Let f ( x ) be continuous and differentiable everywhere, and assume that f ( x ) is analytic. Furthermore assume that for all non-negative integers n , we have f ( n ) ( 0 ) = n , where f ( n ) ( x ) denotes the n th derivative of f at x . Find f ( 1 ) .
Give your answer to 3 decimal places.
Bonus : As a fun extension, calculate f ( 1 ) if we say that for non-negative n , we have f ( n ) ( 0 ) = n k for some defined positive integer k .
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.
Perfect solution!
We can see that n = 1 ∑ ∞ ( n − 1 ) ! x n = x n = 0 ∑ ∞ n ! x n = x ⋅ e x .
It is easy to show by induction that f ( n ) ( x ) = ( x + n ) e x . Then f ( n ) ( 0 ) = n and f ( 1 ) = e .
In general, for selected k , we have f ( x ) = T k e x , where T k is the k th Touchard Polynomial (the proof is essentially the same as above).
You should clarify that n is a non-negative integer.
In your solution, the "0" isn't something that we randomly add on, but that we're using f ( 0 ) = 0 .
Otherwise, f ( x ) = 1 0 0 0 0 0 0 0 0 + n = 1 ∑ ∞ n ! n x n would also satisfy the conditions.
Log in to reply
Thanks for the feedback! Problem was edited accordingly. And yes, I didn't explicitly state it, but that is where that zero came from (just separated it to make the final summation more evident).
Also, doesn't the assumption of the problem (for all non-negative integers n , f ( n ) ( 0 ) = n ), imply that f ( 0 ) = 0 , the case for n = 0 ?
What if we take a nonzero function g ( x ) with g ( n ) ( 0 ) = 0 for all non-negative n and add it to your f ( x ) ? This problem is flawed as it stands... f ( 1 ) could be anything.
Log in to reply
I'm not sure this reasoning is valid... let's say g ( x ) = f ( x ) , then f ( x ) + g ( x ) = 2 f ( x ) = 2 x e x , but this function does not meet the criteria of our desired function. Namely, f ′ ( x ) = ( 2 x + 2 ) e x , and so f ′ ( 0 ) = 2 , not 1 .
Log in to reply
No, I'm talking about adding a nonzero g ( x ) with g ( n ) ( 0 ) = 0 for all non-negative n ; f ( x ) does not have that property.
Log in to reply
@Otto Bretscher – I misread the first time, and just noticed that... So we can take:
f ( x ) = x e x + e − 1 / x 2 , x = 0
f ( x ) = x e x , x = 0
This satisfies the conditions of the problem. However, this function is no longer analytic at 0. As I recall, all flat functions (functions g ( x ) such that g ( n ) ( x ) = 0 for every non-negative n for some x ) are not analytical at the point x . I was assuming that f ( x ) is analytic.
Log in to reply
@Anthony Kirckof – I believe your last remark ("As I recall,...") is not correct: The function might be analytic at x but "take off" later.
For example, we might define f ( x ) = x e x for x ≤ 2 1 and f ( x ) = x e x + e − x − 0 . 5 1 for x > 2 1 , which is analytic at 0 (but fails to be analytic at x = 2 1 ).
@Otto Bretscher – Problem now edited to include the statement of being analytic.
Log in to reply
@Anthony Kirckof – I'm afraid that "analytic at 0" is not enough... just require the function to be analytic.
Problem Loading...
Note Loading...
Set Loading...
If we write out the Taylor Polynomial, we find that
f ( x ) = 0 + n = 1 ∑ ∞ n ! n x n = n = 1 ∑ ∞ ( n − 1 ) ! x n .
Using the convergence test, we see that f converges everywhere, so we just need to calculate f ( 1 ) as above:
f ( 1 ) = n = 1 ∑ ∞ ( n − 1 ) ! 1 .
This sum converges (by definition) to e .
To solve the generalization, let k be a defined positive integer. We start with the Taylor Polynomial once again:
f ( x ) = n = 1 ∑ ∞ n ! n k x n .
We proceed by induction, and claim that for the given integer k , f ( 1 ) = B k e , where B k is the k th Bell number. The case of k = 1 is covered above. Assuming that f ( 1 ) = B k e for 1 ≤ k ≤ j for some positive integer j , the Taylor polynomial for j + 1 then becomes:
f ( 1 ) = n = 1 ∑ ∞ n ! n j + 1 = 0 + n = 1 ∑ ∞ ( n + 1 ) ! ( n + 1 ) j + 1 = n = 1 ∑ ∞ n ! ( n + 1 ) j = n = 1 ∑ ∞ i = 0 ∑ j n ! ( i j ) n i = i = 0 ∑ j n = 1 ∑ ∞ n ! ( i j ) n i = i = 0 ∑ j ( i j ) n = 1 ∑ ∞ n ! n i = i = 0 ∑ j ( i j ) B i e .
Here, we note the recurrence formula for Bell numbers, B j + 1 = i = 0 ∑ j ( i j ) B i . With that in mind, we proceed and find that f ( 1 ) = B j + 1 e , completing the proof. Thus f ( 1 ) = B k e .