2
$\begingroup$

The following series seems to become an arithmetic series $$a_{n+2} = a_n^2-n\cdot a_{n+1}$$ If $$a_1=3, a_2=4$$ $$(a_3=5, a_4=6, ..., a_n = n+2)$$

Can $a_n=n+2$ be derived from the original recursion equation?

And are there any other initial values that will also lead to an arithmetic series?

$\endgroup$
5
$\begingroup$

Of course.

If $a_n=n+2$ and $a_{n+1}=(n+1)+2$ then $$a_{n+2}=(n+2)^2-n(n +3)=n+4. $$

For the generalizing question, assume $a_n=bn+c$ for all $n$. Then $$b(n+2)+c = (bn+c)^2-n(b(n+1)+c), $$ i.e., $$bn+(2b+c)=(2bc-b-c)n+c^2$$ for all $n$, which implies $$ b=2bc-b-c,\qquad 2b+c=c^2,$$ hence $2bc=2b+c=c^2 $. One trivial solution is $b=c=0$ (i.e., $a_n=0$ for all $n$). If $c\ne 0$, we are led to $c=2b$ and $c=2$ (i.e., $a_n=n+2$for all $n$).

$\endgroup$
  • 1
    $\begingroup$ Thanks, Hagen. Would there be some way to go from $a_{n+2}=a_n^2-n\cdot a_{n+1}$ and get to $a_n=n+2$ somehow? (not assuming $a_n = bn+c$) $\endgroup$ – Kay K. Nov 24 '15 at 17:57
  • 2
    $\begingroup$ There's a mistake in the simplification of the $n^2$ terms, but the conclusion remains. $\endgroup$ – Yves Daoust Nov 24 '15 at 18:02
3
$\begingroup$

If the terms form an arithmetic sequence,

$$a_n=\alpha n+\beta.$$

Then by the recurrence,

$$\alpha(n+2)+\beta=(\alpha n+\beta)^2-n(\alpha(n+1)+\beta)=(\alpha^2-\alpha)n^2+(2\alpha\beta-\alpha-\beta)n+\beta^2.$$

The term in $n^2$ can only vanish when $\alpha=0$ or $\alpha=1$, which give $$\beta=-\beta n+\beta^2\implies \beta=0,$$

or $$n+2+\beta=(\beta-1)n+\beta^2\implies \beta=2.$$

There are no other possibilities.

$\endgroup$

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.