Loading [MathJax]/jax/output/HTML-CSS/jax.js

Sunday, June 27, 2021

Beta function

dirichlet beta

Derivative of Dirichlet Beta function


We prove the following infinite product which turned out to be a modified version of the Dirichlet Beta function derivative. 133557799111113131515=(πeγΓ(34)4)π4 Proof

Take logarithm in both sides ln(3)3+ln(5)5ln(7)7+ln(9)9ln(11)11+ln(13)13...=π4[ln(π)γ4lnΓ(34)] Then, we want to prove n=1(1)n+1ln(2n1)(2n1)=π4[ln(π)γ4lnΓ(34)] Consider the Dirichlet beta function β(v)=n=1(1)n+1(2n1)v Then β(v)=n=1(1)nln(2n1)(2n1)v Therefore we want to find β(1)=n=1(1)n+1ln(2n1)(2n1) There are various ways to find the derivative, one of the most straigforward is using the Kummer's Fourier series for the log-gamma function: lnΓ(t)=12ln(πsinπt)+[γ+ln(2π)](12t)+1πn=1lnnnsin(2πnt) It turned out that if t=14 then sin(πn2)={sin(π2)=1 if n=1sin(π)=0 if n=2sin(3π2)=1 if n=3sin(2π)=0 if n=4} Then n=1lnnnsin(πn2)=n=1(1)n+1ln(2n1)(2n1) Therefore, n=1(1)n+1ln(2n1)(2n1)=π[lnΓ(14)12ln(πsinπ4)[γ+ln(2π)](14)]=π[ln(π)12ln(2)lnΓ(34)12ln(π)14ln(2)14γln(2π)4]=π4[ln(π)γ4Γ(34)] Therefore 133557799111113131515=(πeγΓ(34)4)π4

Thursday, June 24, 2021

Application of Ramanujan's master theorem

Bessel function

Mellin transform of the Bessel function

We show the following result 0(x2)sJs(x)dx=Γ(s) Proof
We will use the Ramanujan's master theorem to prove this result: Assume that the function f has an expansion of the form f(x)=k=0(ϕ(k)k!)(x)k for some analytic function ϕ(k), then the Mellin transform of f(x) is given by {Mf(x)}=0xs1f(x)dx=Γ(s)ϕ(s) Back to our problem, recall the expansion series of the Bessel function: Js(x)=(x2)sj=0(x2/4)jj!Γ(1+j+s) Then 0(x2)s1Js(x)dx=0(x2)2s1j=0(x2/4)jj!Γ(1+j+s)dx let w=x24 then dw=x2dx, therefore 0(x2)2s1j=0(x2/4)jj!Γ(1+j+s)dx=0ws1j=0(w)jj!Γ(1+j+s)dw Clearly f(w)=j=0(w)jj!Γ(1+j+s) and ϕ(j)=1Γ(1+j+s) By Ramanujan's master theorem : 0ws1j=0(w)jj!Γ(1+j+s)dw=Γ(s)ϕ(s)=Γ(s)Γ(1)=Γ(s) Then 0(x2)sJs(x)dx=Γ(s)

Integral of the day

Integral involving the compostion of sin function with itself

We prove the following integral: π20sinsinxdx=113!!2+15!!217!!2+... Proof
Recall the series expansion of sin(x) sinx=n=0(1)n(2n+1)!x2n+1xR Therefore π20sinsinxdx=π20n=0(1)n(2n+1)!sin2n+1(x)dx=n=0(1)n(2n+1)!π20sin2n+1(x)dx=n=0(1)n(2n+1)!10w2n+1(1w2)12dw(wsinx)=n=0(1)n(2n+1)!10tn(1t)122dt(tw2)=n=0(1)n(2n+1)!B(n+1,12)2=n=0(1)n(2n+1)!Γ(n+1)Γ(12)2Γ(n+1+12)=n=0(1)n(2n+1)!n!π2Γ(n+1+12)(Γ(12)=π)) Now recall the Legrende duplication formula πΓ(2z)=22z1Γ(z)Γ(z+12) Applying the formula to Γ(n+1+12) n=0(1)n(2n+1)!n!π2Γ(n+1+12)=n=0(1)n(2n+1)!22nn!Γ(n+1)Γ(2n+2)=n=0(1)n(2n+1)!22nn!2(2n+1)!=n=0(1)n22nn!2(2n+1)!2=n=0(1)n1(2n+1)!!2 π20sinsinxdx=n=0(1)n1(2n+1)!!2

Wednesday, June 23, 2021

Feynman's trick II

log integral

Formula involving integrals of logarithm function

We prove the following formula involving the integral of powers of the function ln(1x). 10lnn(1x)ln(x)xdx=(1)n+1n![(n+12)ζ(n+2)12n1k=1ζ(k+1)ζ(n+1k)] Some examples for the first naturals: n=110ln(1x)ln(x)xdx=ζ(3)n=210ln2(1x)ln(x)xdx=3ζ(4)+ζ(2)ζ(2)=π4180n=310ln3(1x)ln(x)xdx=12ζ(5)6ζ(2)ζ(3)=12ζ(5)π2ζ(3)n=410ln4(1x)ln(x)xdx=60ζ(6)+24ζ(2)ζ(4)+12ζ(3)ζ(3)=12ζ2(3)2π6105n=510ln5(1x)ln(x)xdx=360ζ(7)120ζ(2)ζ(5)120ζ(3)ζ(4)=4π4ζ(3)320π2ζ(5)+360ζ(7) Proof 10lnn(1x)ln(x)xdx=10[dndtn|t=0+(1x)t][dds|s=0+xs]xdx=dndtn|t=0+dds|s=0+10(1x)txsxdx=dndtn|t=0+dds|s=0+10(1x)txs1dx=dndtn|t=0+dds|s=0+B(t+1,s) Now, recall the series expansion of the beta function (proof in Appendix 1): B(x,y)=j=0(1y)jj!(x+j)y>0 Therefore dndtn|t=0+dds|s=0+B(t+1,s)=dndtn|t=0+dds|s=0+j=0(1s)jj!(t+1+j)=dndtn|t=0+[lims0+ddsj=0(1s)jj!(t+1+j)]=dndtn|t=0+[lims0+j=0(1s)j(ψ(0)(1s)ψ(0)(1s+j))j!(t+1+j)]=dndtn|t=0+[j=0(1)j(ψ(0)(1)ψ(0)(1+j))j!(t+1+j)]=dndtn|t=0+[j=0Γ(j+1)(γψ(0)(1+j))j!(t+1+j)]=dndtn|t=0+[j=0Γ(j+1)Hjj!(t+1+j)]=dndtn|t=0+[j=0Hj(t+1+j)]=limt0+j=0dndtnHj(t+1+j)=limt0+j=0(1)n+1n!Hj(t+1+j)n+1=j=0(1)n+1n!Hj(1+j)n+1 Now we know that Hj+1=1j+1+Hj therefore j=0(1)n+1n!Hj(1+j)n+1=j=0(1)n+1n!Hj+1(1+j)n+1j=0(1)n+1n!(1+j)n+2=(1)n+1n![j=0Hj+1(1+j)n+1j=01(1+j)n+2] If m=j+1 then (1)n+1n![j=0Hj+1(1+j)n+1j=01(1+j)n+2]=(1)n+1n![m=1Hmmn+1m=11mn+2]=(1)n+1n![m=1Hmmn+1ζ(n+2)] Now, recall this beutiful result proven by Euler (Proof in Appendix 2): n=1Hnnr=(1+r2)ζ(r+1)12r2k=1ζ(k+1)ζ(rk) Then (1)n+1n![m=1Hmmn+1ζ(n+2)]=(1)n+1n![(1+n+12)ζ(n+2)12n1k=1ζ(k+1)ζ(n+1k)ζ(n+2)]=(1)n+1n![(n+12)ζ(n+2)12n1k=1ζ(k+1)ζ(n+1k)] Therefore 10lnn(1x)ln(x)xdx=(1)n+1n![(n+12)ζ(n+2)12n1k=1ζ(k+1)ζ(n+1k)]

Apendix 1: Series expansion for the beta function

From the definition of the beta function: B(x,y)=10tx1(1t)y1dt Recall the series expansion of (1t)y1 (1t)y1=j=0(1)j(y1j)tj 10tx1(1x)y1dt=10tx1j=0(1)j(y1j)tjdt=j=0(1)j(y1j)0tx+j1dt=j=0(1)j(y1j)1(x+j) Then we have j=0(1)j(y1j)1(x+j)=j=0(jyj)1(x+j)=j=0(jy)!j!(y)!1(x+j)=j=0Γ(jy+1)j!Γ(1y)1(x+j)=j=0(1y)jj!(x+j) where the last equality is the definnition of the Pochhammer polynomials. Therefore B(x,y)=j=0(1y)jj!(x+j)

Appendix 2: Proof of Euler's theorem for harmonic series

I found a beautiful proof of this result which relies on the residue theorem. I really liked this proof because it shows that the famous result due to Euler is just a residue of certain kernel function. I present here the original proof in the spirit of Flajolet et al. [1] but I complete the omitted parts by the authors in the original paper.

We need the following Lemma due to Cauchy and Lindelöf.

Lemma. Let ξ be a kernel function and let r(s) be a rational function with O(s2) at infinity. Then α{Res(r(s)ξ(s),α)|α is a pole of ξ that is not a pole of r(s)}infinite series=β{Res(r(s)ξ(s),β)|β is a pole of r(s)}finite series

We will apply the lemma to the kernel ξ(s)=(ψ(s)+γ)2

Propositon. If ξ(s)=(ψ(s)+γ)2 then
α{Res(r(s)ξ(s),α)|α is a pole of ξ that is not a pole of r(s)}=2n=1Hnr(n)+n=1r(n)
Proof

The kernel ξ(s) has a countable number of pole at the positive integers.
Let nN, then the function (ψ(s)+γ) has the following expansion at sn.

ψ(s)+γ=1(sn)+Hn+k=1[(1)KHk+1nζ(k+1)](sn)k Therefore
(ψ(s)+γ)2=[1(sn)+Hn+k=1[(1)KHk+1nζ(k+1)]Ak(sn)k]2=(1(sn)+Hn)2+2(1(sn)+Hn)k=1Ak(sn)k+(k=1Ak(sn)k)2=1(sn)2+2Hn(sn)+H2n+2(1(sn)+Hn)k=1Ak(sn)k+(k=1Ak(sn)k)2=1(sn)2+2Hn(sn)+H2n+2A1+O(sn) Now, suppose r(s) is a rational function that can be expanded with its power series (in reality, we have assumed something weaker: that r(s) is O(s2)). Then for sn: r(s)=r(n)+r(n)(sn)+r(n)2!(sn)2+r(n)3!(sn)3+... Therefore r(s)(ψ(s)+γ)2=r(n)(ψ(s)+γ)2+r(n)(sn)(ψ(s)+γ)2+r(n)2!(sn)2(ψ(s)+γ)2+...=r(n)(sn)2+2Hnr(n)(sn)+H2nr(n)+2A1r(n)+O(sn)r(n)+r(n)(sn)+2Hnr(n)+H2nr(n)(sn)+2A1r(n)(sn)+O(sn)r(n)(sn)+...=H2nr(n)+2A1r(n)+2Hnr(n)+r(n)2!+1sn[2Hnr(n)+r(n)]residue at n+r(n)(sn)2+O(sn) Res(r(s)(ψ(s)+γ)2,n)=2Hnr(n)+r(n) Summing each nN{0} n=1Res(r(s)(ψ(s)+γ)2,n)=2n=1Hnr(n)+n=1r(n) or, equivalently α{Res(r(s)ξ(s),α)|α is a pole of ξ that is not a pole of r(s)}=2n=1Hnr(n)+n=1r(n)

We are ready to prove Euler's theorem

Theorem (Euler). For integer 2, n=1Hnnq=(1+q2)ζ(q+1)12q2k=1ζ(k+1)ζ(qk) Proof If we apply the last result to the function r(s)=1sq we have n=1Res((ψ(s)+γ)2sq,n)=2n=1Hnnqqζ(q+1) By the first lemma, β{Res((ψ(s)+γ)2sq,β)|β is a pole of 1sq}=2n=1Hnnqqζ(q+1) Given that r(s)=1sq has a single pole at s=0 of order q we have: Res((ψ(s)+γ)2sq,0)=2n=1Hnnqqζ(q+1)

To find the residue we can expand (ψ(s)+γ)2sq:
First, recall the series expansion of ψ(s)+γ at s=0:
ψ(s)+γ=1sk=1ζ(k+1)sk Therefore (ψ(s)+γ)2=[1sk=1ζ(k+1)sk]2=1s22k=1ζ(k+1)sk1+(k=1ζ(k+1)sk)2=1s22k=1ζ(k+1)sk1+k=1kj=1ζ(j+1)ζ(kj+2)sk+1 Then (ψ(s)+γ)2sq=1s2+q2k=1ζ(k+1)sk1qA+k=1kj=1ζ(j+1)ζ(kj+2)sk+1qB From here is easy to show that the residue in A is the coefficient of sk1q=s1 k1q=1k=q and the residue in B is coefficient of sk+1q=s1 where k+1q=1k=q2 Therefore Res((ψ(s)+γ)2sq,0)=2ζ(q+1)+q2j=1ζ(j+1)ζ(qj) Then 2ζ(q+1)q2j=1ζ(j+1)ζ(qj)=2n=1Hnnqqζ(q+1) n=1Hnnq=(1+q2)ζ(q+1)12q2k=1ζ(k+1)ζ(qk)

[1] Philippe Flajolet, Bruno Salvy. Euler Sums and Contour Integral Representations. [Research Report] RR-2917, INRIA. 1996. ffinria-00073780f

Friday, June 18, 2021

Cauchy integral formula

triple exponential

Triple exponential integral

We show the proof of the following integral: π0eecosxcossinxcos(ecosxsinsinx)dx=πe On the way we will find this beautiful result: π0eeeix+eeeixdx=2πe Proof π0eecosxcossinxcos(ecosxsinsinx)dx=π0eecosxcossinx[eiecosxsinsinx+eiecosxsinsinx2]dx=12π0eecosxcossinx+iecosxsinsinx+eecosxcossinxiecosxsinsinxdx=12π0eecosx(cossinx+isinsinx)+eecosx(cossinxisinsinx)dx=12π0eecosxeisinx+eecosxeisinxdx=12π0eecosx+isinx+eecosxisinxdx=12π0eeeix+eeeixdx=12π0[n=0(eeix)nn!+0(eeix)nn!]dx=12π0n=0(eeix)n+(eeix)nn!dx=12n=01n!π0(eeix)n+(eeix)ndx=12n=01n![π0(eeix)ndx+π0(eeix)ndx]=12n=01n![π0(eeix)ndx0π(eeix)ndx]=12n=01n![π0(eeix)ndx+0π(eeix)ndx]=12n=01n!ππ(eeix)ndx=12n=01n!γeznzidz(where γ(x)=eixx[π,π])=12n=01n!2π(Cauchy integral formula)=πe π0eecosxcossinxcos(ecosxsinsinx)dx=πe Corollary π0eeeix+eeeixdx=2πe

Thursday, June 17, 2021

Riemann zeta function

zeta series

Series involving the zeta function, e and pi

We show the proof of this beutiful series posted by @infseriesbot. exp(ζ(2)2ζ(2)3+ζ(4)4ζ(4)5+ζ(6)6...)=2πe To prove this result we use two concepts:
  1. The Taylor series expansion of the log-Gamma function lnΓ(x)
  2. The Euler's reflection formula for the Gamma function Γ(x)
Lets start with the left hand side ζ(2)2ζ(2)3+ζ(4)4ζ(4)5+ζ(6)6...=n=1(12n12n+1)ζ(2n)=n=1ζ(2n)(2n)(2n+1) What is the series of the right hand side? This series is related to the expansion series of the log-Gamma function: ln(Γ(z))=γ(z1)+k=2(1)kζ(k)k(z1)k|z1|<1 where γ is the Euler-Mascheroni constant. If we let z=1+t and then z=1t in the expansion series we obtain two equations: ln(Γ(t+1))=γt+k=2(1)kζ(k)ktkln(Γ(1t))=γt+k=2ζ(k)ktk Adding both equations: ln(Γ(t+1))+ln(Γ(1t))=k=1ζ(2k)2kt2k Integrating both sides: 10(ln(Γ(t+1))+ln(Γ(1t)))dt=k=1ζ(2k)k10t2kdt1210(ln(Γ(t+1))+ln(Γ(1t)))dt=k=1ζ(2k)2k(2k+1) This is the series we were looking for in terms of a definite integral. Hopefully we can transform it to a maneuverable one. Here is where the Euler's reflection formula for the Gamma function is useful: Γ(t)Γ(1t)=πsin(πt)tZ Therefore 1210(ln(Γ(t+1))+ln(Γ(1t)))dt=1210(ln(Γ(t+1))(Γ(1t)))dt=1210(ln(tΓ(t))(Γ(1t)))dt=1210ln(tπsin(πt))dt=12ln(π)121210ln(sin(tπ))dt We just have to prove that 10ln(sin(tπ))dt=ln(2) 10ln(sin(tπ))dt=120ln(sin(2πt))dt=ln(2)+120(sin(πt))dt+120ln(cos(πt))dt(sin(2t)=2sin(t)cos(t)) We also know that 10ln(sin(tπ))dt=2120ln(sin(πt))dt(using sin(θ)=sin(θ))=120ln(sin(2πt))dt Therefore 10ln(sin(tπ))dtln(2)+120ln(sin(πt))dt+120ln(cos(πt))dt=ln(2)+120ln(sin(πt))dt+120ln(sin(πw))dw(w=12t)=ln(2)+210ln(sin(πt))dt+210ln(sin(πw))dw 10ln(sin(tπ))dt=ln(2)+10ln(sin(tπ))dt+10ln(sin(tπ))dt10ln(sin(πt))dt=ln(2) We conclude from results (1),(2) and (3) our claim: ζ(2)2ζ(2)3+ζ(4)4ζ(4)5+ζ(6)6...=12ln(π)1210ln(sin(tπ))dt=12ln(π)12+12ln(2) exp(ζ(2)2ζ(2)3+ζ(4)4ζ(4)5+ζ(6)6...)=2πe

Wednesday, June 16, 2021

Feynman's trick

Feynman

Integral involving the Feynman's integral trick

Today we show the proof of this result which involves the famous Feynman's integral trick twice. The post of @infseriesbot has a typo (the sign is wrong). First, we differentiate the parameter a under the integral sign: dda0(1cosax)ln(x)x2dx=0dda(1cosax)ln(x)x2dx=0sinaxln(x)xdx=0sinwln(wa)wdw(wax)=0sinwln(w)wdwln(a)0sinwwdw=0sinwln(w)wdwπln(a)2 =0sinwddtwt|t=0+wdwπln(a)2=ddt|t=0+0wt1sinwdwπln(a)2 Here we are using the fact that ddtwt|t=0+=ln(w) and the fact that 0sin(w)wdw=π2 (a famous result obtained through complex analysis). Note that the first integral is the Mellin transform of sinw, so we could use the following expansion series of sin (a consequence of the Euler's formula and the de Moivre theorem) in oder to apply the Ramanujan's master theorem: sinw=n=0(w)n[sin(nπ2)]n! Then, by Ramanujan's master theorem: ddt|t=0+0wt1sinwdw=ddt|t=0+[Γ(t)sin(tπ2)]=limt0+Γ(t)[π2cos(tπ2)+ψ0(t)sin(πt2)]=limt0+Γ(t+1)[π2cos(tπ2)t+ψ0(t)sin(πt2)t] Update: I found a cool way to calculate this limit:
Note that π2cos(tπ2)t=π21tπ323t2!+π525t34!π727t56!+...=π21t+O(t) Also note that ψ0(t)sin(πt2)t=ψ0(t)π2ψ0(t)π323t23!+ψ0(t)π525t45!ψ0(t)π727t67!+...=ψ0(t)π2tψ0(t)(π323t3!π525t35!+π727t57!...)=ψ0(t)π2tψ0(t)O(t) Therefore π2cos(tπ2)t+ψ0(t)sin(πt2)t=π21t+ψ0(t)π2+O(t)tψ0(t)O(t)=π2(1t+ψ0(t))+O(t)tψ0(t)O(t)=π2ψ0(t+1)+O(t)tψ0(t)O(t) limt0+Γ(t+1)[π2cos(tπ2)t+ψ0(t)sin(πt2)t]=limt0+Γ(t+1)[π2ψ0(t+1)+O(t)tψ0(t)O(t)]=π2γ
where limt0+ψ0(t+1)=γ and limt0+tψ0(t)=1

Therefore
dda0(1cosax)ln(x)x2dx=π2γπln(a)20(1cosax)ln(x)x2dx=π2aγ+aln(a)da+C=aπγ2aπln(a)2+πa2+C Note that if a=0 then C=0 therefore 0(1cosax)ln(x)x2dx=πa2(γ+ln(a)1)

Monday, June 14, 2021

Continued fractions

coth

Continued fraction expansion for coth(z)

We will show the proof for this continued fraction posted by @infseriesbot. e+1e1=coth(12)=2+16+110+114+ Update We add the proof of this result, which is bascially the same function evaluated at other point: 0=coth(iπ2)=2π26π210π214
Recall the continued fraction representation for ratios of confluent hypergeometric functions of the type 0F1(c,z): Let (an) be a sequence of complex numbers defined by an=1(c+n1)(c+n) where c is a complex constant such that cZ{0}. Then 1+Kn=1(anz1)=0F1(c,z)0F1(c+1,z)
Now, for this particular case, we will use the hypergemetric representation of sin(z) and cos(z) sin(z)=z0F1(32,z24) sin(z)=0F1(12,z24) Then tan(x)=sin(x)cos(x)=xx33!+x55!x77!+...1+x22!+x44!x66!+...=x(1x23!+x45!x67!+...)1+x22!+x44!x66!+...=x0F1(32,x24)0F1(12,x24)=x1x23x25x27 If we evaluate at x=iz2 tan(iz2)=iz22z26z210z214=iz2+z26+z210+z214coth(z2)z=cot(iz2)iz=2+z26+z210+z214+ If we evaluate at z=1 e+1e1=coth(12)=2+16+110+114+ Since the result is valid for al the complex plane, if we evaluate at z=iπ then coth(iπ2)=0 and 0=coth(iπ2)=2π26π210π214

Residue theorem

niceintegral

Integral involving the Residue theorem

Today we present here the proof of this nice integral. The solution involves Complex Analysis, geometric series, even functions and trigonometry 0cosxe2πx1dx=0[n=1e2πnxcosx]dx=n=10e2πnxcosxdx=n=10e2πnx(eix+eix2)dx=n=10ex(i2πn)+ex(i2πn)2dx=12n=10ex(i2πn)dx+12n=10ex(i2πn)dx=12n=1ex(i2πn)(1+(i2πn)x)(i2πn)2|0+12n=1ex(i2πn)(1+(i2πn)x)(i2πn)2|0=n=11(i2πn)2+n=11(i2πn)2=n=1[1(i2πn)2+1(i2πn)2]=2+n=0[1(i2πn)2+1(i2πn)2]=1+12n=1(i2πn)2+12n=1(i2πn)2=112Res(πcotπz(i2πz)2,i2π)12Res(πcotπ(i2πz)2,i2π)=112limzi2π[ddz(zi2π)2(πcotπz)(i2πz)2]12limzi2π[ddz(z+i2π)2(πcotπz)(i2πz)2]=118πlimzi2πddzcotπz18πlimzi2πddzcotπz=1+18csc2(i2)+18csc2(i2)=118(2ie12e12)218(2ie12e12)2=1e(e1)2

Thursday, June 10, 2021

Gaussian integral

Mathedemo

Generalization of Gaussian integral

We are going to prove this result posted by @infeseriesbot that is a generalization of the Gaussian integral (s=12). n=0(x2)nn!2sdx=π1sRe(s)>0 It was previously discussed in Twitter a few months ago and turned out that the proof is a consquence of the Ramanujan's master theorem. However, other proofs were provided using the bracketts method, the log-gamma function and contour integration. In the proof we also find the values of s that guarantee the convergence of the integral (Re(s)>0).

Proof. Recall the Ramanujan's master theorem: If a function f(x) has an expansion of the form f(x)=k=0ϕ(k)k!(x)k for some function (say analytic or integrable) φ(k), then the Mellin transform of f(x) is given by {Mf(x)}(t)=0xt1f(x)dx=Γ(t)φ(t) In this particular case n=0(x2)nn!2sdx=20n=0(x2)nn!2sdx=0n=0w121(w)nn!2sdw(wx2)=0w121n=0(w)nn!2sdw=0w121n=0(n!)12s(w)nn!dw this is a Mellin transform {Mf(w)}(12)=0w121f(w)dw  where f(w)=n=0(n!)12s(w)nn!and ϕ(t)=(t)!12s=[Γ(1t)]12s By the Ramanujan's master theorem 0w121n=0(n!)12s(w)nn!dw=Γ(12)[Γ(12)]12s=π1s The result is valid for Re(s)>0. To show this, lets review the convergence of n=0(x2)nn!2sdx. We can use the ratio test for the sequence ((x2)nn!2s)n limn|(x2)n+1(n+1)!2s(x2)nn!2s|=limn|n!2s(n+1)!2s||x2|=limn|n!2s(n+1)!2s||x2|=limnexp(2Re(slnn!)2Re(sln(n+1)!))|x2|=limnexp(2Re(s)(lnn!ln(n+1)!))|x2|=exp(limnRe(s)2(lnn!ln(n+1)!))|x2|=exp(Re(s)limn2ln(n!(n+1)!))|x2|=exp(Re(s)())|x2| limn|(x2)n+1(n+1)!2s(x2)nn!2s|={0 if Re(s)>0|x2| if Re(s)=0 if Re(s)<0 Therefore n=0(x2)nn!2sdx<if{Re(s)>0Re(s)=0|x|<1 Then, we have obtained the desired result: n=0(x2)nn!2sdx=π1sRe(s)>0

Series of the day

Series involving the digamma and the zeta functions The sum 1(n+1)pnq ...