New Year Integral!

Calculus Level 4

0 1 ln ( x ) ln ( 1 x ) d x = a + π b c \int_{0}^{1}\ln(x)\ln(1-x)\ dx = a + \frac {\pi^b}c

The equation above holds true for integers a a , b b , and c c . Find a + b + c a+b+c .


The answer is -2.

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.

4 solutions

Guilherme Niedu
Jan 2, 2020

I'll go for a long, detailed solution, with the same approach as @Mark Hennings :

I = 0 1 ln ( x ) ln ( 1 x ) d x \large \displaystyle I = \int_{0}^{1} \ln(x) \ln(1-x) dx

By the Maclaurin series of ln ( 1 x ) \ln(1-x) for x 1 |x| \leq 1 and x 1 x \neq 1 :

I = 0 1 ln ( x ) k = 1 x k k d x \large \displaystyle I = - \int_{0}^{1} \ln(x) \sum_{k=1}^{\infty} \frac{x^k}{k} dx

I = k = 1 1 k 0 1 ln ( x ) x k d x \large \displaystyle I = -\sum_{k=1}^{\infty} \frac{1}{k} \int_{0}^{1} \ln(x) \cdot x^k dx

Make x = e u x = e^{-u} , d x = e u d u dx = -e^{-u} du :

I = k = 1 1 k 0 ( u ) e k u ( e u ) d x \large \displaystyle I = -\sum_{k=1}^{\infty} \frac{1}{k} \int_{\infty}^{0} (-u) \cdot e^{-ku} \cdot (-e^{-u}) dx

I = k = 1 1 k 0 u e ( k + 1 ) u d x \large \displaystyle I = \sum_{k=1}^{\infty} \frac{1}{k} \int_{0}^{\infty} u \cdot e^{-(k+1)u} dx

Make u = t k + 1 u = \frac{t}{k+1} , d u = 1 k + 1 d t du = \frac{1}{k+1} dt

I = k = 1 1 k 0 t k + 1 e t 1 k + 1 d t \large \displaystyle I = \sum_{k=1}^{\infty} \frac{1}{k} \int_{0}^{\infty} \frac{t}{k+1} \cdot e^{-t} \frac{1}{k+1} dt

I = k = 1 1 k 1 ( k + 1 ) 2 0 t e t d t \large \displaystyle I = \sum_{k=1}^{\infty} \frac{1}{k} \frac{1}{(k+1)^2} \int_{0}^{\infty} t e^{-t} dt

The integral is equal to Γ ( 2 ) \Gamma(2) , which is equal to 1 ! = 1 1!=1 ( Γ ( s ) \Gamma(s) is the Gamma function )

I = k = 1 1 k 1 ( k + 1 ) 2 \large \displaystyle I = \sum_{k=1}^{\infty} \frac{1}{k} \frac{1}{(k+1)^2}

Expanding in partial fractions:

I = k = 1 1 k 1 k + 1 1 ( k + 1 ) 2 \large \displaystyle I = \sum_{k=1}^{\infty} \frac{1}{k} - \frac{1}{k+1} - \frac{1}{(k+1)^2}

I = k = 1 1 k 1 k + 1 k = 1 1 ( k + 1 ) 2 \large \displaystyle I = \sum_{k=1}^{\infty} \frac{1}{k} - \frac{1}{k+1} - \sum_{k=1}^{\infty} \frac{1}{(k+1)^2}

The first sum will telescope and be equal to 1 1 . In the second sum, make n = k + 1 n = k+1

I = 1 ( n = 2 1 n 2 ) \large \displaystyle I = 1 - \left ( \sum_{n=2}^{\infty} \frac{1}{n^2} \right )

I = 1 ( n = 1 1 n 2 1 ) \large \displaystyle I = 1 - \left ( \sum_{n=1}^{\infty} \frac{1}{n^2} - 1 \right )

I = 2 n = 1 1 n 2 \large \displaystyle I = 2 - \sum_{n=1}^{\infty} \frac{1}{n^2}

The sum is the Riemann zeta function ζ ( s ) \zeta(s) evaluated at s = 2 s=2 , which is equal to π 2 6 \frac{\pi^2}{6}

I = 2 π 2 6 \color{#20A900} \boxed{ \large \displaystyle I = 2 - \frac{\pi^2}{6} }

Thus:

a = 2 , b = 2 , c = 6 a + b + c = 2 \color{#3D99F6} a = 2, b = 2, c = -6 \rightarrow \boxed{ \large \displaystyle a+b+c = -2}

I thought that the integers are positive here.

William Ly - 6 months, 2 weeks ago
Mark Hennings
Jan 1, 2020

Since ln ( 1 x ) = n = 1 1 n x n 0 x < 1 \ln(1-x) = - \sum_{n=1}^\infty \tfrac{1}{n}x^n \hspace{2cm} 0 \le x < 1 we can use the Monotone Convergence Theorem to deduce that 0 1 ln x ln ( 1 x ) d x = n = 1 1 n 0 1 x n ln x d x = n = 1 1 n ( n + 1 ) 2 = n = 1 ( 1 n 1 n + 1 1 ( n + 1 ) 2 ) = 1 n = 2 1 n 2 = 2 ζ ( 2 ) = 2 1 6 π 2 \begin{aligned} \int_0^1 \ln x \ln(1-x)\,dx & =\; -\sum_{n=1}^\infty \frac{1}{n}\int_0^1 x^n \, \ln x\,dx \; = \; \sum_{n=1}^\infty \frac{1}{n(n+1)^2} \; = \; \sum_{n=1}^\infty \left(\frac{1}{n} - \frac{1}{n+1} - \frac{1}{(n+1)^2} \right) \\ & = \; 1 - \sum_{n=2}^\infty \frac{1}{n^2} \; = \; 2 - \zeta(2) \; = \; 2 - \tfrac16\pi^2 \end{aligned} making the answer 2 + 2 6 = 2 2 + 2 - 6 = \boxed{-2} .

Sir could it be done without series expansion?

Vilakshan Gupta - 1 year, 5 months ago

Log in to reply

Well, you could note that 0 1 x u 1 ( 1 x ) v 1 ln x ln ( 1 x ) d x = 2 u v 0 1 x u 1 ( 1 x ) v 1 d x = 2 u v B ( u , v ) \int_0^1 x^{u-1}(1-x)^{v-1}\,\ln x \, \ln(1-x)\,dx \; = \; \frac{\partial^2}{\partial u \partial v}\int_0^1 x^{u-1}(1-x)^{v-1}\,dx \; = \; \frac{\partial^2}{\partial u \partial v}B(u,v) for u , v > 0 u,v > 0 , and hence deduce that 0 1 ln x ln ( 1 x ) d x = 2 u v B ( u , v ) u = v = 1 \int_0^1 \ln x \ln(1-x)\,dx \; = \; \frac{\partial^2}{\partial u \partial v}B(u,v)\Big|_{u=v=1} to make sense of that, you will need to know the values of the digamma function ψ \psi and its derivative at small integer values. I think the series expansion is easier.

Mark Hennings - 1 year, 5 months ago
Naren Bhandari
Jun 3, 2021

Since 0 1 x n 1 log ( 1 x ) d x = H n n \int_0^1 x^{n-1} \log(1-x) dx =-\frac{H_n}{n} and the required integral is lim n 1 + n 0 1 x n 1 log ( 1 x ) = lim n 1 + n H n n = lim n 1 + ( H n n 2 + H n ( 2 ) n ζ ( 2 ) n ) = 2 ζ ( 2 ) \lim_{n\to 1^+}\frac{\partial}{\partial n}\int_0^{1} x^{n-1}\log(1-x)=\lim_{n\to 1^+}\frac{\partial}{\partial n} \frac{H_n}{n}= \lim_{n\to 1^+}\left(\frac{H_n}{n^2}+\frac{H_n^{(2)}}{n}-\frac{\zeta(2)}{n}\right)=2-\zeta(2) Note that 0 1 x n 1 log ( 1 x ) = k 1 0 1 x n + k 1 k d x = k 1 1 k ( n + k ) = H n n \int_0^{1} x^{n-1}\log(1-x) =-\sum_{k\geq 1} \int_0^1x^{n+k-1}{k}dx=-\sum_{k\geq 1} \frac{1}{k(n+k)} =-\frac{H_n}{n} For more approaches for the first logarithmic integrals result see here my solution .

The value of the given integral is 2 π 2 6 2-\dfrac{π^2}{6} . So, a = 2 , b = 2 , c = 6 a=2, b=2,c=-6 , and a + b + c = 2 + 2 6 = 2 a+b+c=2+2-6=\boxed {-2}

Please explain why the answer is 2 π 2 6 2-\dfrac{\pi^2}{6}

Isaac YIU Math Studio - 1 year, 5 months ago

1 pending report

Vote up reports you agree with

×

Problem Loading...

Note Loading...

Set Loading...