Limit behaviour of Fibonacci sequence

  • Context: Undergrad 
  • Thread starter Thread starter PeroK
  • Start date Start date
PeroK
Science Advisor
Homework Helper
Insights Author
Gold Member
2025 Award
Messages
29,687
Reaction score
21,502
TL;DR
Ratio of terms in the Fibonacci sequence.
I thought this might be quite interesting to anyone who hasn't seen it before.

Consider any Fibonacci sequence, where ##F_1 > F_0 \ge 0## and ##F_{n+1} = F_n + F_{n-1} \ (n > 1)##. We want to show that the ratio of consecutive terms, ##a_n = \dfrac{F_{n+1}}{F_n}## tends to the Golden Ratio: ##\varphi = \dfrac{1 + \sqrt 5}{2}##.

First, note that ##1 \le a_n \le 2 \ (n \ge 1)##. The sequence is increasing and no term can be more than twice the previous term.

Note that:
$$a_{n + 1} = \frac{F_{n+2}}{F_{n+1}} = \frac{F_{n+1} + F_n}{F_{n+1}} = \frac{(a_n + 1)F_n}{a_nF_n} = \frac{a_n + 1}{a_n} = 1 + \frac 1 {a_n}$$Hence:
$$a_{n+1 } > \varphi \ \Leftrightarrow \ 1 + \frac 1 {a_n} > \varphi \ \Leftrightarrow \ a_n < \frac 1{\varphi - 1}= \varphi$$This last equality follows from the characteristic quadratic equation that defines ##\varphi##:
$$\varphi^2 - \varphi -1 = 0 \ \implies \ \varphi(\varphi-1) = 1 \ \implies \ \varphi = \frac 1 {\varphi - 1}$$So:

If ##a_n < \varphi##, then ##a_{n+1} > \varphi##; and, if ##a_n > \varphi##, then ##a_{n+1} < \varphi##.

This means that ##a_n## is composed of two subsequences, alternating between above and below the Golden Ratio.

Note also that if ##a_n = \varphi##, then ##a_{n+1} = \varphi## and the sequence is constant thereafter. This can only happen, however, when ##F_1 = \varphi F_0##.

If you know a bit of real analysis, you see that it is enough to show that the subsequence less than ##\varphi## is strictly increasing; and, the subsequence above ##\varphi## is strictly decreasing. This guarantees that both subsequences converge to ##\varphi## - as that is the only possible limit.

To see this, we need to compare ##a_{n+2}## with ##a_n##:
$$a_{n+2} = \frac{F_{n+3}}{F_{n+2}} = \frac{F_{n+2} + F_{n+1}}{F_{n+1} + F_n} = \frac{(2a_n + 1)F_n}{(a_n+1)F_n} = \frac{2a_n + 1}{a_n +1}$$Hence:
$$a_n - a_{n+2} = \frac{a_n^2 + a_n - 2a_n - 1}{a_n + 1} = \frac{a_n^2 - a_n - 1}{a_n + 1}$$And we see that:

If ##1 < a_n < \varphi##, then ##a_n^2 - a_n - 1 < 0## and ##1 < a_n < a_{n+2} < \varphi##; and

if ##\varphi < a_n##, then ##a_n^2 - a_n - 1 > 0## and ##\varphi < a_{n+2} < a_n##.

And, as required, we have two monotonic subsequences that must both converge to ##\varphi##.

Note that to prove this last bit rigorously, you need some work, which may not add very much here.
 
Last edited:
Mathematics news on Phys.org
PeroK said:
TL;DR: Ratio of terms in the Fibonacci sequence.

I thought this might be quite interesting to anyone who hasn't seen it before.

Consider any Fibonacci sequence, where and . We want to show that the ratio of consecutive terms, tends to the Golden Ration .

First, note that . The sequence is increasing and no term can be more than twice the previous term.

If , then:
And
(As satisfies , which is its defining equation.

The first question is when is ? Note that:
So
We can see that the sequence alternates between less than and greater than the Golden Ratio. Unless, of course it equals the Golden ratio, in which case it remains constant thereafter. (Note, however, that can only happen if .)

The next thing we want to show is that each of these odd and even subsequences is increasing for the lower subsequence, and decreasing for the higher sequence. We need to ask when is ? Note that:
So:
Is stuff missing from this post or is it my browser?
 
kuruman said:
Is stuff missing from this post or is it my browser?
The Latex disappears. It comes and it goes. The software on this site has been rubbish for years!
 
I might just give up. It's hopeless. Waste of time.
 
testing...

Example of in-line LaTeX: ##v(t) = L \frac{di(t)}{dt}##

Example of stand-alone LaTeX:
$$v(t) = L \frac{di(t)}{dt}$$
 
It's working for me...
 
berkeman said:
It's working for me...
It works for a bit, but once there is a significant amount, it gives up. The preview stops working.
 
PeroK said:
It works for a bit, but once there is a significant amount, it gives up. The preview stops working.
To clarify, you are saying that once of a latex threshold, preview stops working and it removes the code or? Do you get any errors? Anything in the browser console?
 
Claude appears to understand the problem. The key is this happening on posts with lots of equations and going back and forth with preview, edit. There is a lot of render thrash happening and the browser ejects. Working on a fix. Hope we don't run into a regression.
 
  • #10
Greg Bernhardt said:
To clarify, you are saying that once of a latex threshold, preview stops working and it removes the code or? Do you get any errors? Anything in the browser console?
No error. Simply the text with all the Latex missing. Missing from the preview and also when I post.
 
  • #11
Greg Bernhardt said:
Claude appears to understand the problem. The key is this happening on posts with lots of equations and going back and forth with preview, edit. There is a lot of render thrash happening and the browser ejects. Working on a fix. Hope we don't run into a regression.
That's what it seems like. I don't have an offline Latex editor. Is that what other people have? I've always relied on the site.
 
  • #13
PeroK said:
I might just give up. It's hopeless. Waste of time.
Been there myself. If it helps any, I learned by experience, after losing my hard work several times, to copy and paste posts with many complex equations in a backup text file as I go along. I assemble the backup in small sections each of which I review and edit separately as if they were replies themselves. There is no need for an offline editor if you do it online in segments. When I am satisfied everything, I paste the assembled text file all at once in the thread.
 
  • Like
Likes   Reactions: PeroK
  • #14
kuruman said:
Been there myself. If it helps any, I learned by experience, after losing my hard work several times, to copy and paste posts with many complex equations in a backup text file as I go along. I assemble the backup in small sections each of which I review and edit separately as if they were replies themselves. There is no need for an offline editor if you do it online in segments. When I am satisfied everything, I paste the assembled text file all at once in the thread.
Yes, I didn't lose anything, except the time messing around.
 
  • #15
Moved!
 
  • #16
Regardless of the formatting failure, the Fibonacci sequence has an explicit formula known as Binet's formula, which contains sqrt(5). Pretty neat how square roots can appear out of nowhere. Also, it looks like the LATEX works. I can see @PeroK 's LATEX just right.
 
  • Like
Likes   Reactions: PeroK

Similar threads

  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 6 ·
Replies
6
Views
5K
  • · Replies 2 ·
Replies
2
Views
3K
  • · Replies 15 ·
Replies
15
Views
2K
  • · Replies 3 ·
Replies
3
Views
4K