Вход на сайт

Просмотр новости

Найдите то, что Вас интересует

Limit behaviour of Fibonacci sequence

Дата публикации: 12-06-2026 15:58:10



Основное содержимое страницы с новостью.

Skip to main content
  • Level: Undergrad 
  • Thread starter Thread starter PeroK
  • Start date Start date Jun 7, 2026
Science Advisor
Homework Helper
Gold Member
2025 Award
Messages
29,902
Reaction score
21,780
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.

Discussion
Science Advisor
Homework Helper
Messages
3,369
Reaction score
1,912
There is a much easier way to arrive at these results.

From ##a_{n+1} = 1 + a_n^{-1}## and the definition of ##\varphi## we can obtain $$
a_{n+1} - \varphi = (-1)\frac{a_n - \varphi}{a_n\varphi}.$$ Assuming ##a_n \geq 1## we have ##a_n \varphi \geq \varphi > 1## so that $$
|a_{n+1} - \varphi| < |a_n - \varphi|.$$ This is sufficient to conclude that ##a_n \to \varphi##. The factor of ##-1## on the right means that if ##a_n < \varphi## then ##a_{n+1} > \varphi## etc.

Science Advisor
Homework Helper
Gold Member
2025 Award
Messages
29,902
Reaction score
21,780
There is a much easier way to arrive at these results.

From ##a_{n+1} = 1 + a_n^{-1}## and the definition of ##\varphi## we can obtain $$
a_{n+1} - \varphi = (-1)\frac{a_n - \varphi}{a_n\varphi}.$$ Assuming ##a_n \geq 1## we have ##a_n \varphi \geq \varphi > 1## so that $$

That is useful.
I didn't think this was sufficient, as we could have the odd terms and even terms both converging to different limits less than and greater than ##\varphi##.

The additional equation that I thought was needed was:
$$a_n - a_{n+2} = \frac{a_n^2 - a_n - 1}{a_n + 1}$$This ties the odd or even sequence to the function values. Because ##x^2 - x - 1## has only one zero on ##(1, 2)##, the only way that ##a_n^2 - a_n - 1## can tend to zero is if ##a_n## tends to ##\varphi##.

Science Advisor
Homework Helper
Messages
3,369
Reaction score
1,912
A slight correction is indeed required.

From ##a_n \geq 1## we have ##a_n \varphi \geq \varphi ## so that $$
|a_{n+1} - \varphi| \leq \frac{1}{\varphi} |a_n - \varphi|$$ which together with ##\varphi > 1## is sufficient for ##a_n \to \varphi##.

Science Advisor
Homework Helper
Messages
3,369
Reaction score
1,912
Alternatively, we can also consider that the general solution of ##F_{n+2} = F_{n+1} + F_n## is $$\begin{split}
F_n &= A \left( \frac{1 + \sqrt{5}}{2}\right)^n + B\left( \frac{1 - \sqrt{5}}{2}\right)^n \\
&= A \varphi^n + B (-\varphi^{-1})^n \\
&= \varphi^n \left(A + B (-\varphi^{-2})^{n}\right)
\end{split}
$$ for some constants ##A## and ##B##. For an integer sequence, ##B## must be the conjugate of ##A## in ##\mathbb{Q}[\sqrt{5}]## so that if ##A = a + b\sqrt{5}## then ##B = a - b\sqrt{5}## for rational ##a## and ##b##. For the case ##F_0 = 0, F_1 = 1## we in fact have ##a = 0## and ##b = \frac15##.

From the last expression for ##F_n## we have $$
\frac{F_{n+1}}{F_n} = \varphi \frac{ A +B \left( - \varphi^{-2} \right)^{n+1} }
{ A + B \left( - \varphi^{-2} \right)^{n} }$$ and some basic analysis shows that as ##n \to \infty## for ##A \neq 0## the fraction tends to 1. It is also easy to see that ##(F_n)## alternates between terms which are greater than and less than ##A\varphi^n##, with ##(F_{n+1}/F_n)## therefore alternatiing between terms which are greater than and less than ##\varphi##.

Схожие новости

#Наименование новостиТональностьИнформативностьДата публикации
1Recursive pattern in Collatz sequence valuation jumps01023-07-2025
2About a natural extension of Collatz conjecture01024-03-2026
30001-01-1970
40001-01-1970
50001-01-1970
60001-10-2026
70030-09-2026
80030-09-2026
90030-09-2026
10Series with rotating phase01021-09-2026

Классификация: . Схожих патентов: 0. Схожих новостей: 10. Тональность: 0. Информативность: 10. Источник: www.physicsforums.com.