This site is supported by donations to The OEIS Foundation.

# User:Bogart B. Strauss

This is my approach to the elementary factorization of integers:

Given an integer $Q=(2^{b})*(3^{c})*(6v+(-1)^{r})*(-1)^{a}$ with a, b, c, v, and r equal to the positive integers and zero.

Since we can find factors of 2 and 3 with a glance and the sign of the integer is of no importance for factorization, they are considered trivial and thus we are left with $6v+(-1)^{r}$ which combines two cases (r-even and r-odd). If this set of numbers equals R, then R*R is an element of R.

$6v+(-1)^{r}$ is the sequence $1,5,7,11,13,17,19,23,25,29,31,35,37,41,43$ .....sequence A007310 of the OEIS.

Considering that this can be formulated with a single variable, let's define a new sequence z.

$z=(1/2)*(6x-(-1)^{x}+3)$ with x equal to the positive integers or zero.

Using a sieving method (if you are curious, email me), we can then derive something to equate $6v+(-1)^{r}$ with.

$6v+(-1)^{r}=z^{2}+z*(6y+3-(-1)^{x})$ with y equal to the positive integers or zero.

Thanks to Fermat realizing that a difference of squares with integer bases turns into a product of integers, we can then add that methodology to what we already have.

$6v+(-1)^{r}=z^{2}+z*(6y+3-(-1)^{x})=n^{2}-m^{2}$ where n and m are positive integers and m can equal zero with $n>m$ .

Considering that the factors will be found in divisor pairs, we can state that $6v+(-1)^{r}=z*f$ .....with the set of all f being equal to the set of all z.

Notable relations:

$n-m=z$ $n+m=f$ $n=3k+(1/4)*(1+(-1)^{r})*(3-(-1)^{x})$ where k is a positive integer or zero

$m=3y+(1/4)*(1-(-1)^{r})*(3+(-1)^{x})$ Things start getting strange when we look at the resulting patterns from n and m as they relate to $6v+(-1)^{r}$ . There is a clear split into 6 cases (3 for r-even, 3 for r-odd) and a monolith of a difference of squares formula:

$6v+(-1)^{r}$ can be split in three: $18p+(-1)^{r}$ , $18p+7*(-1)^{r}$ , and $18p+13*(-1)^{r}$ .

6v+(-1)^r=18p+(6q+1)*(-1)^r=( ((1/4)*(18d+18+9*(-1)^r+(6*(-1)^q-1-q*(1+(-1)^q))*(-1)^(d+r)))^2 - ((1/2)*(3y+3k+(3y-3k)*(-1)^r))^2 )*(-1)^r where p and d are positive integers or zero while q={0,1,2}. I wouldn't read too much into this formula without figuring a desired q and r first.

I could not derive any more obvious patterns from this point and had to mess around with, well, anything. That anything is 3 times the number to be factored as another difference of squares.

$3*(6v+(-1)^{r})=g^{2}-h^{2}$ where g is a positive integer not divisible by 3 and h is an integer not divisible by 3 with $g>h$ .

Notable resulting relations:

$g=2n-m$ $h=n-2m$ $6v+(-1)^{r}=gm+nh=(1/2)*(gn+mh)$ From that last formula, we realize that either $gn$ or $mh$ has only one possible factor of two or "critical 2". The reason is that $6v+(-1)^{r}$ is always an odd number and thus it is always a sum of an odd number and an even number.

If we list the first 15 numbers in the sequence $6v+(-1)^{r}$ , we can identify a pattern of critical 2's. Every unique divisor pair has a unique combination of h,m,n,g.

1*1____h=1, m=0, n=1, g=2

1*5____h=-1, m=2, n=3, g=4

1*7____h=-2, m=3, n=4, g=5

1*11___h=-4, m=5, n=6, g=7

1*13___h=-5, m=6, n=7, g=8

1*17___h=-7, m=8, n=9, g=10

1*19___h=-8, m=9, n=10, g=11

1*23___h=-10, m=11, n=12, g=13

1*25___h=-11, m=12, n=13, g=14

5*5____h=5, m=0, n=5, g=10

1*29___h=-13, m=14, n=15, g=16

1*31___h=-14, m=15, n=16, g=17

1*35___h=-16, m=17, n=18, g=19

5*7____h=4, m=1, n=6, g=11

1*37___h=-17, m=18, n=19, g=20

1*41___h=-19, m=20, n=21, g=22

1*43___h=-20, m=21, n=22, g=23

The pattern of critical 2's is $g,m,h,n,m,g,n,h$ repeating. Also notice that the critical 2's are unique to the number to be factored and thus relates each divisor pair of the number. This splits things up into 8 different cases (four for r-even and four for r-odd).

I can no longer reasonably subject anyone to a formula as we now have a total of 48 unique cases to deal with. Here's an example of just one case with B equaling positive integers and zero:

$6v+(-1)^{r}=18p+1$ (from the three-split)$=24B+1$ (from the four-split)$=72B+1$ (from the forty eight split)

m is an element of $12B$ n is an element of $18B+or-1$ h is an element of $+or-z$ g is an element of $2*z$ Recall that z is defined earlier. You can see here that the critical 2 is g for all numbers equal to $72B+1$ .

So $145=17^{2}-12^{2}=(17+12)*(17-12)=29*5$ . While $3*145=22^{2}-(-7)^{2}=(22-7)*(22+7)=15*29$ .

That's as far as I've gotten with the elementary factoring of integers and I encourage anyone to try and dig deeper, I certainly am. Patterns exist, we merely have yet to find them. But here it seems that there will always be a "next" pattern.

I will be transferring more summaries of my work here soon. The next piece will be an N choose K type of formula for figuring out the number of divisor pairs with a given prime factorization. This one will be a while as I am trying for a better notation for the extension of Sigma summation I have to use.

This is no place for philosophical statements or meandering through things I may perhaps misunderstand. For the things I may be mistaken about, I have yet to understand the proofs otherwise. I don't believe, I think. It does not escape me that this is a very precious responsibility.