e x + e x + e x + e x + ⋅ ⋅ ⋅ = 1 0
Find the value of x satisfying the infinitely nested function above.
Give your answer to 3 decimal places.
Enter 0.666 if you come to the conclusion that no such x exists.
Clarification : e denotes Euler's number , e ≈ 2 . 7 1 8 2 8 .
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.
@Otto Bretscher ,we really liked your comment, and have converted it into a solution. If you subscribe to this solution, you will receive notifications about future comments.
I approached this slightly differently than the other solution, so I may as well post it:
e x + e x + e x + e x + e = 1 0 e x + 1 0 = 1 0 x + 1 0 = ln ( 1 0 ) x = ln ( 1 0 ) − 1 0 x ≅ − 7 . 6 9 7
The "replace iteratively" procedure assumes that a limit exists. We should first prove why the limit exists, before we can conclude that it is equal to a certain value.
Are you sure that the tower is converging to 10? In problems like this, finding the fixed point is usually the easy part but proving (or disproving) convergence is the hard(er) but also the more interesting part.
Log in to reply
To be truthful, I was never taught how to approach problems of this manner before I joined Brilliant. It was only a few months ago I saw the method I displayed above done on another similar problem, and since then I've used this again and again to approach problems like this.
But, when it comes to proving convergence for problems that are nested like this, I don't know where I would start, but I'd love for you to guide me towards the right direction.
Log in to reply
I will work it out (as a comment to Arthur's solution)... give me a little time.
Log in to reply
@Otto Bretscher – Thank you tremendously :)
Log in to reply
@Andrew Tawfeek – Ok my solution is complete...please let me know whether it makes sense.
e x + 1 0 = 1 0 x ≈ − 7 . 6 9 7
but, as x ≈ − 7 . 6 9 7 , the value of the nested function above is approaching 0 which is not equal to 10.
Proof:
by just trying to substitute and observe the obtained results as x ≈ − 7 . 6 9 7 ,
as x < 0 < e , x + e will turn out negative.
0 < e x + e = e − s o m e t h i n g < 1 0 < e x + e x + e = e x + e x + e = e x + e − s o m e t h i n g < 1 ⋮
it completely suffices that as x ≈ − 7 . 6 9 7 , the infinite nested function e x + e x + e x + e x + ⋅ ⋅ ⋅ will approach zero and not equal 10.
∴ there is no such x exist.
e x + e x + e x + e x + e x + e x + e x + . . . = 1 0
Start by applying the natural log on both sides:
x + e x + e x + e x + e x + e x + e x + e x + . . . = ln ( 1 0 )
The part that comes after the x+ is the same as what we started with, so we can replace that with 10.
x + 1 0 = ln ( 1 0 )
Solve for x.
x = ln ( 1 0 ) − 1 0 = − 7 . 6 9 7
I assume you are using the iteration a 0 = 0 and a n + 1 = e x + a n = e ln ( 1 0 ) − 1 0 + a n . If so, I'm afraid your tower is not going to converge to 10 but to − W ( − e 1 0 1 0 ) ≈ 0 . 0 0 0 4 5 4 2 , where W is Lambert's function . I can provide a detailed explanation upon request.
Log in to reply
Yes sir can you please explain it ! I am curious about convergence s
Log in to reply
As Arthur and Andrew point out, if there is a solution at all, it must be x = ln ( 1 0 ) − 1 0 . However, we must examine whether the power tower with x = ln ( 1 0 ) − 1 0 does indeed converge to 10. It will turn out that it doesn't, so that this problem has no solutions .
It is convenient to introduce the iteration function f ( t ) = e x + t = e ln ( 1 0 ) − 1 0 + t , an exponential growth function. Now we can write our iteration simply as a n + 1 = f ( a n ) , with the seed a 0 = 0 , and we are asked to find out whether lim n → ∞ a n = 1 0 .
It turns out that f ( t ) has two fixed points , where f ( t ) = t , namely, b = − W ( − e 1 0 1 0 ) ≈ 0 . 0 0 0 4 5 4 2 and c = 1 0 .
I suggest that the reader graphs the functions f ( t ) and t on the same axes to get a sense for what's going on here. Draw a cobweb for various initial values if you are familiar with the concept...that helps a lot in really understanding the issue of convergence.
Now we can show by induction that a n < b for all n . We know that a 0 = 0 < b . Assuming a n < b we conclude that a n + 1 = f ( a n ) < f ( b ) = b since f ( t ) is increasing.
Since a n < b < 0 . 0 0 0 5 for all n , the limit of a n cannot be 10. It turns out that the limit is actually b , but that observation is not crucial for our solution (left as an exercise).
Let me mention that this convergence behaviour is closely related to the fact that f ′ ( b ) = f ( b ) = b < 1 while f ′ ( 1 0 ) = f ( 1 0 ) = 1 0 > 1 : The equilibrium b is stable while c = 1 0 isn't.
Note to @Calvin Lin : Maybe you will be kind enough to convert this comment to a solution if you find it useful. I got the problem "wrong". ;)
Log in to reply
@Otto Bretscher – Thank you for commenting on this! Apparently math isn't always as easy as it appears.
Log in to reply
@Arthur von Da – Es ist besser so...This convergence stuff is interesting!
@Arthur von Da – I will take the liberty to change the wording of the problem to implement the correct answer.
@Arthur von Da – I think this twist makes your problem even more interesting!
@Otto Bretscher – I see now why it doesn't converge to 10, thank you!
I feel like I understand more about these kinds of problems now. After reading your solution, I looked back at my own work for this problem, then facepalmed over something that had flown completely over my head.
Starting off with f ( t ) = e l n ( 1 0 ) − 1 0 + t , a n + 1 = f ( a n ) , and a 0 = 0 :
a 0 = 0 a 1 = e l n ( 1 0 ) − 1 0 + ( 0 ) = e 1 0 1 0 e ( 0 ) ≅ 0 . 0 0 0 4 5 4 a 2 = 0 . 0 0 0 4 5 4 e a 1 ≅ 0 . 0 0 0 4 5 4 a 3 ≅ 0 . 0 0 0 4 5 4 . . .
Just evaluating the first few iterations showed me this isn't even in arm's reach of 10!
Although I do have a question. After each iteration, I've noticed that 1 < e t < 2 . For example, during a 2 , e e 1 0 1 0 = 1 . 0 0 0 4 5 4 1 0 2 3 7 gets multiplied with e 1 0 1 0 ≅ 0 . 0 0 0 4 5 4 . Wouldn't the fact that e t stays greater than 1 cause this iterative function to grow?
Log in to reply
@Andrew Tawfeek – I'm not sure I fully understand what you mean.
There are two separate issues here:
The iteration function f ( t ) is increasing as it is an exponential growth function.
The sequence a n is increasing as well, meaning that a n + 1 > a n for all n ; this is true because f ( t ) > t for t < b . (But if the seed a 0 were above b and below c , then the sequence a n would be decreasing.)
A great way to visualize and understand these things is in terms of cobwebs . Draw three cobwebs: One for a seed below b , as in our problem, one for a seed between b and c , and finally one for a seed above c = 1 0 .
By the way: It's great that you did some numerical experiments... thanks!
Log in to reply
@Otto Bretscher – Oh I understand now! I had misunderstood your previous post and thought that the sequence was decreasing. My apologies!
@Otto Bretscher – Thanks a lot, now I understood what is really going on here, what an oversight
Arthur, I did the same way you did, but why is it giving 0.666 as the answer, I am confused due to this. Please clarify
Log in to reply
Read my solution listed as a comment below. It turns out that ln ( 1 0 ) − 1 0 isn't the solution after all.
Problem Loading...
Note Loading...
Set Loading...
Relevant wiki: Nested Functions
As Arthur and Andrew point out, if there is a solution at all, it must be x = ln ( 1 0 ) − 1 0 . However, we must examine whether the power tower with x = ln ( 1 0 ) − 1 0 does indeed converge to 10. It will turn out that it doesn't, so that this problem has no solutions .
It is convenient to introduce the iteration function f ( t ) = e x + t = e ln ( 1 0 ) − 1 0 + t , an exponential growth function. Now we can write our iteration simply as a n + 1 = f ( a n ) , with the seed a 0 = 0 , and we are asked to find out whether lim n → ∞ a n = 1 0 .
It turns out that f ( t ) has two fixed points , where f ( t ) = t , namely, b = − W ( − e 1 0 1 0 ) ≈ 0 . 0 0 0 4 5 4 2 and c = 1 0 .
I suggest that the reader graphs the functions f ( t ) and t on the same axes to get a sense for what's going on here. Draw a cobweb for various initial values if you are familiar with the concept...that helps a lot in really understanding the issue of convergence.
Now we can show by induction that a n < b for all n . We know that a 0 = 0 < b . Assuming a n < b we conclude that a n + 1 = f ( a n ) < f ( b ) = b since f ( t ) is increasing.
Since a n < b < 0 . 0 0 0 5 for all n , the limit of a n cannot be 10. It turns out that the limit is actually b , but that observation is not crucial for our solution (left as an exercise).
Let me mention that this convergence behavior is closely related to the fact that f ′ ( b ) = f ( b ) = b < 1 while f ′ ( 1 0 ) = f ( 1 0 ) = 1 0 > 1 : The equilibrium b is stable while c = 1 0 isn't.