Aren't you amazed at this?

Calculus Level 5

0 1 0 1 0 1 ( { x y } { y z } { z x } ) 2 d x d y d z \int_0^1 \int_0^1 \int_0^1 \left(\left\{\frac{x}{y}\right\}\left\{\frac{y}{z}\right\}\left\{\frac{z}{x}\right\}\right)^2 \; dx\; dy\; dz

The integral above is equal to A π 2 B ζ ( C ) D + π 2 ζ ( E ) F + π 4 ζ ( G ) H + π 6 I + ζ 2 ( J ) K A-\dfrac{\pi^2}{B}-\dfrac{\zeta(C)}{D}+\dfrac{\pi^2\zeta(E)}{F}+\dfrac{\pi^4 \zeta(G)}{H}+\dfrac{\pi^6}{I}+\dfrac{\zeta^2 (J)}{K} for positive integers A , B , C , D , E , F , G , H , I , J , K . A, B, C, D, E, F, G, H, I, J, K.

What is A + B + C + D + E + F + G + H + I + J + K ? A+B+C+D+E+F+G+H+I+J+K?


Notation: { } \{ \cdot\} denotes the fractional part function .


Note: The solution of this particular problem isn't available anywhere on the web. This is termed as an open problem in the book Fractional Parts, Series, Integrals by Springer.


The answer is 7713.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

1 solution

Denote our Integral by I, then by noting the cyclic symmetry we can break it down into intervals as,

0 1 0 1 0 1 ( { x y } { y z } { z x } ) 2 d x d y d z = 3 0 < x < y < z < 1 ( x y y z { z x } ) 2 d x d y d z + 3 0 < x < z < y < 1 ( x y { y z } { z x } ) 2 d x d y d z = I 1 + I 2 \displaystyle \begin{aligned} &\int_0^1 \int_0^1 \int_0^1\left(\left\{\dfrac{x}{y}\right\}\left\{\dfrac{y}{z}\right\}\left\{\dfrac{z}{x}\right\}\right)^2 \; dx\; dy\; dz \\ &= 3\int\limits_{0<x<y<z<1}\left(\dfrac{x}{y}\dfrac{y}{z}\left\{\dfrac{z}{x}\right\}\right)^2 \; dx\; dy\; dz + 3\int\limits_{0<x<z<y<1} \left(\dfrac{x}{y}\left\{\dfrac{y}{z}\right\}\left\{\dfrac{z}{x}\right\}\right)^2 \; dx\; dy\; dz \\ &=I_1+I_2 \end{aligned}

We will calculate I 1 I_1 and I 2 I_2 separately here,

I 1 = 3 0 < x < y < z < 1 ( x y y z { z x } ) 2 d x d y d z = 3 z = 0 1 y = x z x = 0 z x 2 z 2 { z x } 2 d x d y d z = 3 z = 0 1 x = 0 z ( z x ) x 2 z 2 { z x } 2 d x d y d z = 3 z = 0 1 t = 0 1 z 2 t 2 ( 1 t ) { 1 t } 2 d t d z By x=zt = z = 0 1 3 z 2 d z t = 0 1 ( t 2 t 3 ) { 1 t } 2 d t = 0 1 t 2 { 1 t } 2 d t 0 1 t 3 { 1 t } 2 d t = S 2 S 3 \displaystyle \begin{aligned} &I_1 \\ &= 3\int\limits_{0<x<y<z<1}\left(\dfrac{x}{y}\dfrac{y}{z}\left\{\dfrac{z}{x}\right\}\right)^2 \; dx\; dy\; dz \\ &= 3\int_{z=0}^1 \int_{y=x}^{z}\int_{x=0}^z \dfrac{x^2}{z^2} \left\{\dfrac{z}{x}\right\}^2 \; dx\; dy\; dz \\ &= 3 \int_{z=0}^1 \int_{x=0}^z \dfrac{(z-x)x^2}{z^2} \left\{\dfrac{z}{x}\right\}^2 \; dx\; dy\; dz \\ &= 3\int_{z=0}^1 \int_{t=0}^1 z^2 t^2(1-t) \left\{\dfrac{1}{t}\right\}^2 \; dt\; dz \quad \color{#3D99F6}{\text{By x=zt}} \\ &= \int_{z=0}^1 3z^2 \; dz \int_{t=0}^1 (t^2-t^3) \left\{\dfrac{1}{t}\right\}^2 \; dt \\ &= \int_{0}^1 t^2 \left\{\dfrac{1}{t}\right\}^2\; dt-\int_{0}^1 t^3 \left\{\dfrac{1}{t}\right\}^2\; dt \\ &= \mathfrak{S}_2-\mathfrak{S}_3 \end{aligned}

Next for I 2 I_2 we have,

I 2 = 3 0 < x < z < y < 1 ( x y { y z } { z x } ) 2 d x d y d z = 3 y = 0 1 x = 0 z z = 0 y ( x y { y z } { z x } ) 2 d x d y d z = 3 y = 0 1 z = 0 y { y z } 2 1 y 2 x = 0 z x 2 { z x } 2 d x d y d z = 3 y = 0 1 z = 0 y { y z } 2 z 3 y 2 t = 0 1 t 2 { 1 t } 2 d t d y d z By x=zt = S 2 y = 0 1 3 y 2 d y t = 0 1 t 3 { 1 t } 2 d t By z=yt = S 2 S 3 \displaystyle \begin{aligned} &I_2 \\ &= 3\int\limits_{0<x<z<y<1} \left(\dfrac{x}{y}\left\{\dfrac{y}{z}\right\}\left\{\dfrac{z}{x}\right\}\right)^2 \; dx\; dy\; dz \\ &= 3\int_{y=0}^1 \int_{x=0}^z \int_{z=0}^y \left(\dfrac{x}{y}\left\{\dfrac{y}{z}\right\}\left\{\dfrac{z}{x}\right\}\right)^2 \; dx\; dy\; dz \\ &= 3\int_{y=0}^1 \int_{z=0}^y \left\{\dfrac{y}{z}\right\}^2 \dfrac{1}{y^2} \int_{x=0}^z x^2 \left\{\dfrac{z}{x}\right\}^2 \; dx\; dy\; dz \\ &= 3\int_{y=0}^1 \int_{z=0}^y \left\{\dfrac{y}{z}\right\}^2 \dfrac{z^3}{y^2} \int_{t=0}^1 t^2 \left\{\dfrac{1}{t}\right\}^2 \; dt \; dy\; dz \quad \color{#3D99F6}{\text{By x=zt}} \\ &= \mathfrak{S}_2 \int_{y=0}^1 3y^2 \; dy \int_{t=0}^1 t^3 \left\{\dfrac{1}{t}\right\}^2\; dt \quad \color{#3D99F6}{\text{By z=yt}} \\ &= \mathfrak{S_2}\mathfrak{S_3}\end{aligned}

It remains to calculate S 2 \mathfrak{S_2} and S 3 \mathfrak{S_3} which can be evaluated as,

S 2 = t = 0 1 t 2 { 1 t } 2 d t = 1 { x } 2 x 4 d x = k = 1 0 1 x 2 ( k + x ) 4 d x = k = 1 1 3 k ( k + 1 ) 3 = 1 ζ ( 2 ) 3 ζ ( 3 ) 3 \displaystyle \begin{aligned} &\mathfrak{S}_2 \\ &= \int_{t=0}^1 t^2 \left\{\dfrac{1}{t}\right\}^2\; dt \\ &= \int_{1}^\infty \dfrac{\{x\}^2}{x^4}\; dx \\ &= \sum_{k=1}^\infty \int_0^1 \dfrac{x^2}{(k+x)^4}\; dx \\ &= \sum_{k=1}^\infty \dfrac{1}{3k(k+1)^3} \\ &= 1-\dfrac{\zeta(2)}{3}-\dfrac{\zeta(3)}{3}\end{aligned}

S 3 = t = 0 1 t 3 { 1 t } 2 d t = 1 { x } 2 x 5 d x = k = 1 0 1 x 2 ( k + x ) 5 d x = k = 1 4 k + 1 12 k 2 ( k + 1 ) 4 = 1 2 ζ ( 3 ) 6 ζ ( 4 ) 4 \displaystyle \begin{aligned} &\mathfrak{S}_3 \\ &= \int_{t=0}^1 t^3 \left\{\dfrac{1}{t}\right\}^2\; dt \\ &= \int_{1}^\infty \dfrac{\{x\}^2}{x^5}\; dx \\ &= \sum_{k=1}^{\infty}\int_0^1 \dfrac{x^2}{(k+x)^5}\; dx \\ &= \sum_{k=1}^\infty \dfrac{4k+1}{12k^2(k+1)^4} \\ &= \dfrac{1}{2}-\dfrac{\zeta(3)}{6}-\dfrac{\zeta(4)}{4}\end{aligned}

Now combining we have,

0 1 0 1 0 1 ( { x y } { y z } { z x } ) 2 d x d y d z = I 1 + I 2 = S 2 S 3 + S 2 S 3 = 1 π 2 12 ζ ( 3 ) 2 + π 2 ζ ( 3 ) 108 + π 4 ζ ( 3 ) 1080 + π 6 6480 + ζ 2 ( 3 ) 18 \displaystyle \begin{aligned} &\int_0^1 \int_0^1 \int_0^1\left(\left\{\dfrac{x}{y}\right\}\left\{\dfrac{y}{z}\right\}\left\{\dfrac{z}{x}\right\}\right)^2 \; dx\; dy\; dz \\ &= I_1+I_2 \\ &= \mathfrak{S}_2-\mathfrak{S}_3+\mathfrak{S}_2\mathfrak{S}_3 \\ &= 1-\dfrac{\pi^2}{12}-\dfrac{\zeta(3)}{2}+\dfrac{\pi^2\zeta(3)}{108}+\dfrac{\pi^4 \zeta(3)}{1080}+\dfrac{\pi^6}{6480}+\dfrac{\zeta^2 (3)}{18}\end{aligned}

This makes the answer 1 + 12 + 3 + 2 + 3 + 108 + 6480 + 3 + 18 + 3 + 1080 = 7713 \boxed{1+12+3+2+3+108+6480+3+18+3+1080=7713}

I used pretty much the same method. Isn't the open problem in Furdui's book one where the integrand has not been squared? I think that would make things a little harder, but haven't tried it yet.

Mark Hennings - 4 years ago

Log in to reply

That was a general n n th power. I have tried for the case of 4 variables only till now. Are you talking about that one?

Log in to reply

There are n n variables in Open Problem 2.58, but the fractional parts are not squared. Or is there another OP I have missed?

Mark Hennings - 4 years ago

Log in to reply

@Mark Hennings Oh I messed up this with the power integral. You were right, but I don't think the case where it is not squares would be difficult as it is. I mean to say we will end up with some easy integrals at the end maybe. I have done the fourth case but not sure of posting it on brilliant. I don't know if it would be ok to post it on brilliant as a note.

Log in to reply

@Aditya Narayan Sharma The result without the squares (for three variables) is OK to work out in terms of π \pi and ζ ( 3 ) \zeta(3) . The point of OP 2.58 is to have a single expression for an integral of n n variables, and not just to be able to do this for three or four variables. Doing these "small n n " cases is good, but leaves the full OP unsolved, unless enough insight is obtained about how to go from three to four dimensions to show us how to do the general case.

Mark Hennings - 4 years ago

Aditya narayan sharma ... Your solution is similar to Prof. Roberto Tauraso in AMM ... You should solve it in a different way...

Shivam Sharma - 3 years, 12 months ago

Log in to reply

Well, ultimately, that depends on whether Aditya worked the technique out for himself. I developed much the same solution method, and never read the solutions in AMM. To be quite honest many of the problems on Brilliant have been proposed/solved before; that does not make them uninteresting,

Mark Hennings - 3 years, 12 months ago

Log in to reply

Thank you @Mark Hennings Sir for the comment, @Shivam Sharma As Mark Sir said indeed there are tons of problems on brilliant which have been solved before even they appeared anywhere else but that doesn't restrict us to use the methods by which they are meant to be solved, I just came to know from you that these problem already had a solution maybe somewhere in AMM

Aditya Narayan Sharma - 3 years, 12 months ago

This question is proposed and created by Cornel Ioan Vaneal in AMM ...

Shivam Sharma - 3 years, 12 months ago

@Aditya Narayan Sharma - But you should solve it by different method... Mathematics says that everything should be different and creative...

Shivam Sharma - 3 years, 12 months ago

@Mark Hennings - Person should be honest ... in his works... otherwise he/she will deceive himself/herself ... I think you know about this... Anyways , leave this topic...

Shivam Sharma - 3 years, 12 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...