Calculus and Patterns

Calculus Level 5

The area bounded by the curves ( ln x ) n (\ln{x})^{n} and ( ln x ) n 1 (\ln{x})^{n-1} is of the form ( 1 ) n + 1 ( a n e b n ) (-1)^{n+1}\left(a_n e-b_n\right) , where n N n\in\mathbb{N} and a n a_n and b n b_n are integers.

Let

  • lim n a n b n = α e \displaystyle \lim_{n \to \infty} \frac{a_n}{b_n}=\frac{\alpha}{e}
  • Sum of the digits of a 25 a_{25} be β \beta
  • Sum of distinct prime factors of b 20 b_{20} be γ \gamma

Submit your answer as α + β + γ \alpha+\beta+\gamma . You may use a calculator.

Hint: Find the closed forms of a n a_n and b n b_n .


The answer is 190.

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

Leonel Castillo
Jan 29, 2018

Finding an Integral expression for the area:

We first need to find a way to write this area as an integral. For the bounds of integration solve log n x = log n 1 x \log^n x = \log^{n-1} x . Let u = log x u = \log x to get the polynomial equation u n u n 1 = 0 u n 1 ( u 1 ) = 0 u = 0 , 1 u^n - u^{n-1} = 0 \implies u^{n-1} (u - 1) = 0 \implies u=0,1 for n > 1 n>1 . If n = 1 n=1 no area is bound between the curves. So we have to solve log x = 0 , 1 x = 1 , e \log x = 0,1 \implies x = 1, e . Also, as the function ranges from 0 0 to 1 1 we will have that log n 1 x > log n x \log^{n-1} x > \log^n x so an expression for this area is:

1 e log n 1 x log n x d x \large \int_1^e \log^{n-1} x - \log^{n} x dx .


Introducing a new sequence:

Computing this integral ended up being remarkably difficult but as the problem hints, it is possible to write it in terms of two arithmetic functions. My procedure is as follows: First, let u = log x u = \log x . Then d x = e u d u dx = e^u du and thus the integral is equal to 0 1 x n 1 e x x n e x d x \large \int_0^1 x^{n-1} e^x - x^{n} e^x dx . To solve this, I will introduce the following sequence. Let s n = 0 1 x n e x d x s_n = \int_0^1 x^n e^x dx . This sequence has the property that the integral we want to compute is just s n 1 s n s_{n-1} - s_n so if we can find a general formula for s n s_n we can tackle the original problem.


Recurrence relation for s n s_n

We have that s n = 0 1 x n e x d x s_n = \int_0^1 x^n e^x dx . Let u = x n , d v = e x u = x^n, dv = e^x so d u = n x n 1 , d v = e x du = nx^{n-1}, dv = e^x to get, by integration by parts, that this is equal to e n 0 1 x n 1 e x d x = e n s n 1 e - n \int_0^1 x^{n-1} e^x dx = e - n s_{n-1} . So the recurrence relation is s n = e n s n 1 s_n = e - n s_{n-1} . We can completely define the sequence in with this recurrence relation if we only compute s 1 = 0 1 x e x d x = 1 s_1 = \int_0 ^1 x e^x dx = 1 .

I now show some computed values: s 2 = e 2 s 3 = 2 e + 6 s 4 = 9 e 24 s 5 = 44 e + 120 s_2 = e - 2 \\ s_3 = -2e + 6 \\ s_4 = 9e - 24 \\ s_5 = -44e + 120 . At a quick glance one can immediately notice that the integer term is ( 1 ) n + 1 n ! (-1)^{n+1} n! . But the e e term is much harder to find by just looking, some extra work has to be put in. If one writes out what the e e term's coefficient is they will find the following expression: 1 n + n ( n 1 ) n ( n 1 ) ( n 2 ) + . . . ± n ! 2 ! = 1 ( 1 ) n + 1 k = 2 n 1 n ! k ! ( 1 ) k 1 - n + n(n-1) - n(n-1)(n-2) + ... \pm \frac{n!}{2!} = 1 - (-1)^{n+1}\sum_{k=2}^{n-1} \frac{n!}{k!} (-1)^k

With this we have finally solved s n = ( 1 ) n + 1 n ! + e ( 1 ( 1 ) n + 1 k = 2 n 1 n ! k ! ( 1 ) k ) \large s_n = (-1)^{n+1} n! + e \left( 1 - (-1)^{n+1}\sum_{k=2}^{n-1} \frac{n!}{k!} (-1)^k \right )


General formula for the area

With the previous result finding the area is just an exercise in algebra but we end up with: ( 1 ) n + 1 ( ( k = 2 n 1 n ! k ! ( 1 ) k + k = 2 n 2 ( n 1 ) ! k ! ( 1 ) k ) e ( n ! + ( n 1 ) ! ) ) \Large (-1)^{n+1} \left( \left( \sum_{k=2}^{n-1} \frac{n!}{k!} (-1)^k + \sum_{k=2}^{n-2} \frac{(n-1)!}{k!} (-1)^k \right)e - (n! + (n-1)!) \right) with, specifically:

a n = k = 2 n 1 n ! k ! ( 1 ) k + k = 2 n 2 ( n 1 ) ! k ! ( 1 ) k a_n = \sum_{k=2}^{n-1} \frac{n!}{k!} (-1)^k + \sum_{k=2}^{n-2} \frac{(n-1)!}{k!} (-1)^k

b n = n ! + ( n 1 ) ! = ( n 1 ) ! ( n + 1 ) b_n = n! + (n-1)! = (n-1)!(n+1)


Solving for α \alpha

For the fraction limit, I will consider separately both of the sums that make up a n a_n . Diving the first part of a n a_n by b n b_n yields n n + 1 k = 2 n 1 ( 1 ) k k ! \frac{n}{n+1} \sum_{k=2}^{n-1} \frac{(-1)^k}{k!} . As n n \to \infty the fraction outside goes to 1 and the sum goes to e 1 e^{-1} so this part of the limit equals 1 e \frac{1}{e} . Now notice that for the second part of a n a_n we just get 1 n + 1 k = 2 n 2 ( 1 ) k k ! \frac{1}{n+1} \sum_{k=2}^{n-2} \frac{(-1)^k}{k!} . The sum approaches the constant e 1 e^{-1} but the fraction outside goes to 0 so this part vanishes. Leaving us with the limit being 1 e \frac{1}{e} and thus α = 1 \alpha = 1 .


Solving for γ \gamma

Notice that b 20 = 19 ! ( 21 ) b_{20} = 19! (21) . 21 is not a prime number, and 19 ! 19! will contain all the prime numbers from 1 to 19 so γ = 2 + 3 + 5 + 7 + 11 + 13 + 17 + 19 = 77 \gamma = 2 + 3 + 5 + 7 + 11 + 13 + 17 + 19 = 77 .

Solving for β \beta

Here I did not figure out any clever way to compute this in an easy way. I used a computer program to compute this using the formula I gave for a n a_n . The result is β = 112 \beta = 112 .

a n = ! ( n + 1 ) n a_n=\dfrac{!(n+1)}{n}

! n !n means subfactorial of n n .

Digvijay Singh - 3 years, 4 months ago

Log in to reply

Ah, I did not know of that function. But comparing the values it equals the expression I gave.

Leonel Castillo - 3 years, 4 months ago

Log in to reply

It also equals 1 n [ ( n + 1 ) ! e ] \dfrac{1}{n}\left[ \dfrac{(n+1)!}{e} \right]

Now that's a function you know of.

Digvijay Singh - 3 years, 4 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...