Kenny's Identity

Calculus Level 5

L = lim x 0 n = 1 r = 0 ( 1 ) r n + r ( n + r r ) x n + r 1 L=\lim_{x\to0}\sum_{n=1}^{\infty}\sum_{r=0}^{\infty}\frac{(-1)^r}{n+r}\dbinom{n+r}{r}x^{n+r-1}

Find the value of 10000 L \lfloor10000L\rfloor .


This problem is original.


The answer is 10000.

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

Kenny Lau
Oct 12, 2015

L \quad L

= lim x 0 n = 1 r = 0 ( 1 ) r n + r ( n + r r ) x n + r 1 \displaystyle =\lim_{x\to0}\sum_{n=1}^{\infty}\sum_{r=0}^{\infty}\frac{(-1)^r}{n+r}\dbinom{n+r}{r}x^{n+r-1}

= lim x 0 1 x n = 1 x n r = 0 ( 1 ) r n + r ( n + r r ) x r \displaystyle =\lim_{x\to0}\frac1x\sum_{n=1}^{\infty}x^n\sum_{r=0}^{\infty}\frac{(-1)^r}{n+r}\dbinom{n+r}{r}x^r

= lim x 0 1 x n = 1 x n r = 0 ( 1 ) r n ( n + r 1 r ) x r \displaystyle =\lim_{x\to0}\frac1x\sum_{n=1}^{\infty}x^n\sum_{r=0}^{\infty}\frac{(-1)^r}n\dbinom{n+r-1}{r}x^r [See (1)]

= lim x 0 1 x n = 1 x n n r = 0 ( n + r 1 r ) ( x ) r \displaystyle =\lim_{x\to0}\frac1x\sum_{n=1}^{\infty}\frac{x^n}n\sum_{r=0}^{\infty}\dbinom{n+r-1}{r}(-x)^r

= lim x 0 1 x n = 1 x n n 1 ( 1 + x ) n \displaystyle =\lim_{x\to0}\frac1x\sum_{n=1}^{\infty}\frac{x^n}n\frac1{(1+x)^n} [See (2)]

= lim x 0 1 x n = 1 1 n x n ( 1 + x ) n \displaystyle =\lim_{x\to0}\frac1x\sum_{n=1}^{\infty}\frac1n\frac {x^n}{(1+x)^n}

= lim x 0 1 x n = 1 1 n ( x 1 + x ) n \displaystyle =\lim_{x\to0}\frac1x\sum_{n=1}^{\infty}\frac1n\left(\frac x{1+x}\right)^n

= lim x 0 1 x ( ln ( 1 x 1 + x ) ) \displaystyle =\lim_{x\to0}\frac1x\left(-\ln\left(1-\frac x{1+x}\right)\right) [See (3)]

= lim x 0 1 x ( ln ( 1 1 + x ) ) \displaystyle =\lim_{x\to0}\frac1x\left(-\ln\left(\frac1{1+x}\right)\right)

= lim x 0 1 x ln ( 1 + x ) \displaystyle =\lim_{x\to0}\frac1x\ln(1+x)

= lim x 0 ln ( 1 + x ) 1 x \displaystyle =\lim_{x\to0}\ln(1+x)^{\frac1x} [See (4)]

= ln e \displaystyle =\ln e

= 1 \displaystyle =1


(1): 1 n + r ( n + r r ) \dfrac1{n+r}\dbinom{n+r}{r}

= ( n + r ) ! n ! r ! ( n + r ) =\dfrac{(n+r)!}{n!r!(n+r)}

= ( n + r 1 ) ! n ! r ! =\dfrac{(n+r-1)!}{n!r!}

= 1 n ( n + r 1 ) ! ( n 1 ) ! r ! =\dfrac1n\dfrac{(n+r-1)!}{(n-1)!r!}

= 1 n ( n + r 1 r ) =\dfrac1n\dbinom{n+r-1}{r}


(2): ( 1 x ) n r = 0 ( n + r 1 r ) x r (1-x)^{-n}\equiv\displaystyle\sum_{r=0}^{\infty}\dbinom{n+r-1}{r}x^r ( 1 < x < 1 -1<x<1 ), substituting x : x x:-x .

This can be proven using induction.

For a non-induction but informal approach, consider 1 1 x = 1 x 1 x = 1 + x + x 2 + x 3 + \dfrac1{1-x}=\dfrac{1-x^\infty}{1-x}=1+x+x^2+x^3+\cdots .

Then 1 ( 1 x ) n = ( 1 + x + x 2 + x 3 + ) n \dfrac1{(1-x)^n}=(1+x+x^2+x^3+\cdots)^n .

Consider the coefficient of x r x^r in this infini-nomial (binomial, trinomial, ...) expansion.

This is equivalent to choosing a non-negative number from the first bracket, then one from the second, until the n-th bracket, and then adding them to r r .

Consider dividing r r stars by n 1 n-1 bars instead, which is still equivalent.

There are in total n + r 1 n+r-1 objects in which r r are stars, hence the coefficient is ( n + r 1 r ) \dbinom{n+r-1}{r} .

See Combinations with Repetition for further information.


(3): ln ( 1 1 x ) = ln ( 1 x ) = n = 1 x n n \displaystyle\ln\left(\frac1{1-x}\right)=-\ln(1-x)=\sum_{n=1}^{\infty}\frac{x^n}n .

Proof:

n = 1 x n n \quad\displaystyle\sum_{n=1}^{\infty}\frac{x^n}n

= n = 1 0 x x n 1 d x =\displaystyle\sum_{n=1}^{\infty}\int_0^xx^{n-1}\mathrm dx

= 0 x n = 1 x n 1 d x =\displaystyle\int_0^x\sum_{n=1}^{\infty}x^{n-1}\mathrm dx

= 0 x n = 0 x n d x =\displaystyle\int_0^x\sum_{n=0}^{\infty}x^n\mathrm dx

= 0 x 1 1 x d x =\displaystyle\int_0^x\frac1{1-x} dx [See (2)]

= [ ln 1 1 x ] 0 x =\displaystyle\left[\ln\left|\frac1{1-x}\right|\right]_0^x

= ln 1 1 x =\displaystyle\ln\left|\frac1{1-x}\right|


(4): lim x 0 ( 1 + x ) 1 x = e \lim_{x\to0}(1+x)^{\frac1x}=e

Such is the definition of e e .

Nice little easy problem indeed, as Kartik observed.

I started with the substitution n + r = m n+r=m . Now m > r ( 1 ) r ( m r ) x m 1 m \sum_{m>r}(-1)^r{m \choose r}\frac{x^{m-1}}{m} = m = 1 ( r = 0 m 1 ( 1 ) r ( m r ) ) x m 1 m =\sum_{m=1}^{\infty}\left(\sum_{r=0}^{m-1}(-1)^r{m \choose r}\right)\frac{x^{m-1}}{m} = m = 1 ( x ) m 1 m = ln ( x + 1 ) x =\sum_{m=1}^{\infty}\frac{(-x)^{m-1}}{m}=\frac{\ln(x+1)}{x} ; in the last step I'm using the Taylor series of ln ( x + 1 ) \ln(x+1) . The limit as x 0 x\rightarrow{0} is 1.

Otto Bretscher - 5 years, 8 months ago

Why makes things so complicated? For n + r 1 0 n + r - 1 \ne 0 , when x 0 x\to 0 , x n + r 1 0 x^{n+r-1} \to 0 . So we only need to consider when n = r 1 = 0 n = r - 1 = 0 which occurs when n = 1 , r = 0 n=1,r= 0 . When x x approaches 0 from the right hand side x n + r 1 x^{n+r-1} approaches 1 when n = 1 , r = 0 n = 1, r = 0 . Thus the series equals to ( 1 ) 0 1 + 0 ( 1 + 0 1 ) = 1 \dfrac{(-1)^0}{1+0} \dbinom{1+0}1 = 1 .

Pi Han Goh - 5 years, 7 months ago

Log in to reply

Because summation?

Kenny Lau - 5 years, 7 months ago

Log in to reply

What do you mean?

Pi Han Goh - 5 years, 7 months ago

Log in to reply

@Pi Han Goh @Kenny Lau I know it's been two years, but it works because all the terms in the sums are always finite. You'd have to check if you had a case like zero divided by zero

Pierre Stöber - 3 years, 2 months ago

Nice little easy problem - what I liked the most though is surely the solution. Very well written! That must have been quite a work. +1!

Kartik Sharma - 5 years, 8 months ago

Log in to reply

Thank you!

Kenny Lau - 5 years, 8 months ago

Awesome question and an equally awesome solution.Upvoted!

Athiyaman Nallathambi - 5 years, 8 months ago

Log in to reply

Thank you!

Kenny Lau - 5 years, 8 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...