I'm trying to define a formula which will assist in the generation of some unusual numeric sequences. The sequences are known ahead of time, so I'm really just looking for a simple arithmetic formula that can be used to calculate them, rather than hard coding the values (so that it's more extensible for future updates to my system). I've tried this a number of ways but am not sure what method to use in order to attempt to define this. Any assistance would be extremely appreciated!

Known Values

when n = 1 the sequence = [0,2,...]

when n = 2 the sequence = [0,-1,3,2,...]

when n = 3 the sequence = [0,-1,-2,4,3,2,...]

when n = 4 the sequence = [0,-1,-2,-3,5,4,3,2,...]

when n = 6 the sequence = [0,-1,-2,-3,-4,-5,7,6,5,4,3,2,...]

when n = 8 the sequence = [0,-1,-2,-3,-4,-5,-6,-7,9,8,7,6,5,4,3,2,...]

All sequences loop on the "..." (i.e. when n=1 the sequence is really [0,2,0,2,0,2,etc.])

Re-stating the task

The task is to define a formula that defines the value of the ith member of the sequence, with prior-knowledge of the value of n.

aka x in f(n,i) = x

Fwiw, this is a real world problem to assist in the generation of dynamic imposition templates for a printing application :)

  • $\begingroup$ So when n=1, there are only 2 terms in the sequence which keep on repeating? $\endgroup$ – Manish Kundu Jan 9 '18 at 2:45
  • $\begingroup$ @ManishKundu infinite terms, repeated as 0,2,0,2,0,2... etc $\endgroup$ – John Doe Jan 9 '18 at 2:46
  • $\begingroup$ It is $i -(2n+1) \lfloor \frac{i+n+1}{ 2n+1} \rfloor$ $\endgroup$ – user798409 Jan 9 '18 at 2:47

We can see that it is something like $$f(n,i)=\begin{cases}1-i&0<i\le n\\2n+2-i&n<i\le2n\end{cases}$$Then the repetition can be dealt with by modular arithmetic (or ceiling/floor functions). Suppose $i>2n$. Then computing $i\mod{2n}$ can be done by $$i\equiv i-2n\left\lfloor\frac{i-1}{2n}\right\rfloor\mod{2n}$$This will be used in place of $i$ in the final formula.

If $i\in[1,n]\mod{2n}$, then $$\frac {i}n\in(2k,2k+1]\implies\left\lceil\frac{i}{n}\right\rceil=2k+1$$and similarly, if $i\in[n+1,2n]\mod{2n}$, then $$\left\lceil\frac{i}{n}\right\rceil=2k$$

So we get $$f(n,i)=\frac{1-(-1)^{\left\lceil\frac{i}{n}\right\rceil}}{2}\left(1-i+2n\left\lfloor\frac {i-1}{2n}\right\rfloor\right)+\frac{1+(-1)^{\left\lceil\frac{i}{n}\right\rceil}}{2}\left(2n+2-i+2n\left\lfloor\frac {i-1}{2n}\right\rfloor\right)$$where the initial fractions are to choose whether we are in the range $i\in[1,n]$ or $i\in[n+1,2n]$ in$\mod{2n}$.

  • $\begingroup$ Thanks for the fast response but when I apply this to the first sequence (n=1), i am getting a sequence of [4,0,...]? $\endgroup$ – SJB Jan 9 '18 at 3:36
  • $\begingroup$ @SJB If $i=n=1$, then $\left\lceil\frac in\right\rceil=1$, and $\left\lfloor\frac i{2n}\right\rfloor=0$, so you should get the first entry to be $0$, as required: $$f(1,1)=\frac{1-(-1)^1}{2}(1-1+2\cdot 0)+\frac{1+(-1)^1}{2}(2+2-1)=0$$ But yes, you're right it was slightly wrong, since it got the $2n$-th term wrong before. I believe it is fixed now. Before, it was giving a sequence [0,4,0,4,...]. Now, I believe the next term is $$f(1,2)=0\cdot(1-2+2)+\frac{1+1}{2}(2+2-2)=2$$(By the way, I am using that the first entry in the sequence has $i=1$, not $i=0$, which some people might use) $\endgroup$ – John Doe Jan 9 '18 at 3:50
  • $\begingroup$ Yeah the base-1 was throwing it off - this is great now! thank you! $\endgroup$ – SJB Jan 9 '18 at 3:53
  • $\begingroup$ No worries, you're welcome :) $\endgroup$ – John Doe Jan 9 '18 at 3:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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