Inspired by Archit Boobna

Calculus Level 4

For every real number z z , define the infinite nested logarithm as the limit of the series

a 1 = ln z a 2 = ln ln z z a 3 = ln ln ln z z z a_1 = \ln z \\ a_ 2 = \ln_{ \ln z } z \\ a_3 = \ln_{ \ln _{ \ln z } z } z \\ \vdots

where a n + 1 = ln a n z a_{n+1} = \ln_{a_n} z .

For what value of Z Z , does the limit of the sequence a n a_n exist for all z > Z z > Z ?


Inspiration, see solution comment

π π \pi ^ \pi e π e^{\pi } π e \pi ^ e e e e^ e

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

Archit Boobna
Apr 13, 2015

To find minimum value of "z", we may use the minimum value of a n { a }_{ n } , as each a n { a }_{ n } decreases as z decreases, so minimum value of a n { a }_{ n } will occur at minimum value of z making it real.

(Why is this true? Differentiate any a n { a }_{ n } with respect to z, it will always be positive if real.)

Let b n = a n + 1 { b }_{ n }={ a }_{ n+1 } . Now all the terms of these series are common. So for them to exist, they both need to exist together. Let's assume b n { b }_{ n } exists for b n 1 { b }_{ n-1 } greater than a certain value k.

Since b n 1 { b }_{ n-1 } is behaving in the same way for b n { b }_{ n } as a n 1 { a }_{ n-1 } is doing for a n { a }_{ n } , min( a n 1 { a }_{ n-1 } ) such that a n { a }_{ n } is real is also "k".

So, min A l l t e r m s a r e r e a l a n 1 = min A l l t e r m s a r e r e a l b n 1 \min _{ All\quad terms\quad are\quad real }{ { a }_{ n-1 } } =\min _{ All\quad terms\quad are\quad real }{ { b }_{ n-1 } } \\ min A l l t e r m s a r e r e a l a n 1 = min A l l t e r m s a r e r e a l a n \therefore \min _{ All\quad terms\quad are\quad real }{ { a }_{ n-1 } } =\min _{ All\quad terms\quad are\quad real }{ { a }_{ n } } \\ min A l l t e r m s a r e r e a l a 1 = min A l l t e r m s a r e r e a l a 2 \therefore \min _{ All\quad terms\quad are\quad real }{ { a }_{ 1 } } =\min _{ All\quad terms\quad are\quad real }{ { a }_{ 2 } } \\

So minima of terms (and consequently z) such that all terms are real will occur when a 1 = a 2 { a }_{ 1 }={ a }_{ 2 }

( H e r e " z " r e p r e s e n t s m i n i m u m v a l u e o f z ) ln z = log ln z z log e z = log ln z z e = ln z min z = e e \left( Here\quad "z"\quad represents\quad minimum\quad value\quad of\quad z \right) \\ \\ \ln { z } =\log _{ \ln { z } }{ z } \\ \therefore \log _{ e }{ z } =\log _{ \ln { z } }{ z } \\ \therefore e=\ln { z } \\ \boxed { \min { z } ={ e }^{ e } }

Please correct me if I have made any mistake.

Archit Boobna - 6 years, 2 months ago

I am not sure but I have never seen a natural logarithm with a base. So I assumed it to be log instead of ln.

Archit Boobna - 6 years, 2 months ago

Having all the terms real does not imply that the corresponding limits exist.

Abhishek Sinha - 6 years, 2 months ago

Log in to reply

The idea is that if even 1 term is negative, the next term will not exist, leading to the limit not existing

Archit Boobna - 6 years, 2 months ago

Log in to reply

True. But here we are comcerned with showing that the limit indeed exists when z crosses a threshold.

Abhishek Sinha - 6 years, 2 months ago

Log in to reply

@Abhishek Sinha So how do we show that?

Archit Boobna - 6 years, 2 months ago

Log in to reply

@Archit Boobna I think it is possible to show the corresponding map f ( x ) = ln z ln x f(x)=\frac{\ln z}{\ln x} to be a contraction mapping into some suitable interval where f f maps into itself. Then the convergence of the sequence will follow.

Abhishek Sinha - 6 years, 2 months ago

Pi never occurs as a result in logarithms, choose "Double e" ✅

Eric Scholz - 2 years, 4 months ago
Raphael Nasif
Apr 17, 2015

Crap, I was just about to post a question just like this...

Now let the series be defined with the complex logarithm. Does there exist z C = min { z z , L R } z_\mathcal{C} = \min \lbrace z \mid z, \mathcal{L} \in \mathbb{R} \rbrace

where lim n a n = L \lim_{n \to \infty}{a_n} = \mathcal{L} ?

The problem is essentially about the stability of a fixed point. Allowing for complex logarithms allow us to define the sequence, but that doesn't imply that we will still obtain convergence to the fixed point.

Calvin Lin Staff - 6 years, 1 month ago

Log in to reply

Haha I was not implying so, just posing the question for others to ponder.

Raphael Nasif - 6 years, 1 month ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...