A calculus problem by Akeel Howell

Calculus Level 5

Let x 0 , x 1 , x 2 , \ x_0, x_1, x_2, \cdots \ be a sequence such that x 0 = 1 x_0 = 1 and for n 0 n \geq 0 , x n + 1 = ln ( e x n x n ) \ x_{n+1} = \ln{(e^{x_n} - x_n)} .

If A = 1 + N = 0 ln ( e n = 0 N x n ) , \displaystyle A = 1 + \sum_{N=0}^{\infty}{\ln{ \left( e - \sum_{n = 0}^{N}{x_n} \right) }},

find 1000 A \displaystyle \lfloor {1000A} \rfloor .


The answer is 1718.

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.

2 solutions

x n + 1 = ln ( e x n x n ) e x n + 1 = e x n x n x n = e x n e x n + 1 \begin{aligned} x_{n+1} & = \ln \left(e^{x_n} - x_n \right) \\ e^{x_{n+1}} & = e^{x_n} - x_n \\ \implies x_n & = e^{x_n} - e^{x_{n+1}} \end{aligned}

Therefore, we have

A = 1 + N = 0 ln ( e n = 0 N x n ) = 1 + N = 0 ln ( e n = 0 N ( e x n e x n + 1 ) ) = 1 + N = 0 ln ( e e x 0 + e x N + 1 ) = 1 + N = 0 ln ( e e 1 + e x N + 1 ) = 1 + N = 0 ln ( e x N + 1 ) = 1 + N = 0 x N + 1 = lim M N = 0 M x N = lim M ( e x 0 e x M + 1 ) See note: lim k x k = 0 = e 1 1.718 \begin{aligned} A & = 1 + \sum_{N=0}^\infty \ln \left(e-\sum_{n=0}^N x_n \right) \\ & = 1 + \sum_{N=0}^\infty \ln \left(e-\sum_{n=0}^N \left(e^{x_n} - e^{x_{n+1}}\right)\right) \\ & = 1 + \sum_{N=0}^\infty \ln \left(e-e^{x_0} + e^{x_{N+1}}\right) \\ & = 1 + \sum_{N=0}^\infty \ln \left(e-e^1 + e^{x_{N+1}}\right) \\ & = 1 + \sum_{N=0}^\infty \ln \left(e^{x_{N+1}}\right) \\ & = 1 + \sum_{N=0}^\infty x_{\color{#3D99F6}N+1} \\ & = \lim_{M \to \infty} \sum_{N=0}^M x_{\color{#D61F06}N} \\ & = \lim_{M \to \infty} \left( e^{x_0} - e^{x_{M+1}} \right) & \small \color{#3D99F6} \text{See note: }\lim_{k \to \infty} x_{k} = 0 \\ & = e - 1 \approx 1.718 \end{aligned}

1000 A = 1718 \implies \lfloor 1000A\rfloor = \boxed{1718}


Note: From x n + 1 = ln ( e x n x n ) < ln ( e x n ) = x n x_{n+1} = \ln \left(e^{x_n} - x_n \right) < \ln \left(e^{x_n} \right) = x_n . This implies that x n x_n is decreasing from x 0 = 1 x_0 = 1 and since ln ( e x n x n ) 0 \ln \left(e^{x_n} - x_n \right) \ge 0 , lim n x n = 0 \displaystyle \implies \lim_{n \to \infty} x_n = 0 .

Very nice solution!

James Wilson - 3 years, 7 months ago
Akeel Howell
Oct 6, 2017

We have that x 0 = 1 x_0 = 1 and x n + 1 = ln ( e x n x n ) x_{n+1} = \ln{(e^{x_n} - x_n)} . Thus, the recursive definition of the sequence gives

x 0 = 1 x 1 = ln ( e 1 ) x 2 = ln ( e 1 ln ( e 1 ) ) x 3 = ln ( e 1 ln ( e 1 ) ln ( e 1 ln ( e 1 ) ) ) x_0 = 1 \ \implies \ x_1 = \ln{(e-1)} \ \implies \ x_2 = \ln{(e-1-\ln{(e-1)})} \ \implies \ x_3 = \ln{(e-1-\ln{(e-1)} - \ln{(e-1-\ln{(e-1)})})} \ \cdots .

Hence, we see that x n + 1 = ln ( e x 0 x 1 x 2 x n ) x_{n+1} = \ln{(e - x_0 - x_1 - x_2 - \cdots - x_n)} .

On expansion of A A , we see that N = 0 ln ( e n = 0 N x n ) = ln ( e 1 ) + ln ( e ( 1 + ln ( e 1 ) ) ) + ln ( e ( 1 + ln ( e 1 ) + ln ( e 1 ln ( e 1 ) ) ) ) + \displaystyle \sum_{N = 0}^{\infty}{\ln{\left( e - \sum_{n = 0}^N{x_n} \right)}} = \ln{(e-1)} + \ln{(e - (1+\ln{(e-1)}))} + \ln{(e - (1 + \ln{(e-1)} + \ln{(e - 1 - \ln{(e - 1)}})))} + \cdots which is the sum of all the terms of the sequence, with the exception of the first.

Since x n + 1 = ln ( e x n x n ) x_{n+1} = \ln{(e^{x_n} - x_n)} , we have that ln ( e x n x n ) = ln ( e x 0 x 1 x 2 x n ) \ \ln{(e^{x_n} - x_n)} = \ln{(e - x_0 - x_1 - x_2 - \cdots - x_n)} .

So x 0 + x 1 + x 2 + = n x n = 1 + n ln ( e x n x n ) = e lim n e x n + 1 = e 1 \displaystyle x_0 + x_1 + x_2 + \cdots = \sum_n x_n = 1+\sum_n {\ln{(e^{x_n} - x_n)}} = e - \lim_{n \to \infty}{e^{x_{n+1}}} = e - 1 .

1 + N = 0 ln ( e n = 0 N x n ) = 1 + e 2 = e 1 \therefore 1 + \displaystyle \sum_{N = 0}^{\infty}{\ln{\left( e - \sum_{n = 0}^N{x_n} \right)}} = 1 + e - 2 = e - 1 .

Thus, 1000 A = 1718 \lfloor {1000A} \rfloor = \ \boxed{1718} .

@Akeel Howell - Nice problem!

I think you could also use x n = e x n e x n + 1 x_n = e^{x_n}-e^{x_{n+1}} to have the inner sum telescope, and the whole sum just becomes x N + 1 \sum{x_{N+1}} and then use the identity to have this sum telescope again.

Christopher Criscitiello - 3 years, 8 months ago

Log in to reply

@Christopher Criscitiello Agreed.

That makes for a beautiful solution. I think @Chew-Seong Cheong does a great job with this.

Akeel Howell - 3 years, 8 months ago

Very nice problem, thanks so much!

James Wilson - 3 years, 7 months ago

Log in to reply

Sure thing. Thank you for solving it!

Akeel Howell - 3 years, 7 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...