Conjectured value of a harmonic sum $sum_n=1^inftyleft(H_n-,2H_2n+H_4nright)^2$Infinite Series $sum_n=1^inftyfracH_nn^32^n$Sum of Squares of Harmonic NumbersA closed form of $sum_n=1^inftyleft[ H_n^2-left(ln n+gamma+frac12n right)^2right]$Closed form for $sum_n=1^inftyfrac(-1)^n n^4 H_n2^n$Closed form for $sum_n=1^inftyfrac(-1)^n n^a H_n2^n$A closed form of the series $sum_n=1^infty fracH_n^2-(gamma + ln n)^2n$Infinite Series $sum_n=1^inftyfracH_nn^32^n$A conjectured result for $sum_n=1^inftyfrac(-1)^n,H_n/5n$Strategies for evaluating sums $sum_n=1^infty fracH_n^(m)z^nn$How to evaluate $lim_nrightarrow infty nleft [ widetildeH_n-H_2n+H_n right ]$?What's about $sum_n=1^infty frace^H_nlog H_nn^3$, where $H_n$ is the nth harmonic number?A closed form of $sum_n=1^inftyleft[ H_n^2-left(ln n+gamma+frac12n right)^2right]$Justify an approximation of $-sum_n=2^infty H_nleft(frac1zeta(n)-1right)$, where $H_n$ denotes the $n$th harmonic number
Good allowance savings plan?
Should QA ask requirements to developers?
Is it true that real estate prices mainly go up?
How did Alan Turing break the enigma code using the hint given by the lady in the bar?
Offered promotion but I'm leaving. Should I tell?
BitNot does not flip bits in the way I expected
How strictly should I take "Candidates must be local"?
How could our ancestors have domesticated a solitary predator?
Unreachable code, but reachable with exception
Make a transparent 448*448 image
Why does Captain Marvel assume the planet where she lands would recognize her credentials?
Is "history" a male-biased word ("his+story")?
Peter's Strange Word
Can you reject a postdoc offer after the PI has paid a large sum for flights/accommodation for your visit?
Can't find the Shader/UVs tab
What to do when during a meeting client people start to fight (even physically) with each others?
How are such low op-amp input currents possible?
Is there an equal sign with wider gap?
Low budget alien movie about the Earth being cooked
2×2×2 rubik's cube corner is twisted!
How do I express some one as a black person?
Do Bugbears' arms literally get longer when it's their turn?
Subset counting for even numbers
Placing subfig vertically
Conjectured value of a harmonic sum $sum_n=1^inftyleft(H_n-,2H_2n+H_4nright)^2$
Infinite Series $sum_n=1^inftyfracH_nn^32^n$Sum of Squares of Harmonic NumbersA closed form of $sum_n=1^inftyleft[ H_n^2-left(ln n+gamma+frac12n right)^2right]$Closed form for $sum_n=1^inftyfrac(-1)^n n^4 H_n2^n$Closed form for $sum_n=1^inftyfrac(-1)^n n^a H_n2^n$A closed form of the series $sum_n=1^infty fracH_n^2-(gamma + ln n)^2n$Infinite Series $sum_n=1^inftyfracH_nn^32^n$A conjectured result for $sum_n=1^inftyfrac(-1)^n,H_n/5n$Strategies for evaluating sums $sum_n=1^infty fracH_n^(m)z^nn$How to evaluate $lim_nrightarrow infty nleft [ widetildeH_n-H_2n+H_n right ]$?What's about $sum_n=1^infty frace^H_nlog H_nn^3$, where $H_n$ is the nth harmonic number?A closed form of $sum_n=1^inftyleft[ H_n^2-left(ln n+gamma+frac12n right)^2right]$Justify an approximation of $-sum_n=2^infty H_nleft(frac1zeta(n)-1right)$, where $H_n$ denotes the $n$th harmonic number
$begingroup$
There is a known asymptotic expansion of harmonic numbers $H_n$ for $ntoinfty$:
$$beginalignH_n&=gamma+ln n+sum_k=1^inftyleft(-fracB_kkcdot n^kright)\
&=gamma+ln n+frac12n-frac112n^2+frac1120n^4-frac1252n^6,+,dots,endaligntag1$$
where $B_k$ are Bernoulli numbers.
We can take a linear combination of harmonic numbers to cancel constant and logarithmic terms, compensate for $O(n^-1)$ term, and get the following series that is possible to evaluate in a closed form (e.g. using generating function):
$$sum_k=1^inftyleft(H_n-,2H_2n+H_4n-frac18nright)=frac18-fracpi16.tag2$$
Rather than compensating for $O(n^-1)$ term, we can take a series with alternating signs, that is also possible to evaluate in a closed form:
$$sum_n=1^infty,(-1)^n,Big(H_n-,2H_2n+H_4nBig)=frac3pi16-fracpi4sqrt2-fracln28.tag3$$
Generalizing, we can consider two families of series:
$$mathcal A_m=sum_n=1^infty,(-1)^n,Big(H_n-,2H_2n+H_4nBig)^m,tag4$$
$$mathcal B_m=sum_n=1^inftyBig(H_n-,2H_2n+H_4nBig)^m,tag5$$
and try to evaluate them in a closed form.
So far I have the following conjectured result:
$$largesum_n=1^inftyBig(H_n-,2H_2n+H_4nBig)^2stackrelnormalsizecolorgray?=fracpi8-fracpi16,ln2-fracpi^296+frac316,ln^22-fracG4,tag$diamond$$$
where $G$ is the Catalan constant.
Could you please help me to prove this result and, possibly, find other values of $mathcal A_m,mathcal B_m$?
calculus sequences-and-series closed-form conjectures harmonic-numbers
$endgroup$
This question has an open bounty worth +50
reputation from TheSimpliFire ending ending at 2019-03-17 09:11:13Z">in 4 days.
One or more of the answers is exemplary and worthy of an additional bounty.
add a comment |
$begingroup$
There is a known asymptotic expansion of harmonic numbers $H_n$ for $ntoinfty$:
$$beginalignH_n&=gamma+ln n+sum_k=1^inftyleft(-fracB_kkcdot n^kright)\
&=gamma+ln n+frac12n-frac112n^2+frac1120n^4-frac1252n^6,+,dots,endaligntag1$$
where $B_k$ are Bernoulli numbers.
We can take a linear combination of harmonic numbers to cancel constant and logarithmic terms, compensate for $O(n^-1)$ term, and get the following series that is possible to evaluate in a closed form (e.g. using generating function):
$$sum_k=1^inftyleft(H_n-,2H_2n+H_4n-frac18nright)=frac18-fracpi16.tag2$$
Rather than compensating for $O(n^-1)$ term, we can take a series with alternating signs, that is also possible to evaluate in a closed form:
$$sum_n=1^infty,(-1)^n,Big(H_n-,2H_2n+H_4nBig)=frac3pi16-fracpi4sqrt2-fracln28.tag3$$
Generalizing, we can consider two families of series:
$$mathcal A_m=sum_n=1^infty,(-1)^n,Big(H_n-,2H_2n+H_4nBig)^m,tag4$$
$$mathcal B_m=sum_n=1^inftyBig(H_n-,2H_2n+H_4nBig)^m,tag5$$
and try to evaluate them in a closed form.
So far I have the following conjectured result:
$$largesum_n=1^inftyBig(H_n-,2H_2n+H_4nBig)^2stackrelnormalsizecolorgray?=fracpi8-fracpi16,ln2-fracpi^296+frac316,ln^22-fracG4,tag$diamond$$$
where $G$ is the Catalan constant.
Could you please help me to prove this result and, possibly, find other values of $mathcal A_m,mathcal B_m$?
calculus sequences-and-series closed-form conjectures harmonic-numbers
$endgroup$
This question has an open bounty worth +50
reputation from TheSimpliFire ending ending at 2019-03-17 09:11:13Z">in 4 days.
One or more of the answers is exemplary and worthy of an additional bounty.
$begingroup$
Some possibly related questions are linked from here.
$endgroup$
– Piotr Shatalin
Aug 22 '15 at 18:18
3
$begingroup$
BTW, another similar series can be evaluated using generating functions: $sum_n=1^inftyfrac1nleft(H_n-,2H_2n+H_4nright)=frac34 ln^2 2-fracpi^248.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 19:07
2
$begingroup$
And another: $sum_n=1^inftyfrac(-1)^nnleft(H_n-,2H_2n+H_4nright)=frac12 ln^2!left(1+sqrt2right)+frac18 ln^2 2-frac5pi^296.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 19:43
$begingroup$
One way to tackle it, at least the case $m=2$, would be to write $H_n=int_0^1frac1-x^n1-xdx$ and to convert the series into a double integral which probably could be evaluated more easily. But I think this is what we call "Brute Force".
$endgroup$
– Redundant Aunt
Aug 22 '15 at 19:56
4
$begingroup$
One more simple series where $O(n^-1)$ term is cancelled using harmonic numbers only: $sum_n=1^inftyleft(H_n-4H_2n+5H_4n-2H_8nright)=fracpi4sqrt2 - frac3pi16.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 22:14
add a comment |
$begingroup$
There is a known asymptotic expansion of harmonic numbers $H_n$ for $ntoinfty$:
$$beginalignH_n&=gamma+ln n+sum_k=1^inftyleft(-fracB_kkcdot n^kright)\
&=gamma+ln n+frac12n-frac112n^2+frac1120n^4-frac1252n^6,+,dots,endaligntag1$$
where $B_k$ are Bernoulli numbers.
We can take a linear combination of harmonic numbers to cancel constant and logarithmic terms, compensate for $O(n^-1)$ term, and get the following series that is possible to evaluate in a closed form (e.g. using generating function):
$$sum_k=1^inftyleft(H_n-,2H_2n+H_4n-frac18nright)=frac18-fracpi16.tag2$$
Rather than compensating for $O(n^-1)$ term, we can take a series with alternating signs, that is also possible to evaluate in a closed form:
$$sum_n=1^infty,(-1)^n,Big(H_n-,2H_2n+H_4nBig)=frac3pi16-fracpi4sqrt2-fracln28.tag3$$
Generalizing, we can consider two families of series:
$$mathcal A_m=sum_n=1^infty,(-1)^n,Big(H_n-,2H_2n+H_4nBig)^m,tag4$$
$$mathcal B_m=sum_n=1^inftyBig(H_n-,2H_2n+H_4nBig)^m,tag5$$
and try to evaluate them in a closed form.
So far I have the following conjectured result:
$$largesum_n=1^inftyBig(H_n-,2H_2n+H_4nBig)^2stackrelnormalsizecolorgray?=fracpi8-fracpi16,ln2-fracpi^296+frac316,ln^22-fracG4,tag$diamond$$$
where $G$ is the Catalan constant.
Could you please help me to prove this result and, possibly, find other values of $mathcal A_m,mathcal B_m$?
calculus sequences-and-series closed-form conjectures harmonic-numbers
$endgroup$
There is a known asymptotic expansion of harmonic numbers $H_n$ for $ntoinfty$:
$$beginalignH_n&=gamma+ln n+sum_k=1^inftyleft(-fracB_kkcdot n^kright)\
&=gamma+ln n+frac12n-frac112n^2+frac1120n^4-frac1252n^6,+,dots,endaligntag1$$
where $B_k$ are Bernoulli numbers.
We can take a linear combination of harmonic numbers to cancel constant and logarithmic terms, compensate for $O(n^-1)$ term, and get the following series that is possible to evaluate in a closed form (e.g. using generating function):
$$sum_k=1^inftyleft(H_n-,2H_2n+H_4n-frac18nright)=frac18-fracpi16.tag2$$
Rather than compensating for $O(n^-1)$ term, we can take a series with alternating signs, that is also possible to evaluate in a closed form:
$$sum_n=1^infty,(-1)^n,Big(H_n-,2H_2n+H_4nBig)=frac3pi16-fracpi4sqrt2-fracln28.tag3$$
Generalizing, we can consider two families of series:
$$mathcal A_m=sum_n=1^infty,(-1)^n,Big(H_n-,2H_2n+H_4nBig)^m,tag4$$
$$mathcal B_m=sum_n=1^inftyBig(H_n-,2H_2n+H_4nBig)^m,tag5$$
and try to evaluate them in a closed form.
So far I have the following conjectured result:
$$largesum_n=1^inftyBig(H_n-,2H_2n+H_4nBig)^2stackrelnormalsizecolorgray?=fracpi8-fracpi16,ln2-fracpi^296+frac316,ln^22-fracG4,tag$diamond$$$
where $G$ is the Catalan constant.
Could you please help me to prove this result and, possibly, find other values of $mathcal A_m,mathcal B_m$?
calculus sequences-and-series closed-form conjectures harmonic-numbers
calculus sequences-and-series closed-form conjectures harmonic-numbers
asked Aug 22 '15 at 17:46
Vladimir ReshetnikovVladimir Reshetnikov
24.5k5120235
24.5k5120235
This question has an open bounty worth +50
reputation from TheSimpliFire ending ending at 2019-03-17 09:11:13Z">in 4 days.
One or more of the answers is exemplary and worthy of an additional bounty.
This question has an open bounty worth +50
reputation from TheSimpliFire ending ending at 2019-03-17 09:11:13Z">in 4 days.
One or more of the answers is exemplary and worthy of an additional bounty.
$begingroup$
Some possibly related questions are linked from here.
$endgroup$
– Piotr Shatalin
Aug 22 '15 at 18:18
3
$begingroup$
BTW, another similar series can be evaluated using generating functions: $sum_n=1^inftyfrac1nleft(H_n-,2H_2n+H_4nright)=frac34 ln^2 2-fracpi^248.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 19:07
2
$begingroup$
And another: $sum_n=1^inftyfrac(-1)^nnleft(H_n-,2H_2n+H_4nright)=frac12 ln^2!left(1+sqrt2right)+frac18 ln^2 2-frac5pi^296.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 19:43
$begingroup$
One way to tackle it, at least the case $m=2$, would be to write $H_n=int_0^1frac1-x^n1-xdx$ and to convert the series into a double integral which probably could be evaluated more easily. But I think this is what we call "Brute Force".
$endgroup$
– Redundant Aunt
Aug 22 '15 at 19:56
4
$begingroup$
One more simple series where $O(n^-1)$ term is cancelled using harmonic numbers only: $sum_n=1^inftyleft(H_n-4H_2n+5H_4n-2H_8nright)=fracpi4sqrt2 - frac3pi16.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 22:14
add a comment |
$begingroup$
Some possibly related questions are linked from here.
$endgroup$
– Piotr Shatalin
Aug 22 '15 at 18:18
3
$begingroup$
BTW, another similar series can be evaluated using generating functions: $sum_n=1^inftyfrac1nleft(H_n-,2H_2n+H_4nright)=frac34 ln^2 2-fracpi^248.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 19:07
2
$begingroup$
And another: $sum_n=1^inftyfrac(-1)^nnleft(H_n-,2H_2n+H_4nright)=frac12 ln^2!left(1+sqrt2right)+frac18 ln^2 2-frac5pi^296.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 19:43
$begingroup$
One way to tackle it, at least the case $m=2$, would be to write $H_n=int_0^1frac1-x^n1-xdx$ and to convert the series into a double integral which probably could be evaluated more easily. But I think this is what we call "Brute Force".
$endgroup$
– Redundant Aunt
Aug 22 '15 at 19:56
4
$begingroup$
One more simple series where $O(n^-1)$ term is cancelled using harmonic numbers only: $sum_n=1^inftyleft(H_n-4H_2n+5H_4n-2H_8nright)=fracpi4sqrt2 - frac3pi16.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 22:14
$begingroup$
Some possibly related questions are linked from here.
$endgroup$
– Piotr Shatalin
Aug 22 '15 at 18:18
$begingroup$
Some possibly related questions are linked from here.
$endgroup$
– Piotr Shatalin
Aug 22 '15 at 18:18
3
3
$begingroup$
BTW, another similar series can be evaluated using generating functions: $sum_n=1^inftyfrac1nleft(H_n-,2H_2n+H_4nright)=frac34 ln^2 2-fracpi^248.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 19:07
$begingroup$
BTW, another similar series can be evaluated using generating functions: $sum_n=1^inftyfrac1nleft(H_n-,2H_2n+H_4nright)=frac34 ln^2 2-fracpi^248.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 19:07
2
2
$begingroup$
And another: $sum_n=1^inftyfrac(-1)^nnleft(H_n-,2H_2n+H_4nright)=frac12 ln^2!left(1+sqrt2right)+frac18 ln^2 2-frac5pi^296.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 19:43
$begingroup$
And another: $sum_n=1^inftyfrac(-1)^nnleft(H_n-,2H_2n+H_4nright)=frac12 ln^2!left(1+sqrt2right)+frac18 ln^2 2-frac5pi^296.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 19:43
$begingroup$
One way to tackle it, at least the case $m=2$, would be to write $H_n=int_0^1frac1-x^n1-xdx$ and to convert the series into a double integral which probably could be evaluated more easily. But I think this is what we call "Brute Force".
$endgroup$
– Redundant Aunt
Aug 22 '15 at 19:56
$begingroup$
One way to tackle it, at least the case $m=2$, would be to write $H_n=int_0^1frac1-x^n1-xdx$ and to convert the series into a double integral which probably could be evaluated more easily. But I think this is what we call "Brute Force".
$endgroup$
– Redundant Aunt
Aug 22 '15 at 19:56
4
4
$begingroup$
One more simple series where $O(n^-1)$ term is cancelled using harmonic numbers only: $sum_n=1^inftyleft(H_n-4H_2n+5H_4n-2H_8nright)=fracpi4sqrt2 - frac3pi16.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 22:14
$begingroup$
One more simple series where $O(n^-1)$ term is cancelled using harmonic numbers only: $sum_n=1^inftyleft(H_n-4H_2n+5H_4n-2H_8nright)=fracpi4sqrt2 - frac3pi16.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 22:14
add a comment |
2 Answers
2
active
oldest
votes
$begingroup$
So basically, I'll evaluate a bunch of integrals, trying to avoid polylogs as much as possible.
First thing is to notice that $displaystyle H_n-2H_2n+H_4n=int_0^1 fracx^2n-x^4n1+xdx$.
I noticed that $H_n-2H_2n+H_4n=H_4n-H_2n-(H_2n-H_n)=H_4n^--H_2n^-$, where $H_n^-=sum_k=1^n frac(-1)^k+1k$ is called
a skew harmonic number (at least by Khristo N. Boyadzhiev. link.) Knowing they have a simple intergal representation I found the above. My answer is influenced by Boyadzhiev's work.
If I make any unexplainable substitution, it's most likely $t=frac1-x1+x$.
Also, I'm not very good with Latex, so alignment should be awful. Hopefully there are no typos.
Below, easy enough to prove, is what I take for granted:
$ -lnsin x=ln2+sum_n=1^infty fraccos(2nx)n ,-lncos x=ln2+sum_n=1^infty frac(-1)^ncos(2nx)n tag1$
$$ int_0^fracpi2 cos x cos(nx)dx=begincases fracpi4 &n=1\0 &n ,,textodd\ frac(-1)^1+n/2n^2-1 &n ,,texteven endcases tag2$$
$$ int_0^1 fracln(1-x)a+xdx=-operatornameLi_2left(frac1a+1right)tag3$$
Starting,
$$sum_n=0^infty(H_n-2H_2n+H_4n)^2=sum_n=0^inftyint_0^1int_0^1frac(x^2n-x^4n)(u^2n-u^4n)(1+x)(1+u)dxdu
\=smallint_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^2)-2int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^4)+int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^4u^4)
\=I_22-2I_24+I_44$$
Computing $I_22$.
Substitute $u=fracyx$ ,change the order of integration, evaluate the inner integral, and substitue $t=frac1-x1+x$ to get
$$beginalign I_22=int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^2)=int_0^1int_0^xfracdydx(1+x)(x+y)(1-y^2)
\=int_0^1 frac11-y^2int_1^y fracdx(1+x)(x+y) dy=int_0^1 fraclnleft(frac(1+x)^24xright)(1+x)(1-x^2),dx
\=frac-14int_0^1 frac(1+t)t^2ln(1-t^2)dt=-frac14int_0^1fracln(1-t^2)t^2dt-frac14int_0^1fracln(1-t^2)tdt
\=frac14sum_n=0^infty frac1(n+1)(2n+1)+frac14sum_n=0^infty frac1(n+1)(2n+2)=fracln22+fracpi^248.endalign$$
Computing $I_44$.
Start the same as with $I_22$ to get $displaystyle I_44=int_0^1 fraclnleft(frac(1+x)^24xright)(1-x)(1-x^4),dx=frac-18int_0^1 fracln(1-t^2)t^2(1+t^2)(1+t)^3dt$.
We can calculate these integrals:
$$beginalign int_0^1 fracln(1-x^2)1+x^2dx=int_0^1 fracln(1+x)1+x^2dx+int_0^1 fracln(1-x)1+x^2dx tag4
\=int_0^1 fracln(1+x)1+x^2dx +int_0^1 fraclnleft(frac2t1+tright)1+t^2dt
\=fracpi4ln2+sum_n=0^infty (-1)^nint_0^1ln(t) t^2ndt=fracpi4ln2-G. endalign$$
$$beginalign int_0^1 fracln(1-x^2)x^2(1+x^2)dx=int_0^1 fracln(1-x^2)x^2dx-int_0^1 fracln(1-x^2)1+x^2dx tag5
\=-sum_n=0^infty frac1n+1int_0^1 x^2ndx-fracpi4ln2+G=G-fracpi4ln2-2ln2.endalign$$
$$beginalign int_0^1 fracxln(1-x^2)1+x^2dx=frac12int_0^1 fracln(1-x)1+xdx tag6
\=-frac12 operatornameLi_2left(frac12right)=fracln^2 24-fracpi^224.endalign$$
$$beginalign int_0^1 fracln(1-x^2)x(1+x^2)dx=int_0^1 fracln(1-x^2)xdx-int_0^1 fracxln(1-x^2)1+x^2dx tag7
\=-sum_n=0^inftyfrac1n+1int_0^1 x^2n+1dx-fracln^2 24+fracpi^224=-fracpi^224-fracln^2 24.endalign$$
Altogether,
$$I_44=frac-18int_0^1 fracln(1-x^2)x^2(1+x^2)(1+3x+3x^2+x^3)dx
\=-fracpi16ln2+fracln24+fracln^2 216+fracpi^248+fracG4.$$
Computing $I_24$.
Substitute $u=fracyx^2$, change the order of integration, let $yto y^2$, evaluate the inner integral,and substitue $t=frac1-x1+x$:
$$beginalign* I_24=int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^4u^2)=int_0^1int_0^x^2 fracdydx(1+x)(x^2+y)(1-y^2)
\=int_0^1 frac11-y^2int_sqrty^1fracdx(1+x)(x^2+y)dy=2int_0^1fracy1-y^4int_y^1fracdx(1+x)(x^2+y^2)dy
\=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1-x^4)dx-2int_0^1fracxlnleft(frac(1+x)sqrt1+x^22sqrt2xright)(1+x^2)(1-x^4)dx
=I_241-I_242. endalign*$$
Evaulation of $I_241$.
Substitute $t=frac1-x1+x$ to get $displaystyle I_241=frac14int_0^1 fractan^-1(t)t(1+t^2)^2(1+t)^4dt$.
We can calculate these integrals.In the following, let $x=tantheta$:
$$beginalign int_0^1 fractan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetacos^2(theta)dtheta=fracpi^264+fracpi16-frac18.tag8endalign$$
$$beginalign int_0^1 fracxtan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetatanthetacos^2thetadtheta=frac12int_0^fracpi4thetasin(2theta)dtheta=frac18.tag9endalign$$
$$beginalign int_0^1 fracx^2tan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetasin^2(theta)dtheta=fracpi^264-fracpi16+frac18.tag10endalign$$
$$beginalign int_0^1 fracx^3tan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetasin^3(theta)sectheta,dtheta tag11
\=int_0^fracpi4thetatantheta ,dtheta-int_0^fracpi4thetasinthetacostheta ,dtheta=fracpi8ln2-frac18+int_0^fracpi4lncostheta ,dtheta
\=fracpi8ln2-frac18-int_0^fracpi4ln2 ,dtheta-sum_n=1^infty frac(-1)^nnint_0^fracpi4cos(2ntheta),dtheta
\=-fracpi8ln2-frac18+frac12sum_n=1^inftyfrac(-1)^n+1n^2sin(fracpi n2)=fracG2-fracpi8ln2-frac18.endalign$$
$$beginalign int_0^1 fractan^-1(x)x(1+x^2)^2dx=int_0^fracpi4thetacos^3(theta)csctheta,dtheta tag12
\=int_0^fracpi4thetacottheta ,dtheta-int_0^fracpi4thetasinthetacostheta ,dtheta=-frac18-fracpi8ln2-int_0^fracpi4lnsintheta ,dtheta
\=-frac18-fracpi8ln2+int_0^fracpi4ln2 ,dtheta+sum_n=1^inftyfrac1nint_0^fracpi4cos(2ntheta),dtheta
\=-frac18+fracpi8ln2+frac12sum_n=1^infty fracsin(fracpi n2)n^2=fracG2+fracpi8ln2-frac18.endalign$$
Altogether,
$$I_241=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1-x^4)dx= frac14int_0^1 fractan^-1(x)x(1+x^2)^2(1+4x+6x^2+4x^3+x^4)dx
\=fracpi^232+frac18+fracG4$$
Evaulation of $I_242$.
Substitute $t=frac1-x1+x$ to get
$$ I_242=frac14int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)t(1+t^2)^2(1-t^2)(1+t)^2dt
\=frac12int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)(1+t^2)^2(1-t^2)dt+frac14int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)t(1+t^2)(1-t^2)dt
\=frac12int_0^1 fraclnleft(frac1+x^21-x^2right)(1+x^2)^2(1-x^2)dx-frac14int_0^1fracln(1+x^2)(1+x^2)^2(1-x^2)dx\+frac18int_0^1fracln(1+x^2)(1-x^2)x(1+x^2)dx-frac14int_0^1fracln(1-x^2)(1-x^2)x(1+x^2)dx
$$
Calculating these integrals:
$$beginalign int_0^1 fraclnleft(frac1+x^21-x^2right)(1+x^2)^2(1-x^2)dx=-frac12int_0^1 fraclnleft(frac1-x1+xright)sqrtx(1+x)frac1-x1+xdx tag13
\=-frac12int_0^1fractln tsqrt1-t^2dt=-frac18int_0^1fracln tsqrt1-tdt=-frac18int_0^1 t^-1/2ln(1-t)dt
\=frac14sum_n=0^infty frac1(n+1)(2n+3)=frac12-fracln22.endalign$$
$$beginalign int_0^1fracln(1+x^2)(1-x^2)x(1+x^2)dx=frac12int_0^1fracln(1+x)(1-x)x(1+x)dx tag14
\=frac12int_0^1fracln(1+x)xdx-int_0^1fracln(1+x)1+xdx
\=frac12sum_n=0^inftyfrac(-1)^n+1n+1int_0^1 x^n dx-frac12ln^2(1+x)bigg_0^1=fracpi^224-fracln^2 22.endalign$$
$$beginalign int_0^1fracln(1-x^2)(1-x^2)x(1+x^2)dx=frac12int_0^1fracln(1-x)xdx-int_0^1fracln(1-x)1+xdx tag15
\=-fracpi^212-left(fracln^2 22-fracpi^212right)=-fracln^2 22.endalign$$
$$
int_0^1fracln(1+x^2)(1+x^2)^2(1-x^2)dx=-2int_0^fracpi4cos^2(theta)(1-tan^2theta)lncostheta,dtheta tag16
\=-2int_0^fracpi4cos(2theta)lncostheta,dtheta=2ln2int_0^fracpi4cos(2theta)dtheta+sum_n=1^infty frac(-1)^nnint_0^fracpi2costheta cos(ntheta)dtheta
\=ln2-fracpi4+sum_n=1^infty frac(-1)^2n2nfrac(-1)^n+1(2n)^2-1=fracln22-fracpi4+frac12.$$
Altogether, $displaystyle I_242=fracpi^2192+fracpi16+fracln^2 216-frac3ln28+frac18$,
leading to $displaystyle I_24=I_241-I_242=frac5pi^2192-fracpi16-fracln^2 216+frac3ln28+fracG4$,
and finally, confirming the conjecture,
$$sum_n=0^infty(H_n-2H_2n+H_4n)^2=I_22-2I_24+I_44=fracpi8-fracpi16ln2-fracpi^296+frac3ln^2 216-fracG4.$$
I don't know about higher powers. I guess the case $mathcal A_2$ can also be done. If we start the same as with $mathcal B_2$, writing $mathcal A_2=J_22-2J_24+J_44$
we can find that $displaystyle J_22=int_0^1int_0^1fracdxdu(1+x)(1+u)(1+x^2u^2)=2I_44-I_22=-fracpi8ln2+fracG2+fracpi^248+fracln^2 28$
$J_44$ can be reduced to $displaystyle =-frac12int_0^1 fracln(1-x^2)x(x^4+6x^2+1)(1+x)^3 dx$.
already this i can't evaulate fully, as polylogs are inescapable. Factorizing $x^2+6x+1=(x+3+2sqrt2)(x+3-2sqrt2).$
I can get $displaystyle int_0^1 fracln(1-x^2)x(x^4+6x^2+1)dx=-fracpi^212+frac4-3sqrt216operatornameLi_2left(frac2-sqrt24right)+frac4+3sqrt216operatornameLi_2left(frac2+sqrt24right)$
but nothing more.
Edit 1.
After some more work and a fair amount of cancellation, we obtain
$$int_0^1 fracln(1-x^2)x(x^4+6x^2+1)(1+x)^3 dx=frac1+2sqrt24piln2-fracpi^224-frac14lnleft(frac2+sqrt24right)lnleft(frac2-sqrt24right)
-fracsqrt2+12ImoperatornameLi_2left(frac2+sqrt22+fracisqrt22right)-fracsqrt2-12ImoperatornameLi_2left(frac2-sqrt22+fracisqrt22right)$$
I obtained it by calculating $displaystyle int_0^1 fracln(1+x)x+adx=ln2lnleft(fraca+1a-1right)+operatornameLi_2left(frac21-aright)-operatornameLi_2left(frac11-aright)$,
which together with $(3)$ can be used to give a closed form for $displaystyle int_0^1fracln(1-x^2)x+adx$, which in turn, through partial fractions, can be used to give a closed form for $displaystyle int_0^1fracln(1-x^2)x^2+a^2dx$.
Fortunately, things didn't get too ugly as both $3+2sqrt2$ and $3-2sqrt2$ have nice square roots. I will fill in details as soon as I can.
Now we just need to evaluate $J_24$. Starting similarly as with $I_24$,
we have:
$$J_24=int_0^1int_0^1fracdxdu(1+x)(1+u)(1+x^4u^2)
\=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1+x^4)dx-2int_0^1fracxlnleft(frac(1+x)sqrt1+x^22sqrt2xright)(1+x^2)(1+x^4)dx
\=J_241-J_242$$
Through $t=frac1-x1+x$, $J_241$ turns to $displaystyle int_0^1 fractan^-1(x)(1+x^2)(x^4+6x^2+1)(1+x)^4,dx$. I don't have any idea about that yet. Edit 1.
$endgroup$
add a comment |
$begingroup$
Just my thoughts for now: I would try to exploit Parseval's identity. For first, we have:
$$ H_n-2H_2n+H_4n =int_0^1frac-x^n+2x^2n-x^4n1-x,dx tag1 $$
and:
$$ sum_ngeq 1frac-x^n+2x^2n-x^4n1-xe^niy = frac11-xleft(frac-11-e^iyx+frac21-e^iyx^2+frac-11-e^iyx^4right).tag2$$
The poles of the RHS (as a function of $x$) are located at $xinlefte^-iy,pm e^-iy/2,pm e^-iy/4,pm i e^-iy/4right$.
By using the residue theorem we may compute an explicit representation for:
$$ g(y) = sum_ngeq 1left(H_n-2H_2n+H_4nright)e^niy,tag3 $$
then Parseval's theorem gives:
$$ sum_ngeq 1left(H_n-2H_2n+H_4nright)^2 = frac12piint_-pi^pig(y)g(-y),dy tag4$$
and the resulting integral should be not to difficult to evaluate in terms of dilogarithms.
Another chance may be to apply summation by parts (like I did in this question), but it looks lengthy.
$endgroup$
$begingroup$
It’s a very nice idea, but I think you are underestimating the difficulty of evaluating the integral. I checked the representation for $g$ using software: the result is a ton of terms of rational functions of trig’s and inverse trig’s and that is before computing $g(y)g(-y)$. This approach might work, but it looks like a lot of work.
$endgroup$
– Winther
Sep 1 '15 at 15:26
$begingroup$
@Winther: I know, it is a tough nut to crack with bare hands, but with some human-guided simplifications, it just boils down to computing some dilogarithmic integrals related with the fourth roots of unity. I haven't really delved into summation by parts, yet. It looks promising, maybe it is the simple way.
$endgroup$
– Jack D'Aurizio
Sep 1 '15 at 15:35
add a comment |
Your Answer
StackExchange.ifUsing("editor", function ()
return StackExchange.using("mathjaxEditing", function ()
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix)
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
);
);
, "mathjax-editing");
StackExchange.ready(function()
var channelOptions =
tags: "".split(" "),
id: "69"
;
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function()
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled)
StackExchange.using("snippets", function()
createEditor();
);
else
createEditor();
);
function createEditor()
StackExchange.prepareEditor(
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader:
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
,
noCode: true, onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
);
);
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f1406122%2fconjectured-value-of-a-harmonic-sum-sum-n-1-infty-lefth-n-2h-2nh-4n%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
2 Answers
2
active
oldest
votes
2 Answers
2
active
oldest
votes
active
oldest
votes
active
oldest
votes
$begingroup$
So basically, I'll evaluate a bunch of integrals, trying to avoid polylogs as much as possible.
First thing is to notice that $displaystyle H_n-2H_2n+H_4n=int_0^1 fracx^2n-x^4n1+xdx$.
I noticed that $H_n-2H_2n+H_4n=H_4n-H_2n-(H_2n-H_n)=H_4n^--H_2n^-$, where $H_n^-=sum_k=1^n frac(-1)^k+1k$ is called
a skew harmonic number (at least by Khristo N. Boyadzhiev. link.) Knowing they have a simple intergal representation I found the above. My answer is influenced by Boyadzhiev's work.
If I make any unexplainable substitution, it's most likely $t=frac1-x1+x$.
Also, I'm not very good with Latex, so alignment should be awful. Hopefully there are no typos.
Below, easy enough to prove, is what I take for granted:
$ -lnsin x=ln2+sum_n=1^infty fraccos(2nx)n ,-lncos x=ln2+sum_n=1^infty frac(-1)^ncos(2nx)n tag1$
$$ int_0^fracpi2 cos x cos(nx)dx=begincases fracpi4 &n=1\0 &n ,,textodd\ frac(-1)^1+n/2n^2-1 &n ,,texteven endcases tag2$$
$$ int_0^1 fracln(1-x)a+xdx=-operatornameLi_2left(frac1a+1right)tag3$$
Starting,
$$sum_n=0^infty(H_n-2H_2n+H_4n)^2=sum_n=0^inftyint_0^1int_0^1frac(x^2n-x^4n)(u^2n-u^4n)(1+x)(1+u)dxdu
\=smallint_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^2)-2int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^4)+int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^4u^4)
\=I_22-2I_24+I_44$$
Computing $I_22$.
Substitute $u=fracyx$ ,change the order of integration, evaluate the inner integral, and substitue $t=frac1-x1+x$ to get
$$beginalign I_22=int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^2)=int_0^1int_0^xfracdydx(1+x)(x+y)(1-y^2)
\=int_0^1 frac11-y^2int_1^y fracdx(1+x)(x+y) dy=int_0^1 fraclnleft(frac(1+x)^24xright)(1+x)(1-x^2),dx
\=frac-14int_0^1 frac(1+t)t^2ln(1-t^2)dt=-frac14int_0^1fracln(1-t^2)t^2dt-frac14int_0^1fracln(1-t^2)tdt
\=frac14sum_n=0^infty frac1(n+1)(2n+1)+frac14sum_n=0^infty frac1(n+1)(2n+2)=fracln22+fracpi^248.endalign$$
Computing $I_44$.
Start the same as with $I_22$ to get $displaystyle I_44=int_0^1 fraclnleft(frac(1+x)^24xright)(1-x)(1-x^4),dx=frac-18int_0^1 fracln(1-t^2)t^2(1+t^2)(1+t)^3dt$.
We can calculate these integrals:
$$beginalign int_0^1 fracln(1-x^2)1+x^2dx=int_0^1 fracln(1+x)1+x^2dx+int_0^1 fracln(1-x)1+x^2dx tag4
\=int_0^1 fracln(1+x)1+x^2dx +int_0^1 fraclnleft(frac2t1+tright)1+t^2dt
\=fracpi4ln2+sum_n=0^infty (-1)^nint_0^1ln(t) t^2ndt=fracpi4ln2-G. endalign$$
$$beginalign int_0^1 fracln(1-x^2)x^2(1+x^2)dx=int_0^1 fracln(1-x^2)x^2dx-int_0^1 fracln(1-x^2)1+x^2dx tag5
\=-sum_n=0^infty frac1n+1int_0^1 x^2ndx-fracpi4ln2+G=G-fracpi4ln2-2ln2.endalign$$
$$beginalign int_0^1 fracxln(1-x^2)1+x^2dx=frac12int_0^1 fracln(1-x)1+xdx tag6
\=-frac12 operatornameLi_2left(frac12right)=fracln^2 24-fracpi^224.endalign$$
$$beginalign int_0^1 fracln(1-x^2)x(1+x^2)dx=int_0^1 fracln(1-x^2)xdx-int_0^1 fracxln(1-x^2)1+x^2dx tag7
\=-sum_n=0^inftyfrac1n+1int_0^1 x^2n+1dx-fracln^2 24+fracpi^224=-fracpi^224-fracln^2 24.endalign$$
Altogether,
$$I_44=frac-18int_0^1 fracln(1-x^2)x^2(1+x^2)(1+3x+3x^2+x^3)dx
\=-fracpi16ln2+fracln24+fracln^2 216+fracpi^248+fracG4.$$
Computing $I_24$.
Substitute $u=fracyx^2$, change the order of integration, let $yto y^2$, evaluate the inner integral,and substitue $t=frac1-x1+x$:
$$beginalign* I_24=int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^4u^2)=int_0^1int_0^x^2 fracdydx(1+x)(x^2+y)(1-y^2)
\=int_0^1 frac11-y^2int_sqrty^1fracdx(1+x)(x^2+y)dy=2int_0^1fracy1-y^4int_y^1fracdx(1+x)(x^2+y^2)dy
\=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1-x^4)dx-2int_0^1fracxlnleft(frac(1+x)sqrt1+x^22sqrt2xright)(1+x^2)(1-x^4)dx
=I_241-I_242. endalign*$$
Evaulation of $I_241$.
Substitute $t=frac1-x1+x$ to get $displaystyle I_241=frac14int_0^1 fractan^-1(t)t(1+t^2)^2(1+t)^4dt$.
We can calculate these integrals.In the following, let $x=tantheta$:
$$beginalign int_0^1 fractan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetacos^2(theta)dtheta=fracpi^264+fracpi16-frac18.tag8endalign$$
$$beginalign int_0^1 fracxtan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetatanthetacos^2thetadtheta=frac12int_0^fracpi4thetasin(2theta)dtheta=frac18.tag9endalign$$
$$beginalign int_0^1 fracx^2tan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetasin^2(theta)dtheta=fracpi^264-fracpi16+frac18.tag10endalign$$
$$beginalign int_0^1 fracx^3tan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetasin^3(theta)sectheta,dtheta tag11
\=int_0^fracpi4thetatantheta ,dtheta-int_0^fracpi4thetasinthetacostheta ,dtheta=fracpi8ln2-frac18+int_0^fracpi4lncostheta ,dtheta
\=fracpi8ln2-frac18-int_0^fracpi4ln2 ,dtheta-sum_n=1^infty frac(-1)^nnint_0^fracpi4cos(2ntheta),dtheta
\=-fracpi8ln2-frac18+frac12sum_n=1^inftyfrac(-1)^n+1n^2sin(fracpi n2)=fracG2-fracpi8ln2-frac18.endalign$$
$$beginalign int_0^1 fractan^-1(x)x(1+x^2)^2dx=int_0^fracpi4thetacos^3(theta)csctheta,dtheta tag12
\=int_0^fracpi4thetacottheta ,dtheta-int_0^fracpi4thetasinthetacostheta ,dtheta=-frac18-fracpi8ln2-int_0^fracpi4lnsintheta ,dtheta
\=-frac18-fracpi8ln2+int_0^fracpi4ln2 ,dtheta+sum_n=1^inftyfrac1nint_0^fracpi4cos(2ntheta),dtheta
\=-frac18+fracpi8ln2+frac12sum_n=1^infty fracsin(fracpi n2)n^2=fracG2+fracpi8ln2-frac18.endalign$$
Altogether,
$$I_241=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1-x^4)dx= frac14int_0^1 fractan^-1(x)x(1+x^2)^2(1+4x+6x^2+4x^3+x^4)dx
\=fracpi^232+frac18+fracG4$$
Evaulation of $I_242$.
Substitute $t=frac1-x1+x$ to get
$$ I_242=frac14int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)t(1+t^2)^2(1-t^2)(1+t)^2dt
\=frac12int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)(1+t^2)^2(1-t^2)dt+frac14int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)t(1+t^2)(1-t^2)dt
\=frac12int_0^1 fraclnleft(frac1+x^21-x^2right)(1+x^2)^2(1-x^2)dx-frac14int_0^1fracln(1+x^2)(1+x^2)^2(1-x^2)dx\+frac18int_0^1fracln(1+x^2)(1-x^2)x(1+x^2)dx-frac14int_0^1fracln(1-x^2)(1-x^2)x(1+x^2)dx
$$
Calculating these integrals:
$$beginalign int_0^1 fraclnleft(frac1+x^21-x^2right)(1+x^2)^2(1-x^2)dx=-frac12int_0^1 fraclnleft(frac1-x1+xright)sqrtx(1+x)frac1-x1+xdx tag13
\=-frac12int_0^1fractln tsqrt1-t^2dt=-frac18int_0^1fracln tsqrt1-tdt=-frac18int_0^1 t^-1/2ln(1-t)dt
\=frac14sum_n=0^infty frac1(n+1)(2n+3)=frac12-fracln22.endalign$$
$$beginalign int_0^1fracln(1+x^2)(1-x^2)x(1+x^2)dx=frac12int_0^1fracln(1+x)(1-x)x(1+x)dx tag14
\=frac12int_0^1fracln(1+x)xdx-int_0^1fracln(1+x)1+xdx
\=frac12sum_n=0^inftyfrac(-1)^n+1n+1int_0^1 x^n dx-frac12ln^2(1+x)bigg_0^1=fracpi^224-fracln^2 22.endalign$$
$$beginalign int_0^1fracln(1-x^2)(1-x^2)x(1+x^2)dx=frac12int_0^1fracln(1-x)xdx-int_0^1fracln(1-x)1+xdx tag15
\=-fracpi^212-left(fracln^2 22-fracpi^212right)=-fracln^2 22.endalign$$
$$
int_0^1fracln(1+x^2)(1+x^2)^2(1-x^2)dx=-2int_0^fracpi4cos^2(theta)(1-tan^2theta)lncostheta,dtheta tag16
\=-2int_0^fracpi4cos(2theta)lncostheta,dtheta=2ln2int_0^fracpi4cos(2theta)dtheta+sum_n=1^infty frac(-1)^nnint_0^fracpi2costheta cos(ntheta)dtheta
\=ln2-fracpi4+sum_n=1^infty frac(-1)^2n2nfrac(-1)^n+1(2n)^2-1=fracln22-fracpi4+frac12.$$
Altogether, $displaystyle I_242=fracpi^2192+fracpi16+fracln^2 216-frac3ln28+frac18$,
leading to $displaystyle I_24=I_241-I_242=frac5pi^2192-fracpi16-fracln^2 216+frac3ln28+fracG4$,
and finally, confirming the conjecture,
$$sum_n=0^infty(H_n-2H_2n+H_4n)^2=I_22-2I_24+I_44=fracpi8-fracpi16ln2-fracpi^296+frac3ln^2 216-fracG4.$$
I don't know about higher powers. I guess the case $mathcal A_2$ can also be done. If we start the same as with $mathcal B_2$, writing $mathcal A_2=J_22-2J_24+J_44$
we can find that $displaystyle J_22=int_0^1int_0^1fracdxdu(1+x)(1+u)(1+x^2u^2)=2I_44-I_22=-fracpi8ln2+fracG2+fracpi^248+fracln^2 28$
$J_44$ can be reduced to $displaystyle =-frac12int_0^1 fracln(1-x^2)x(x^4+6x^2+1)(1+x)^3 dx$.
already this i can't evaulate fully, as polylogs are inescapable. Factorizing $x^2+6x+1=(x+3+2sqrt2)(x+3-2sqrt2).$
I can get $displaystyle int_0^1 fracln(1-x^2)x(x^4+6x^2+1)dx=-fracpi^212+frac4-3sqrt216operatornameLi_2left(frac2-sqrt24right)+frac4+3sqrt216operatornameLi_2left(frac2+sqrt24right)$
but nothing more.
Edit 1.
After some more work and a fair amount of cancellation, we obtain
$$int_0^1 fracln(1-x^2)x(x^4+6x^2+1)(1+x)^3 dx=frac1+2sqrt24piln2-fracpi^224-frac14lnleft(frac2+sqrt24right)lnleft(frac2-sqrt24right)
-fracsqrt2+12ImoperatornameLi_2left(frac2+sqrt22+fracisqrt22right)-fracsqrt2-12ImoperatornameLi_2left(frac2-sqrt22+fracisqrt22right)$$
I obtained it by calculating $displaystyle int_0^1 fracln(1+x)x+adx=ln2lnleft(fraca+1a-1right)+operatornameLi_2left(frac21-aright)-operatornameLi_2left(frac11-aright)$,
which together with $(3)$ can be used to give a closed form for $displaystyle int_0^1fracln(1-x^2)x+adx$, which in turn, through partial fractions, can be used to give a closed form for $displaystyle int_0^1fracln(1-x^2)x^2+a^2dx$.
Fortunately, things didn't get too ugly as both $3+2sqrt2$ and $3-2sqrt2$ have nice square roots. I will fill in details as soon as I can.
Now we just need to evaluate $J_24$. Starting similarly as with $I_24$,
we have:
$$J_24=int_0^1int_0^1fracdxdu(1+x)(1+u)(1+x^4u^2)
\=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1+x^4)dx-2int_0^1fracxlnleft(frac(1+x)sqrt1+x^22sqrt2xright)(1+x^2)(1+x^4)dx
\=J_241-J_242$$
Through $t=frac1-x1+x$, $J_241$ turns to $displaystyle int_0^1 fractan^-1(x)(1+x^2)(x^4+6x^2+1)(1+x)^4,dx$. I don't have any idea about that yet. Edit 1.
$endgroup$
add a comment |
$begingroup$
So basically, I'll evaluate a bunch of integrals, trying to avoid polylogs as much as possible.
First thing is to notice that $displaystyle H_n-2H_2n+H_4n=int_0^1 fracx^2n-x^4n1+xdx$.
I noticed that $H_n-2H_2n+H_4n=H_4n-H_2n-(H_2n-H_n)=H_4n^--H_2n^-$, where $H_n^-=sum_k=1^n frac(-1)^k+1k$ is called
a skew harmonic number (at least by Khristo N. Boyadzhiev. link.) Knowing they have a simple intergal representation I found the above. My answer is influenced by Boyadzhiev's work.
If I make any unexplainable substitution, it's most likely $t=frac1-x1+x$.
Also, I'm not very good with Latex, so alignment should be awful. Hopefully there are no typos.
Below, easy enough to prove, is what I take for granted:
$ -lnsin x=ln2+sum_n=1^infty fraccos(2nx)n ,-lncos x=ln2+sum_n=1^infty frac(-1)^ncos(2nx)n tag1$
$$ int_0^fracpi2 cos x cos(nx)dx=begincases fracpi4 &n=1\0 &n ,,textodd\ frac(-1)^1+n/2n^2-1 &n ,,texteven endcases tag2$$
$$ int_0^1 fracln(1-x)a+xdx=-operatornameLi_2left(frac1a+1right)tag3$$
Starting,
$$sum_n=0^infty(H_n-2H_2n+H_4n)^2=sum_n=0^inftyint_0^1int_0^1frac(x^2n-x^4n)(u^2n-u^4n)(1+x)(1+u)dxdu
\=smallint_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^2)-2int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^4)+int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^4u^4)
\=I_22-2I_24+I_44$$
Computing $I_22$.
Substitute $u=fracyx$ ,change the order of integration, evaluate the inner integral, and substitue $t=frac1-x1+x$ to get
$$beginalign I_22=int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^2)=int_0^1int_0^xfracdydx(1+x)(x+y)(1-y^2)
\=int_0^1 frac11-y^2int_1^y fracdx(1+x)(x+y) dy=int_0^1 fraclnleft(frac(1+x)^24xright)(1+x)(1-x^2),dx
\=frac-14int_0^1 frac(1+t)t^2ln(1-t^2)dt=-frac14int_0^1fracln(1-t^2)t^2dt-frac14int_0^1fracln(1-t^2)tdt
\=frac14sum_n=0^infty frac1(n+1)(2n+1)+frac14sum_n=0^infty frac1(n+1)(2n+2)=fracln22+fracpi^248.endalign$$
Computing $I_44$.
Start the same as with $I_22$ to get $displaystyle I_44=int_0^1 fraclnleft(frac(1+x)^24xright)(1-x)(1-x^4),dx=frac-18int_0^1 fracln(1-t^2)t^2(1+t^2)(1+t)^3dt$.
We can calculate these integrals:
$$beginalign int_0^1 fracln(1-x^2)1+x^2dx=int_0^1 fracln(1+x)1+x^2dx+int_0^1 fracln(1-x)1+x^2dx tag4
\=int_0^1 fracln(1+x)1+x^2dx +int_0^1 fraclnleft(frac2t1+tright)1+t^2dt
\=fracpi4ln2+sum_n=0^infty (-1)^nint_0^1ln(t) t^2ndt=fracpi4ln2-G. endalign$$
$$beginalign int_0^1 fracln(1-x^2)x^2(1+x^2)dx=int_0^1 fracln(1-x^2)x^2dx-int_0^1 fracln(1-x^2)1+x^2dx tag5
\=-sum_n=0^infty frac1n+1int_0^1 x^2ndx-fracpi4ln2+G=G-fracpi4ln2-2ln2.endalign$$
$$beginalign int_0^1 fracxln(1-x^2)1+x^2dx=frac12int_0^1 fracln(1-x)1+xdx tag6
\=-frac12 operatornameLi_2left(frac12right)=fracln^2 24-fracpi^224.endalign$$
$$beginalign int_0^1 fracln(1-x^2)x(1+x^2)dx=int_0^1 fracln(1-x^2)xdx-int_0^1 fracxln(1-x^2)1+x^2dx tag7
\=-sum_n=0^inftyfrac1n+1int_0^1 x^2n+1dx-fracln^2 24+fracpi^224=-fracpi^224-fracln^2 24.endalign$$
Altogether,
$$I_44=frac-18int_0^1 fracln(1-x^2)x^2(1+x^2)(1+3x+3x^2+x^3)dx
\=-fracpi16ln2+fracln24+fracln^2 216+fracpi^248+fracG4.$$
Computing $I_24$.
Substitute $u=fracyx^2$, change the order of integration, let $yto y^2$, evaluate the inner integral,and substitue $t=frac1-x1+x$:
$$beginalign* I_24=int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^4u^2)=int_0^1int_0^x^2 fracdydx(1+x)(x^2+y)(1-y^2)
\=int_0^1 frac11-y^2int_sqrty^1fracdx(1+x)(x^2+y)dy=2int_0^1fracy1-y^4int_y^1fracdx(1+x)(x^2+y^2)dy
\=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1-x^4)dx-2int_0^1fracxlnleft(frac(1+x)sqrt1+x^22sqrt2xright)(1+x^2)(1-x^4)dx
=I_241-I_242. endalign*$$
Evaulation of $I_241$.
Substitute $t=frac1-x1+x$ to get $displaystyle I_241=frac14int_0^1 fractan^-1(t)t(1+t^2)^2(1+t)^4dt$.
We can calculate these integrals.In the following, let $x=tantheta$:
$$beginalign int_0^1 fractan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetacos^2(theta)dtheta=fracpi^264+fracpi16-frac18.tag8endalign$$
$$beginalign int_0^1 fracxtan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetatanthetacos^2thetadtheta=frac12int_0^fracpi4thetasin(2theta)dtheta=frac18.tag9endalign$$
$$beginalign int_0^1 fracx^2tan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetasin^2(theta)dtheta=fracpi^264-fracpi16+frac18.tag10endalign$$
$$beginalign int_0^1 fracx^3tan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetasin^3(theta)sectheta,dtheta tag11
\=int_0^fracpi4thetatantheta ,dtheta-int_0^fracpi4thetasinthetacostheta ,dtheta=fracpi8ln2-frac18+int_0^fracpi4lncostheta ,dtheta
\=fracpi8ln2-frac18-int_0^fracpi4ln2 ,dtheta-sum_n=1^infty frac(-1)^nnint_0^fracpi4cos(2ntheta),dtheta
\=-fracpi8ln2-frac18+frac12sum_n=1^inftyfrac(-1)^n+1n^2sin(fracpi n2)=fracG2-fracpi8ln2-frac18.endalign$$
$$beginalign int_0^1 fractan^-1(x)x(1+x^2)^2dx=int_0^fracpi4thetacos^3(theta)csctheta,dtheta tag12
\=int_0^fracpi4thetacottheta ,dtheta-int_0^fracpi4thetasinthetacostheta ,dtheta=-frac18-fracpi8ln2-int_0^fracpi4lnsintheta ,dtheta
\=-frac18-fracpi8ln2+int_0^fracpi4ln2 ,dtheta+sum_n=1^inftyfrac1nint_0^fracpi4cos(2ntheta),dtheta
\=-frac18+fracpi8ln2+frac12sum_n=1^infty fracsin(fracpi n2)n^2=fracG2+fracpi8ln2-frac18.endalign$$
Altogether,
$$I_241=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1-x^4)dx= frac14int_0^1 fractan^-1(x)x(1+x^2)^2(1+4x+6x^2+4x^3+x^4)dx
\=fracpi^232+frac18+fracG4$$
Evaulation of $I_242$.
Substitute $t=frac1-x1+x$ to get
$$ I_242=frac14int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)t(1+t^2)^2(1-t^2)(1+t)^2dt
\=frac12int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)(1+t^2)^2(1-t^2)dt+frac14int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)t(1+t^2)(1-t^2)dt
\=frac12int_0^1 fraclnleft(frac1+x^21-x^2right)(1+x^2)^2(1-x^2)dx-frac14int_0^1fracln(1+x^2)(1+x^2)^2(1-x^2)dx\+frac18int_0^1fracln(1+x^2)(1-x^2)x(1+x^2)dx-frac14int_0^1fracln(1-x^2)(1-x^2)x(1+x^2)dx
$$
Calculating these integrals:
$$beginalign int_0^1 fraclnleft(frac1+x^21-x^2right)(1+x^2)^2(1-x^2)dx=-frac12int_0^1 fraclnleft(frac1-x1+xright)sqrtx(1+x)frac1-x1+xdx tag13
\=-frac12int_0^1fractln tsqrt1-t^2dt=-frac18int_0^1fracln tsqrt1-tdt=-frac18int_0^1 t^-1/2ln(1-t)dt
\=frac14sum_n=0^infty frac1(n+1)(2n+3)=frac12-fracln22.endalign$$
$$beginalign int_0^1fracln(1+x^2)(1-x^2)x(1+x^2)dx=frac12int_0^1fracln(1+x)(1-x)x(1+x)dx tag14
\=frac12int_0^1fracln(1+x)xdx-int_0^1fracln(1+x)1+xdx
\=frac12sum_n=0^inftyfrac(-1)^n+1n+1int_0^1 x^n dx-frac12ln^2(1+x)bigg_0^1=fracpi^224-fracln^2 22.endalign$$
$$beginalign int_0^1fracln(1-x^2)(1-x^2)x(1+x^2)dx=frac12int_0^1fracln(1-x)xdx-int_0^1fracln(1-x)1+xdx tag15
\=-fracpi^212-left(fracln^2 22-fracpi^212right)=-fracln^2 22.endalign$$
$$
int_0^1fracln(1+x^2)(1+x^2)^2(1-x^2)dx=-2int_0^fracpi4cos^2(theta)(1-tan^2theta)lncostheta,dtheta tag16
\=-2int_0^fracpi4cos(2theta)lncostheta,dtheta=2ln2int_0^fracpi4cos(2theta)dtheta+sum_n=1^infty frac(-1)^nnint_0^fracpi2costheta cos(ntheta)dtheta
\=ln2-fracpi4+sum_n=1^infty frac(-1)^2n2nfrac(-1)^n+1(2n)^2-1=fracln22-fracpi4+frac12.$$
Altogether, $displaystyle I_242=fracpi^2192+fracpi16+fracln^2 216-frac3ln28+frac18$,
leading to $displaystyle I_24=I_241-I_242=frac5pi^2192-fracpi16-fracln^2 216+frac3ln28+fracG4$,
and finally, confirming the conjecture,
$$sum_n=0^infty(H_n-2H_2n+H_4n)^2=I_22-2I_24+I_44=fracpi8-fracpi16ln2-fracpi^296+frac3ln^2 216-fracG4.$$
I don't know about higher powers. I guess the case $mathcal A_2$ can also be done. If we start the same as with $mathcal B_2$, writing $mathcal A_2=J_22-2J_24+J_44$
we can find that $displaystyle J_22=int_0^1int_0^1fracdxdu(1+x)(1+u)(1+x^2u^2)=2I_44-I_22=-fracpi8ln2+fracG2+fracpi^248+fracln^2 28$
$J_44$ can be reduced to $displaystyle =-frac12int_0^1 fracln(1-x^2)x(x^4+6x^2+1)(1+x)^3 dx$.
already this i can't evaulate fully, as polylogs are inescapable. Factorizing $x^2+6x+1=(x+3+2sqrt2)(x+3-2sqrt2).$
I can get $displaystyle int_0^1 fracln(1-x^2)x(x^4+6x^2+1)dx=-fracpi^212+frac4-3sqrt216operatornameLi_2left(frac2-sqrt24right)+frac4+3sqrt216operatornameLi_2left(frac2+sqrt24right)$
but nothing more.
Edit 1.
After some more work and a fair amount of cancellation, we obtain
$$int_0^1 fracln(1-x^2)x(x^4+6x^2+1)(1+x)^3 dx=frac1+2sqrt24piln2-fracpi^224-frac14lnleft(frac2+sqrt24right)lnleft(frac2-sqrt24right)
-fracsqrt2+12ImoperatornameLi_2left(frac2+sqrt22+fracisqrt22right)-fracsqrt2-12ImoperatornameLi_2left(frac2-sqrt22+fracisqrt22right)$$
I obtained it by calculating $displaystyle int_0^1 fracln(1+x)x+adx=ln2lnleft(fraca+1a-1right)+operatornameLi_2left(frac21-aright)-operatornameLi_2left(frac11-aright)$,
which together with $(3)$ can be used to give a closed form for $displaystyle int_0^1fracln(1-x^2)x+adx$, which in turn, through partial fractions, can be used to give a closed form for $displaystyle int_0^1fracln(1-x^2)x^2+a^2dx$.
Fortunately, things didn't get too ugly as both $3+2sqrt2$ and $3-2sqrt2$ have nice square roots. I will fill in details as soon as I can.
Now we just need to evaluate $J_24$. Starting similarly as with $I_24$,
we have:
$$J_24=int_0^1int_0^1fracdxdu(1+x)(1+u)(1+x^4u^2)
\=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1+x^4)dx-2int_0^1fracxlnleft(frac(1+x)sqrt1+x^22sqrt2xright)(1+x^2)(1+x^4)dx
\=J_241-J_242$$
Through $t=frac1-x1+x$, $J_241$ turns to $displaystyle int_0^1 fractan^-1(x)(1+x^2)(x^4+6x^2+1)(1+x)^4,dx$. I don't have any idea about that yet. Edit 1.
$endgroup$
add a comment |
$begingroup$
So basically, I'll evaluate a bunch of integrals, trying to avoid polylogs as much as possible.
First thing is to notice that $displaystyle H_n-2H_2n+H_4n=int_0^1 fracx^2n-x^4n1+xdx$.
I noticed that $H_n-2H_2n+H_4n=H_4n-H_2n-(H_2n-H_n)=H_4n^--H_2n^-$, where $H_n^-=sum_k=1^n frac(-1)^k+1k$ is called
a skew harmonic number (at least by Khristo N. Boyadzhiev. link.) Knowing they have a simple intergal representation I found the above. My answer is influenced by Boyadzhiev's work.
If I make any unexplainable substitution, it's most likely $t=frac1-x1+x$.
Also, I'm not very good with Latex, so alignment should be awful. Hopefully there are no typos.
Below, easy enough to prove, is what I take for granted:
$ -lnsin x=ln2+sum_n=1^infty fraccos(2nx)n ,-lncos x=ln2+sum_n=1^infty frac(-1)^ncos(2nx)n tag1$
$$ int_0^fracpi2 cos x cos(nx)dx=begincases fracpi4 &n=1\0 &n ,,textodd\ frac(-1)^1+n/2n^2-1 &n ,,texteven endcases tag2$$
$$ int_0^1 fracln(1-x)a+xdx=-operatornameLi_2left(frac1a+1right)tag3$$
Starting,
$$sum_n=0^infty(H_n-2H_2n+H_4n)^2=sum_n=0^inftyint_0^1int_0^1frac(x^2n-x^4n)(u^2n-u^4n)(1+x)(1+u)dxdu
\=smallint_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^2)-2int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^4)+int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^4u^4)
\=I_22-2I_24+I_44$$
Computing $I_22$.
Substitute $u=fracyx$ ,change the order of integration, evaluate the inner integral, and substitue $t=frac1-x1+x$ to get
$$beginalign I_22=int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^2)=int_0^1int_0^xfracdydx(1+x)(x+y)(1-y^2)
\=int_0^1 frac11-y^2int_1^y fracdx(1+x)(x+y) dy=int_0^1 fraclnleft(frac(1+x)^24xright)(1+x)(1-x^2),dx
\=frac-14int_0^1 frac(1+t)t^2ln(1-t^2)dt=-frac14int_0^1fracln(1-t^2)t^2dt-frac14int_0^1fracln(1-t^2)tdt
\=frac14sum_n=0^infty frac1(n+1)(2n+1)+frac14sum_n=0^infty frac1(n+1)(2n+2)=fracln22+fracpi^248.endalign$$
Computing $I_44$.
Start the same as with $I_22$ to get $displaystyle I_44=int_0^1 fraclnleft(frac(1+x)^24xright)(1-x)(1-x^4),dx=frac-18int_0^1 fracln(1-t^2)t^2(1+t^2)(1+t)^3dt$.
We can calculate these integrals:
$$beginalign int_0^1 fracln(1-x^2)1+x^2dx=int_0^1 fracln(1+x)1+x^2dx+int_0^1 fracln(1-x)1+x^2dx tag4
\=int_0^1 fracln(1+x)1+x^2dx +int_0^1 fraclnleft(frac2t1+tright)1+t^2dt
\=fracpi4ln2+sum_n=0^infty (-1)^nint_0^1ln(t) t^2ndt=fracpi4ln2-G. endalign$$
$$beginalign int_0^1 fracln(1-x^2)x^2(1+x^2)dx=int_0^1 fracln(1-x^2)x^2dx-int_0^1 fracln(1-x^2)1+x^2dx tag5
\=-sum_n=0^infty frac1n+1int_0^1 x^2ndx-fracpi4ln2+G=G-fracpi4ln2-2ln2.endalign$$
$$beginalign int_0^1 fracxln(1-x^2)1+x^2dx=frac12int_0^1 fracln(1-x)1+xdx tag6
\=-frac12 operatornameLi_2left(frac12right)=fracln^2 24-fracpi^224.endalign$$
$$beginalign int_0^1 fracln(1-x^2)x(1+x^2)dx=int_0^1 fracln(1-x^2)xdx-int_0^1 fracxln(1-x^2)1+x^2dx tag7
\=-sum_n=0^inftyfrac1n+1int_0^1 x^2n+1dx-fracln^2 24+fracpi^224=-fracpi^224-fracln^2 24.endalign$$
Altogether,
$$I_44=frac-18int_0^1 fracln(1-x^2)x^2(1+x^2)(1+3x+3x^2+x^3)dx
\=-fracpi16ln2+fracln24+fracln^2 216+fracpi^248+fracG4.$$
Computing $I_24$.
Substitute $u=fracyx^2$, change the order of integration, let $yto y^2$, evaluate the inner integral,and substitue $t=frac1-x1+x$:
$$beginalign* I_24=int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^4u^2)=int_0^1int_0^x^2 fracdydx(1+x)(x^2+y)(1-y^2)
\=int_0^1 frac11-y^2int_sqrty^1fracdx(1+x)(x^2+y)dy=2int_0^1fracy1-y^4int_y^1fracdx(1+x)(x^2+y^2)dy
\=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1-x^4)dx-2int_0^1fracxlnleft(frac(1+x)sqrt1+x^22sqrt2xright)(1+x^2)(1-x^4)dx
=I_241-I_242. endalign*$$
Evaulation of $I_241$.
Substitute $t=frac1-x1+x$ to get $displaystyle I_241=frac14int_0^1 fractan^-1(t)t(1+t^2)^2(1+t)^4dt$.
We can calculate these integrals.In the following, let $x=tantheta$:
$$beginalign int_0^1 fractan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetacos^2(theta)dtheta=fracpi^264+fracpi16-frac18.tag8endalign$$
$$beginalign int_0^1 fracxtan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetatanthetacos^2thetadtheta=frac12int_0^fracpi4thetasin(2theta)dtheta=frac18.tag9endalign$$
$$beginalign int_0^1 fracx^2tan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetasin^2(theta)dtheta=fracpi^264-fracpi16+frac18.tag10endalign$$
$$beginalign int_0^1 fracx^3tan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetasin^3(theta)sectheta,dtheta tag11
\=int_0^fracpi4thetatantheta ,dtheta-int_0^fracpi4thetasinthetacostheta ,dtheta=fracpi8ln2-frac18+int_0^fracpi4lncostheta ,dtheta
\=fracpi8ln2-frac18-int_0^fracpi4ln2 ,dtheta-sum_n=1^infty frac(-1)^nnint_0^fracpi4cos(2ntheta),dtheta
\=-fracpi8ln2-frac18+frac12sum_n=1^inftyfrac(-1)^n+1n^2sin(fracpi n2)=fracG2-fracpi8ln2-frac18.endalign$$
$$beginalign int_0^1 fractan^-1(x)x(1+x^2)^2dx=int_0^fracpi4thetacos^3(theta)csctheta,dtheta tag12
\=int_0^fracpi4thetacottheta ,dtheta-int_0^fracpi4thetasinthetacostheta ,dtheta=-frac18-fracpi8ln2-int_0^fracpi4lnsintheta ,dtheta
\=-frac18-fracpi8ln2+int_0^fracpi4ln2 ,dtheta+sum_n=1^inftyfrac1nint_0^fracpi4cos(2ntheta),dtheta
\=-frac18+fracpi8ln2+frac12sum_n=1^infty fracsin(fracpi n2)n^2=fracG2+fracpi8ln2-frac18.endalign$$
Altogether,
$$I_241=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1-x^4)dx= frac14int_0^1 fractan^-1(x)x(1+x^2)^2(1+4x+6x^2+4x^3+x^4)dx
\=fracpi^232+frac18+fracG4$$
Evaulation of $I_242$.
Substitute $t=frac1-x1+x$ to get
$$ I_242=frac14int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)t(1+t^2)^2(1-t^2)(1+t)^2dt
\=frac12int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)(1+t^2)^2(1-t^2)dt+frac14int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)t(1+t^2)(1-t^2)dt
\=frac12int_0^1 fraclnleft(frac1+x^21-x^2right)(1+x^2)^2(1-x^2)dx-frac14int_0^1fracln(1+x^2)(1+x^2)^2(1-x^2)dx\+frac18int_0^1fracln(1+x^2)(1-x^2)x(1+x^2)dx-frac14int_0^1fracln(1-x^2)(1-x^2)x(1+x^2)dx
$$
Calculating these integrals:
$$beginalign int_0^1 fraclnleft(frac1+x^21-x^2right)(1+x^2)^2(1-x^2)dx=-frac12int_0^1 fraclnleft(frac1-x1+xright)sqrtx(1+x)frac1-x1+xdx tag13
\=-frac12int_0^1fractln tsqrt1-t^2dt=-frac18int_0^1fracln tsqrt1-tdt=-frac18int_0^1 t^-1/2ln(1-t)dt
\=frac14sum_n=0^infty frac1(n+1)(2n+3)=frac12-fracln22.endalign$$
$$beginalign int_0^1fracln(1+x^2)(1-x^2)x(1+x^2)dx=frac12int_0^1fracln(1+x)(1-x)x(1+x)dx tag14
\=frac12int_0^1fracln(1+x)xdx-int_0^1fracln(1+x)1+xdx
\=frac12sum_n=0^inftyfrac(-1)^n+1n+1int_0^1 x^n dx-frac12ln^2(1+x)bigg_0^1=fracpi^224-fracln^2 22.endalign$$
$$beginalign int_0^1fracln(1-x^2)(1-x^2)x(1+x^2)dx=frac12int_0^1fracln(1-x)xdx-int_0^1fracln(1-x)1+xdx tag15
\=-fracpi^212-left(fracln^2 22-fracpi^212right)=-fracln^2 22.endalign$$
$$
int_0^1fracln(1+x^2)(1+x^2)^2(1-x^2)dx=-2int_0^fracpi4cos^2(theta)(1-tan^2theta)lncostheta,dtheta tag16
\=-2int_0^fracpi4cos(2theta)lncostheta,dtheta=2ln2int_0^fracpi4cos(2theta)dtheta+sum_n=1^infty frac(-1)^nnint_0^fracpi2costheta cos(ntheta)dtheta
\=ln2-fracpi4+sum_n=1^infty frac(-1)^2n2nfrac(-1)^n+1(2n)^2-1=fracln22-fracpi4+frac12.$$
Altogether, $displaystyle I_242=fracpi^2192+fracpi16+fracln^2 216-frac3ln28+frac18$,
leading to $displaystyle I_24=I_241-I_242=frac5pi^2192-fracpi16-fracln^2 216+frac3ln28+fracG4$,
and finally, confirming the conjecture,
$$sum_n=0^infty(H_n-2H_2n+H_4n)^2=I_22-2I_24+I_44=fracpi8-fracpi16ln2-fracpi^296+frac3ln^2 216-fracG4.$$
I don't know about higher powers. I guess the case $mathcal A_2$ can also be done. If we start the same as with $mathcal B_2$, writing $mathcal A_2=J_22-2J_24+J_44$
we can find that $displaystyle J_22=int_0^1int_0^1fracdxdu(1+x)(1+u)(1+x^2u^2)=2I_44-I_22=-fracpi8ln2+fracG2+fracpi^248+fracln^2 28$
$J_44$ can be reduced to $displaystyle =-frac12int_0^1 fracln(1-x^2)x(x^4+6x^2+1)(1+x)^3 dx$.
already this i can't evaulate fully, as polylogs are inescapable. Factorizing $x^2+6x+1=(x+3+2sqrt2)(x+3-2sqrt2).$
I can get $displaystyle int_0^1 fracln(1-x^2)x(x^4+6x^2+1)dx=-fracpi^212+frac4-3sqrt216operatornameLi_2left(frac2-sqrt24right)+frac4+3sqrt216operatornameLi_2left(frac2+sqrt24right)$
but nothing more.
Edit 1.
After some more work and a fair amount of cancellation, we obtain
$$int_0^1 fracln(1-x^2)x(x^4+6x^2+1)(1+x)^3 dx=frac1+2sqrt24piln2-fracpi^224-frac14lnleft(frac2+sqrt24right)lnleft(frac2-sqrt24right)
-fracsqrt2+12ImoperatornameLi_2left(frac2+sqrt22+fracisqrt22right)-fracsqrt2-12ImoperatornameLi_2left(frac2-sqrt22+fracisqrt22right)$$
I obtained it by calculating $displaystyle int_0^1 fracln(1+x)x+adx=ln2lnleft(fraca+1a-1right)+operatornameLi_2left(frac21-aright)-operatornameLi_2left(frac11-aright)$,
which together with $(3)$ can be used to give a closed form for $displaystyle int_0^1fracln(1-x^2)x+adx$, which in turn, through partial fractions, can be used to give a closed form for $displaystyle int_0^1fracln(1-x^2)x^2+a^2dx$.
Fortunately, things didn't get too ugly as both $3+2sqrt2$ and $3-2sqrt2$ have nice square roots. I will fill in details as soon as I can.
Now we just need to evaluate $J_24$. Starting similarly as with $I_24$,
we have:
$$J_24=int_0^1int_0^1fracdxdu(1+x)(1+u)(1+x^4u^2)
\=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1+x^4)dx-2int_0^1fracxlnleft(frac(1+x)sqrt1+x^22sqrt2xright)(1+x^2)(1+x^4)dx
\=J_241-J_242$$
Through $t=frac1-x1+x$, $J_241$ turns to $displaystyle int_0^1 fractan^-1(x)(1+x^2)(x^4+6x^2+1)(1+x)^4,dx$. I don't have any idea about that yet. Edit 1.
$endgroup$
So basically, I'll evaluate a bunch of integrals, trying to avoid polylogs as much as possible.
First thing is to notice that $displaystyle H_n-2H_2n+H_4n=int_0^1 fracx^2n-x^4n1+xdx$.
I noticed that $H_n-2H_2n+H_4n=H_4n-H_2n-(H_2n-H_n)=H_4n^--H_2n^-$, where $H_n^-=sum_k=1^n frac(-1)^k+1k$ is called
a skew harmonic number (at least by Khristo N. Boyadzhiev. link.) Knowing they have a simple intergal representation I found the above. My answer is influenced by Boyadzhiev's work.
If I make any unexplainable substitution, it's most likely $t=frac1-x1+x$.
Also, I'm not very good with Latex, so alignment should be awful. Hopefully there are no typos.
Below, easy enough to prove, is what I take for granted:
$ -lnsin x=ln2+sum_n=1^infty fraccos(2nx)n ,-lncos x=ln2+sum_n=1^infty frac(-1)^ncos(2nx)n tag1$
$$ int_0^fracpi2 cos x cos(nx)dx=begincases fracpi4 &n=1\0 &n ,,textodd\ frac(-1)^1+n/2n^2-1 &n ,,texteven endcases tag2$$
$$ int_0^1 fracln(1-x)a+xdx=-operatornameLi_2left(frac1a+1right)tag3$$
Starting,
$$sum_n=0^infty(H_n-2H_2n+H_4n)^2=sum_n=0^inftyint_0^1int_0^1frac(x^2n-x^4n)(u^2n-u^4n)(1+x)(1+u)dxdu
\=smallint_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^2)-2int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^4)+int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^4u^4)
\=I_22-2I_24+I_44$$
Computing $I_22$.
Substitute $u=fracyx$ ,change the order of integration, evaluate the inner integral, and substitue $t=frac1-x1+x$ to get
$$beginalign I_22=int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^2u^2)=int_0^1int_0^xfracdydx(1+x)(x+y)(1-y^2)
\=int_0^1 frac11-y^2int_1^y fracdx(1+x)(x+y) dy=int_0^1 fraclnleft(frac(1+x)^24xright)(1+x)(1-x^2),dx
\=frac-14int_0^1 frac(1+t)t^2ln(1-t^2)dt=-frac14int_0^1fracln(1-t^2)t^2dt-frac14int_0^1fracln(1-t^2)tdt
\=frac14sum_n=0^infty frac1(n+1)(2n+1)+frac14sum_n=0^infty frac1(n+1)(2n+2)=fracln22+fracpi^248.endalign$$
Computing $I_44$.
Start the same as with $I_22$ to get $displaystyle I_44=int_0^1 fraclnleft(frac(1+x)^24xright)(1-x)(1-x^4),dx=frac-18int_0^1 fracln(1-t^2)t^2(1+t^2)(1+t)^3dt$.
We can calculate these integrals:
$$beginalign int_0^1 fracln(1-x^2)1+x^2dx=int_0^1 fracln(1+x)1+x^2dx+int_0^1 fracln(1-x)1+x^2dx tag4
\=int_0^1 fracln(1+x)1+x^2dx +int_0^1 fraclnleft(frac2t1+tright)1+t^2dt
\=fracpi4ln2+sum_n=0^infty (-1)^nint_0^1ln(t) t^2ndt=fracpi4ln2-G. endalign$$
$$beginalign int_0^1 fracln(1-x^2)x^2(1+x^2)dx=int_0^1 fracln(1-x^2)x^2dx-int_0^1 fracln(1-x^2)1+x^2dx tag5
\=-sum_n=0^infty frac1n+1int_0^1 x^2ndx-fracpi4ln2+G=G-fracpi4ln2-2ln2.endalign$$
$$beginalign int_0^1 fracxln(1-x^2)1+x^2dx=frac12int_0^1 fracln(1-x)1+xdx tag6
\=-frac12 operatornameLi_2left(frac12right)=fracln^2 24-fracpi^224.endalign$$
$$beginalign int_0^1 fracln(1-x^2)x(1+x^2)dx=int_0^1 fracln(1-x^2)xdx-int_0^1 fracxln(1-x^2)1+x^2dx tag7
\=-sum_n=0^inftyfrac1n+1int_0^1 x^2n+1dx-fracln^2 24+fracpi^224=-fracpi^224-fracln^2 24.endalign$$
Altogether,
$$I_44=frac-18int_0^1 fracln(1-x^2)x^2(1+x^2)(1+3x+3x^2+x^3)dx
\=-fracpi16ln2+fracln24+fracln^2 216+fracpi^248+fracG4.$$
Computing $I_24$.
Substitute $u=fracyx^2$, change the order of integration, let $yto y^2$, evaluate the inner integral,and substitue $t=frac1-x1+x$:
$$beginalign* I_24=int_0^1int_0^1fracdxdu(1+x)(1+u)(1-x^4u^2)=int_0^1int_0^x^2 fracdydx(1+x)(x^2+y)(1-y^2)
\=int_0^1 frac11-y^2int_sqrty^1fracdx(1+x)(x^2+y)dy=2int_0^1fracy1-y^4int_y^1fracdx(1+x)(x^2+y^2)dy
\=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1-x^4)dx-2int_0^1fracxlnleft(frac(1+x)sqrt1+x^22sqrt2xright)(1+x^2)(1-x^4)dx
=I_241-I_242. endalign*$$
Evaulation of $I_241$.
Substitute $t=frac1-x1+x$ to get $displaystyle I_241=frac14int_0^1 fractan^-1(t)t(1+t^2)^2(1+t)^4dt$.
We can calculate these integrals.In the following, let $x=tantheta$:
$$beginalign int_0^1 fractan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetacos^2(theta)dtheta=fracpi^264+fracpi16-frac18.tag8endalign$$
$$beginalign int_0^1 fracxtan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetatanthetacos^2thetadtheta=frac12int_0^fracpi4thetasin(2theta)dtheta=frac18.tag9endalign$$
$$beginalign int_0^1 fracx^2tan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetasin^2(theta)dtheta=fracpi^264-fracpi16+frac18.tag10endalign$$
$$beginalign int_0^1 fracx^3tan^-1(x)(1+x^2)^2dx=int_0^fracpi4thetasin^3(theta)sectheta,dtheta tag11
\=int_0^fracpi4thetatantheta ,dtheta-int_0^fracpi4thetasinthetacostheta ,dtheta=fracpi8ln2-frac18+int_0^fracpi4lncostheta ,dtheta
\=fracpi8ln2-frac18-int_0^fracpi4ln2 ,dtheta-sum_n=1^infty frac(-1)^nnint_0^fracpi4cos(2ntheta),dtheta
\=-fracpi8ln2-frac18+frac12sum_n=1^inftyfrac(-1)^n+1n^2sin(fracpi n2)=fracG2-fracpi8ln2-frac18.endalign$$
$$beginalign int_0^1 fractan^-1(x)x(1+x^2)^2dx=int_0^fracpi4thetacos^3(theta)csctheta,dtheta tag12
\=int_0^fracpi4thetacottheta ,dtheta-int_0^fracpi4thetasinthetacostheta ,dtheta=-frac18-fracpi8ln2-int_0^fracpi4lnsintheta ,dtheta
\=-frac18-fracpi8ln2+int_0^fracpi4ln2 ,dtheta+sum_n=1^inftyfrac1nint_0^fracpi4cos(2ntheta),dtheta
\=-frac18+fracpi8ln2+frac12sum_n=1^infty fracsin(fracpi n2)n^2=fracG2+fracpi8ln2-frac18.endalign$$
Altogether,
$$I_241=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1-x^4)dx= frac14int_0^1 fractan^-1(x)x(1+x^2)^2(1+4x+6x^2+4x^3+x^4)dx
\=fracpi^232+frac18+fracG4$$
Evaulation of $I_242$.
Substitute $t=frac1-x1+x$ to get
$$ I_242=frac14int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)t(1+t^2)^2(1-t^2)(1+t)^2dt
\=frac12int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)(1+t^2)^2(1-t^2)dt+frac14int_0^1 fraclnleft(fracsqrt1+t^21-t^2right)t(1+t^2)(1-t^2)dt
\=frac12int_0^1 fraclnleft(frac1+x^21-x^2right)(1+x^2)^2(1-x^2)dx-frac14int_0^1fracln(1+x^2)(1+x^2)^2(1-x^2)dx\+frac18int_0^1fracln(1+x^2)(1-x^2)x(1+x^2)dx-frac14int_0^1fracln(1-x^2)(1-x^2)x(1+x^2)dx
$$
Calculating these integrals:
$$beginalign int_0^1 fraclnleft(frac1+x^21-x^2right)(1+x^2)^2(1-x^2)dx=-frac12int_0^1 fraclnleft(frac1-x1+xright)sqrtx(1+x)frac1-x1+xdx tag13
\=-frac12int_0^1fractln tsqrt1-t^2dt=-frac18int_0^1fracln tsqrt1-tdt=-frac18int_0^1 t^-1/2ln(1-t)dt
\=frac14sum_n=0^infty frac1(n+1)(2n+3)=frac12-fracln22.endalign$$
$$beginalign int_0^1fracln(1+x^2)(1-x^2)x(1+x^2)dx=frac12int_0^1fracln(1+x)(1-x)x(1+x)dx tag14
\=frac12int_0^1fracln(1+x)xdx-int_0^1fracln(1+x)1+xdx
\=frac12sum_n=0^inftyfrac(-1)^n+1n+1int_0^1 x^n dx-frac12ln^2(1+x)bigg_0^1=fracpi^224-fracln^2 22.endalign$$
$$beginalign int_0^1fracln(1-x^2)(1-x^2)x(1+x^2)dx=frac12int_0^1fracln(1-x)xdx-int_0^1fracln(1-x)1+xdx tag15
\=-fracpi^212-left(fracln^2 22-fracpi^212right)=-fracln^2 22.endalign$$
$$
int_0^1fracln(1+x^2)(1+x^2)^2(1-x^2)dx=-2int_0^fracpi4cos^2(theta)(1-tan^2theta)lncostheta,dtheta tag16
\=-2int_0^fracpi4cos(2theta)lncostheta,dtheta=2ln2int_0^fracpi4cos(2theta)dtheta+sum_n=1^infty frac(-1)^nnint_0^fracpi2costheta cos(ntheta)dtheta
\=ln2-fracpi4+sum_n=1^infty frac(-1)^2n2nfrac(-1)^n+1(2n)^2-1=fracln22-fracpi4+frac12.$$
Altogether, $displaystyle I_242=fracpi^2192+fracpi16+fracln^2 216-frac3ln28+frac18$,
leading to $displaystyle I_24=I_241-I_242=frac5pi^2192-fracpi16-fracln^2 216+frac3ln28+fracG4$,
and finally, confirming the conjecture,
$$sum_n=0^infty(H_n-2H_2n+H_4n)^2=I_22-2I_24+I_44=fracpi8-fracpi16ln2-fracpi^296+frac3ln^2 216-fracG4.$$
I don't know about higher powers. I guess the case $mathcal A_2$ can also be done. If we start the same as with $mathcal B_2$, writing $mathcal A_2=J_22-2J_24+J_44$
we can find that $displaystyle J_22=int_0^1int_0^1fracdxdu(1+x)(1+u)(1+x^2u^2)=2I_44-I_22=-fracpi8ln2+fracG2+fracpi^248+fracln^2 28$
$J_44$ can be reduced to $displaystyle =-frac12int_0^1 fracln(1-x^2)x(x^4+6x^2+1)(1+x)^3 dx$.
already this i can't evaulate fully, as polylogs are inescapable. Factorizing $x^2+6x+1=(x+3+2sqrt2)(x+3-2sqrt2).$
I can get $displaystyle int_0^1 fracln(1-x^2)x(x^4+6x^2+1)dx=-fracpi^212+frac4-3sqrt216operatornameLi_2left(frac2-sqrt24right)+frac4+3sqrt216operatornameLi_2left(frac2+sqrt24right)$
but nothing more.
Edit 1.
After some more work and a fair amount of cancellation, we obtain
$$int_0^1 fracln(1-x^2)x(x^4+6x^2+1)(1+x)^3 dx=frac1+2sqrt24piln2-fracpi^224-frac14lnleft(frac2+sqrt24right)lnleft(frac2-sqrt24right)
-fracsqrt2+12ImoperatornameLi_2left(frac2+sqrt22+fracisqrt22right)-fracsqrt2-12ImoperatornameLi_2left(frac2-sqrt22+fracisqrt22right)$$
I obtained it by calculating $displaystyle int_0^1 fracln(1+x)x+adx=ln2lnleft(fraca+1a-1right)+operatornameLi_2left(frac21-aright)-operatornameLi_2left(frac11-aright)$,
which together with $(3)$ can be used to give a closed form for $displaystyle int_0^1fracln(1-x^2)x+adx$, which in turn, through partial fractions, can be used to give a closed form for $displaystyle int_0^1fracln(1-x^2)x^2+a^2dx$.
Fortunately, things didn't get too ugly as both $3+2sqrt2$ and $3-2sqrt2$ have nice square roots. I will fill in details as soon as I can.
Now we just need to evaluate $J_24$. Starting similarly as with $I_24$,
we have:
$$J_24=int_0^1int_0^1fracdxdu(1+x)(1+u)(1+x^4u^2)
\=2int_0^1fractan^-1left(frac1-x1+xright)(1+x^2)(1+x^4)dx-2int_0^1fracxlnleft(frac(1+x)sqrt1+x^22sqrt2xright)(1+x^2)(1+x^4)dx
\=J_241-J_242$$
Through $t=frac1-x1+x$, $J_241$ turns to $displaystyle int_0^1 fractan^-1(x)(1+x^2)(x^4+6x^2+1)(1+x)^4,dx$. I don't have any idea about that yet. Edit 1.
edited Sep 15 '15 at 15:09
answered Sep 14 '15 at 15:51
nospoonnospoon
4,6361532
4,6361532
add a comment |
add a comment |
$begingroup$
Just my thoughts for now: I would try to exploit Parseval's identity. For first, we have:
$$ H_n-2H_2n+H_4n =int_0^1frac-x^n+2x^2n-x^4n1-x,dx tag1 $$
and:
$$ sum_ngeq 1frac-x^n+2x^2n-x^4n1-xe^niy = frac11-xleft(frac-11-e^iyx+frac21-e^iyx^2+frac-11-e^iyx^4right).tag2$$
The poles of the RHS (as a function of $x$) are located at $xinlefte^-iy,pm e^-iy/2,pm e^-iy/4,pm i e^-iy/4right$.
By using the residue theorem we may compute an explicit representation for:
$$ g(y) = sum_ngeq 1left(H_n-2H_2n+H_4nright)e^niy,tag3 $$
then Parseval's theorem gives:
$$ sum_ngeq 1left(H_n-2H_2n+H_4nright)^2 = frac12piint_-pi^pig(y)g(-y),dy tag4$$
and the resulting integral should be not to difficult to evaluate in terms of dilogarithms.
Another chance may be to apply summation by parts (like I did in this question), but it looks lengthy.
$endgroup$
$begingroup$
It’s a very nice idea, but I think you are underestimating the difficulty of evaluating the integral. I checked the representation for $g$ using software: the result is a ton of terms of rational functions of trig’s and inverse trig’s and that is before computing $g(y)g(-y)$. This approach might work, but it looks like a lot of work.
$endgroup$
– Winther
Sep 1 '15 at 15:26
$begingroup$
@Winther: I know, it is a tough nut to crack with bare hands, but with some human-guided simplifications, it just boils down to computing some dilogarithmic integrals related with the fourth roots of unity. I haven't really delved into summation by parts, yet. It looks promising, maybe it is the simple way.
$endgroup$
– Jack D'Aurizio
Sep 1 '15 at 15:35
add a comment |
$begingroup$
Just my thoughts for now: I would try to exploit Parseval's identity. For first, we have:
$$ H_n-2H_2n+H_4n =int_0^1frac-x^n+2x^2n-x^4n1-x,dx tag1 $$
and:
$$ sum_ngeq 1frac-x^n+2x^2n-x^4n1-xe^niy = frac11-xleft(frac-11-e^iyx+frac21-e^iyx^2+frac-11-e^iyx^4right).tag2$$
The poles of the RHS (as a function of $x$) are located at $xinlefte^-iy,pm e^-iy/2,pm e^-iy/4,pm i e^-iy/4right$.
By using the residue theorem we may compute an explicit representation for:
$$ g(y) = sum_ngeq 1left(H_n-2H_2n+H_4nright)e^niy,tag3 $$
then Parseval's theorem gives:
$$ sum_ngeq 1left(H_n-2H_2n+H_4nright)^2 = frac12piint_-pi^pig(y)g(-y),dy tag4$$
and the resulting integral should be not to difficult to evaluate in terms of dilogarithms.
Another chance may be to apply summation by parts (like I did in this question), but it looks lengthy.
$endgroup$
$begingroup$
It’s a very nice idea, but I think you are underestimating the difficulty of evaluating the integral. I checked the representation for $g$ using software: the result is a ton of terms of rational functions of trig’s and inverse trig’s and that is before computing $g(y)g(-y)$. This approach might work, but it looks like a lot of work.
$endgroup$
– Winther
Sep 1 '15 at 15:26
$begingroup$
@Winther: I know, it is a tough nut to crack with bare hands, but with some human-guided simplifications, it just boils down to computing some dilogarithmic integrals related with the fourth roots of unity. I haven't really delved into summation by parts, yet. It looks promising, maybe it is the simple way.
$endgroup$
– Jack D'Aurizio
Sep 1 '15 at 15:35
add a comment |
$begingroup$
Just my thoughts for now: I would try to exploit Parseval's identity. For first, we have:
$$ H_n-2H_2n+H_4n =int_0^1frac-x^n+2x^2n-x^4n1-x,dx tag1 $$
and:
$$ sum_ngeq 1frac-x^n+2x^2n-x^4n1-xe^niy = frac11-xleft(frac-11-e^iyx+frac21-e^iyx^2+frac-11-e^iyx^4right).tag2$$
The poles of the RHS (as a function of $x$) are located at $xinlefte^-iy,pm e^-iy/2,pm e^-iy/4,pm i e^-iy/4right$.
By using the residue theorem we may compute an explicit representation for:
$$ g(y) = sum_ngeq 1left(H_n-2H_2n+H_4nright)e^niy,tag3 $$
then Parseval's theorem gives:
$$ sum_ngeq 1left(H_n-2H_2n+H_4nright)^2 = frac12piint_-pi^pig(y)g(-y),dy tag4$$
and the resulting integral should be not to difficult to evaluate in terms of dilogarithms.
Another chance may be to apply summation by parts (like I did in this question), but it looks lengthy.
$endgroup$
Just my thoughts for now: I would try to exploit Parseval's identity. For first, we have:
$$ H_n-2H_2n+H_4n =int_0^1frac-x^n+2x^2n-x^4n1-x,dx tag1 $$
and:
$$ sum_ngeq 1frac-x^n+2x^2n-x^4n1-xe^niy = frac11-xleft(frac-11-e^iyx+frac21-e^iyx^2+frac-11-e^iyx^4right).tag2$$
The poles of the RHS (as a function of $x$) are located at $xinlefte^-iy,pm e^-iy/2,pm e^-iy/4,pm i e^-iy/4right$.
By using the residue theorem we may compute an explicit representation for:
$$ g(y) = sum_ngeq 1left(H_n-2H_2n+H_4nright)e^niy,tag3 $$
then Parseval's theorem gives:
$$ sum_ngeq 1left(H_n-2H_2n+H_4nright)^2 = frac12piint_-pi^pig(y)g(-y),dy tag4$$
and the resulting integral should be not to difficult to evaluate in terms of dilogarithms.
Another chance may be to apply summation by parts (like I did in this question), but it looks lengthy.
edited Apr 13 '17 at 12:21
Community♦
1
1
answered Aug 30 '15 at 17:36
Jack D'AurizioJack D'Aurizio
291k33284667
291k33284667
$begingroup$
It’s a very nice idea, but I think you are underestimating the difficulty of evaluating the integral. I checked the representation for $g$ using software: the result is a ton of terms of rational functions of trig’s and inverse trig’s and that is before computing $g(y)g(-y)$. This approach might work, but it looks like a lot of work.
$endgroup$
– Winther
Sep 1 '15 at 15:26
$begingroup$
@Winther: I know, it is a tough nut to crack with bare hands, but with some human-guided simplifications, it just boils down to computing some dilogarithmic integrals related with the fourth roots of unity. I haven't really delved into summation by parts, yet. It looks promising, maybe it is the simple way.
$endgroup$
– Jack D'Aurizio
Sep 1 '15 at 15:35
add a comment |
$begingroup$
It’s a very nice idea, but I think you are underestimating the difficulty of evaluating the integral. I checked the representation for $g$ using software: the result is a ton of terms of rational functions of trig’s and inverse trig’s and that is before computing $g(y)g(-y)$. This approach might work, but it looks like a lot of work.
$endgroup$
– Winther
Sep 1 '15 at 15:26
$begingroup$
@Winther: I know, it is a tough nut to crack with bare hands, but with some human-guided simplifications, it just boils down to computing some dilogarithmic integrals related with the fourth roots of unity. I haven't really delved into summation by parts, yet. It looks promising, maybe it is the simple way.
$endgroup$
– Jack D'Aurizio
Sep 1 '15 at 15:35
$begingroup$
It’s a very nice idea, but I think you are underestimating the difficulty of evaluating the integral. I checked the representation for $g$ using software: the result is a ton of terms of rational functions of trig’s and inverse trig’s and that is before computing $g(y)g(-y)$. This approach might work, but it looks like a lot of work.
$endgroup$
– Winther
Sep 1 '15 at 15:26
$begingroup$
It’s a very nice idea, but I think you are underestimating the difficulty of evaluating the integral. I checked the representation for $g$ using software: the result is a ton of terms of rational functions of trig’s and inverse trig’s and that is before computing $g(y)g(-y)$. This approach might work, but it looks like a lot of work.
$endgroup$
– Winther
Sep 1 '15 at 15:26
$begingroup$
@Winther: I know, it is a tough nut to crack with bare hands, but with some human-guided simplifications, it just boils down to computing some dilogarithmic integrals related with the fourth roots of unity. I haven't really delved into summation by parts, yet. It looks promising, maybe it is the simple way.
$endgroup$
– Jack D'Aurizio
Sep 1 '15 at 15:35
$begingroup$
@Winther: I know, it is a tough nut to crack with bare hands, but with some human-guided simplifications, it just boils down to computing some dilogarithmic integrals related with the fourth roots of unity. I haven't really delved into summation by parts, yet. It looks promising, maybe it is the simple way.
$endgroup$
– Jack D'Aurizio
Sep 1 '15 at 15:35
add a comment |
Thanks for contributing an answer to Mathematics Stack Exchange!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f1406122%2fconjectured-value-of-a-harmonic-sum-sum-n-1-infty-lefth-n-2h-2nh-4n%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
$begingroup$
Some possibly related questions are linked from here.
$endgroup$
– Piotr Shatalin
Aug 22 '15 at 18:18
3
$begingroup$
BTW, another similar series can be evaluated using generating functions: $sum_n=1^inftyfrac1nleft(H_n-,2H_2n+H_4nright)=frac34 ln^2 2-fracpi^248.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 19:07
2
$begingroup$
And another: $sum_n=1^inftyfrac(-1)^nnleft(H_n-,2H_2n+H_4nright)=frac12 ln^2!left(1+sqrt2right)+frac18 ln^2 2-frac5pi^296.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 19:43
$begingroup$
One way to tackle it, at least the case $m=2$, would be to write $H_n=int_0^1frac1-x^n1-xdx$ and to convert the series into a double integral which probably could be evaluated more easily. But I think this is what we call "Brute Force".
$endgroup$
– Redundant Aunt
Aug 22 '15 at 19:56
4
$begingroup$
One more simple series where $O(n^-1)$ term is cancelled using harmonic numbers only: $sum_n=1^inftyleft(H_n-4H_2n+5H_4n-2H_8nright)=fracpi4sqrt2 - frac3pi16.$
$endgroup$
– Vladimir Reshetnikov
Aug 22 '15 at 22:14