I am currently in the second year of high school doing a research paper in mathematics and I need some help with it.

I am trying to find numbers that are both Polygonal and Mersenne. Polygonal numbers are numbers of the form $\cfrac{n^2(s-2)-n(s-4)}{2}$, where $s$ is the number of sides that the polygon has, a.k.a. an s-gon, and n is the number of points that one of the outermost sides has, a.k.a. the n-th s-gon. For example, when $n=2$ and $s=3$, we get the second triangular number, $3$. Mersenne numbers are numbers of the form $2^p -1$. The more known type of number is a Mersenne prime, but in this case we do not require the p to be a prime. So if we want to find numbers that are both Mersenne numbers and Polygonal numbers, we need to find integer solutions to the equation $\cfrac{n^2(s-2)-n(s-4)}{2} = 2^p -1$. First, in order to get rid of the redundant solutions, we say that $s \ge 3$ (because 3 is the smallest number of sides that a polygon can have (Euclidian space)) and that $n \ge 3$ (Because the first s-gonal number is always $1$, and the second s-gonal number is always $s$, with which we would be able to asign every Mersenne number atleast $1$ Polygonal number, but that wouldnt be fun :) ) and $p \ge 1$ ($p$ can be whatever positive integer).

So now we have an equation with 3 variables that we seek integer solutions to. I know that I will need to use Diophantines equations and others, but is it too hard for a mere mortal like me unlike the the mastermind that is Wolfram Alpha to solve this equation? Should I change it to instead of using the general formula for Polygonial numbers use the formula for a specific P. number (for example if we take triangular numbers we would get $\cfrac{n^2+n}{2} = 2^p -1$? What kind of questions should I ask after this one (which numbers are both P. and M.), maybe is there an infinite amount of them, which if I understand correctly is a typical question for Diophantine analysis.

Thanks in advance and please ask questions if you need further info., and sorry if there are some grammatical mistakes in the text.

Edit: I am seeking non-trivial solutions to the problem, as in n is not 1 or 2. Sorry about that, should have mentioned it earlier.



Let's consider the approach of fixing either $n$ or $s$.

Fixing $s$ clearly leads to difficulties, for example for $s=3$, the problem becomes equivalent to the Ramanujan-Nagell equation whose solution is non-trivial - but interesting to understand/explore.

Fixing $n$ is probably a good approach, because the numbers now form an A.P. For example, at $n=3$, the numbers take the values of all multiples of 3 which are at least 6 and hence all $2^k-1$ values are taken for which $k$ is even and at least 4. You could get a collection of such "modulo" results for each value of $n$ and try to combine them.

  • $\begingroup$ Thanks for the comment. What about not fixing any of the variables? $\endgroup$ – redelectrons Sep 29 '15 at 18:53

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.