Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

If $$a_{n+1}=a_n^2+1,$$ with initial $a_1=\frac{1}{2}$. How to solve this sequence problem, i.e., how to represent $a_n$ in closed form?

share|cite|improve this question
What is the question? – MJD May 16 '12 at 15:39
Let's not jump into voting to close, before we ask OP for clarifications. Especially if OP is willing to promptly reply and edit the question within few minutes of "clarification request" comments. It's annoying. – user2468 May 16 '12 at 15:51
It is unlikely that there is a closed form. Recursions of the form $a_{n+1}=a_n^2+c$ are very complicated. The Mandelbrot set reflects this. – lhf May 16 '12 at 16:04
My first though is to represent $a_n=\frac{b_n}{c_n}$ in lowest terms. It becomes $c_n=2^{2^{(n-1)}}$, $b_{n+1}=b_n^2+2^{2^{(n-1)}}$, $b_1=1$, which gets you into $\mathbb N$. But I'm not sure it helps any. – Ross Millikan May 16 '12 at 16:04
Why do you think there is a closed form? What would you do if you had one? Perhaps you're trying to solve a problem related to this sequence that does not need a closed form. – lhf May 16 '12 at 16:06
up vote 8 down vote accepted

EDIT: The answer below assumes $a_1$ is an integer, and as such is not directly applicable to your question, but I am leaving it here in case people can make use of the results mentioned in the paper below.

Aho and Sloan proved that for sequences like yours, there is a constant $k$ such that

$$a_n = \lfloor k^{2^n} \rfloor$$

for sufficiently large $n$. $k$ can be defined as a limit of a sequence using $a_n$ itself. If you include $k$ as one of your closed form constants, you are done!

See their paper for details:

Of course, for your special case, one might still be able to find a different "closed form" which might be more appealing to you.

share|cite|improve this answer
But $a_n$ is never an integer, is it? – Did May 16 '12 at 17:47
@Didier: Yes, it isn't. The paper deals with that too. I read the question too quickly and thought $a_1 = 1$. Will update. – Aryabhata May 16 '12 at 18:00
@anon: No, it is my fault. – Aryabhata May 16 '12 at 18:00
@Didier: I added a disclaimer! It might need some work to get those results to work for this problem (if at all they work). But I have left the answer, in case someone wants to try. – Aryabhata May 16 '12 at 20:17

The recurrence $a_{n+1} = a_n^2+c$ has a (known) closed form if and only if $c=0$ or $c=-2$. See this answer for more explanation.

share|cite|improve this answer

Perhaps let $a_n=\tan \theta_n$. Then $\tan \theta_{n+1}=\sec^2 \theta_n$. So, $\tan \theta_2=\sec^2 \theta_1$, $\tan \theta_3=1+\sec^4 \theta_1$, etc. I am not sure if this procedure will produce a closed form.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.