mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   science_man_88 (https://www.mersenneforum.org/forumdisplay.php?f=140)
-   -   Modified Form of LL Test? (https://www.mersenneforum.org/showthread.php?t=23920)

science_man_88 2018-12-21 15:12

Modified Form of LL Test?
 
[URL="https://en.m.wikipedia.org/wiki/Iterated_function"]Iterated function[/URL] on Wikipedia has two examples that are relevant:

\[f(x)=ax+b\to f^n(x)=a^nx+ \frac{a^n-1}{a-1}b\]

and

\[f(x)=ax^2+bx+\frac{b^2-2b-8}{4a}\to f^n(x)=\frac{2\alpha^{2^n}+2\alpha^{-2^n}-b}{2a}\]

where:

\[\alpha=\frac{2ax+b\pm\sqrt{(2ax+b)^2-16}}{4}\]

The first is the way Mersenne numbers iterate with a=2,b=1 . The second is how the the Lucas-Lehmer test variants iterate. The normal Lucas-Lehmer test sequence (prior to mod) is a=1,b=0. The reduced version using \(2x^2-1\) is a=2,b=0. I chose to look at the latter as it shares a possible x value with the Mersenne numbers.
Now getting the b values to equate was a priority for me ( because it simplifies the expressions if they both equal 0, as then b can be taken out of the variables). Allowing b to go to b-1 without changing the value of the first n-th iterate gives the following:

\[a^nx+\frac{a^n-1}{a-1}b+\frac{a^n-1}{a-1}\]

Now we can eliminate all the b values and any direct multiplies by b from both n-th iterates giving:

\[f^n(x)=a^nx+\frac{a^n-1}{a-1}\]

and

\[f^n(x)=\frac{2\alpha^{2^n}+2\alpha^{-2^n}}{2a}\]

where:

\[\alpha=\frac{2ax\pm\sqrt{(2ax)^2-16}}{4}\]


Next thing to note is that a is equal in both cases (namely a=2). Plugging that fact, plus \(a=\frac{x-1}{3}\) for our common x value into the first case, and the whole set of equations in the second reduces to :

\[f^n(x)=\left ( \frac{x-1}{3}\right )^nx+\left ( \frac{x-1}{3}\right )^n-1\]

and

\[f^n(x)=\frac{\alpha^{2^n}+\alpha^{-2^n}}{2}\]

where:

\[\alpha=\frac{4x\pm\sqrt{(4x)^2-16}}{4}=x\pm\sqrt{x^2-1}\]

My main question is:

What's the simplified version of their modular remainder ( second mod first)?

that can allow us to solve for a relation n+3 must have to get 0 out and therefore all odd Mersenne prime exponents.

(Idea for this thread, was from: ewmayer and CRGreathouse)

Batalov 2018-12-21 15:46

Look no farther than Wikipedia - [URL]https://en.wikipedia.org/wiki/Lucas%E2%80%93Lehmer_primality_test#Proof_of_correctness[/URL]
and you will find the same, only with [$]\omega = \alpha[/$]

science_man_88 2018-12-21 17:31

[QUOTE=Batalov;503541]Look no farther than Wikipedia - [URL]https://en.wikipedia.org/wiki/Lucas%E2%80%93Lehmer_primality_test#Proof_of_correctness[/URL]
and you will find the same, only with [$]\omega = \alpha[/$][/QUOTE]

I'm looking for n that cause it to be 0. yes I know they will be such that n+3 is prime but I'm looking for a property that isn't so generic.


All times are UTC. The time now is 19:12.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.