=x<dnypn>0 as the pn are positive and dense so it exists pm(x,y].

f is right-continuous on [a,b]

We pick-up x[a,b]. For any ϵ>0 is exists NN such that 0<n>Npn<ϵ. Let δ>0 be so small that the interval (x,x+δ) doesn’t contain any point in the finite set {p1,,pN}. Then 0<f(y)f(x)n>Npn<ϵ, for any y(x,x+δ) proving the right-continuity of f at x. Continue reading A monotonic function whose points of discontinuity form a dense set

A function whose Maclaurin series converges only at zero

Let’s describe a real function f whose Maclaurin series converges only at zero. For n0 we denote fn(x)=encosn2x and f(x)=n=0fn(x)=n=0encosn2x. For k0, the kth-derivative of fn is f(k)n(x)=enn2kcos(n2x+kπ2) and |f(k)n(x)|enn2k for all xR. Therefore n=0f(k)n(x) is normally convergent and f is an indefinitely differentiable function with f(k)(x)=n=0enn2kcos(n2x+kπ2). Its Maclaurin series has only terms of even degree and the absolute value of the term of degree 2k is (n=0enn4k)x2k(2k)!>e2k(2k)4kx2k(2k)!>(2kxe)2k. The right hand side of this inequality is greater than 1 for ke2x. This means that for any nonzero x the Maclaurin series for f diverges.

Painter’s paradox

Can you paint a surface with infinite area with a finite quantity of paint? For sure… let’s do it!

Consider the 3D surface given in cylindrical coordinates as S(ρ,φ):{x=ρcosφy=ρsinφz=1ρ for (ρ,φ)[1,)×[0,2π). The surface is named Gabriel’s horn.

Volume of Gabriel’s horn

The volume of Gabriel’s horn is V=π1(1ρ2) dρ=π which is finite.

Area of Gabriel’s horn

The area of Gabriel’s horn for (ρ,φ)[1,a)×[0,2π) with a>1 is: A=2πa11ρ1+(1ρ2)2 dρ2πa1dρρ=2πloga. As the right hand side of inequality above diverges to as a, we can conclude that the area of Gabriel’s horn is infinite.

Conclusion

Gabriel’s horn could be filled with a finite quantity of paint… therefore painting a surface with infinite area. Unfortunately the thickness of the paint coat is converging to 0 as z goes to , leading to a paint which won’t be too visible!