Hello everybody,
I am programmer interested in seeking members for PROOF of CONCEPT in "new
approaches" to Factoring Algorithms .

http://www.sourceforge.net/projects/atpg

For previous work see

http://www.lri.fr/~simon/satex/infos/benchs-generators/VZNuri.html

and

http://www.cs.indiana.edu/cgi-pub/sabry/cnf.html

The project works follows.

Hello everybody,
I am programmer interested in seeking members for PROOF of CONCEPT in "new
approaches" to Factoring Algorithms .

http://www.sourceforge.net/projects/atpg


For previous work see

http://www.lri.fr/~simon/satex/infos/benchs-generators/VZNuri.html

and

http://www.cs.indiana.edu/cgi-pub/sabry/cnf.html

The project works follows.

1) Build a multiplier circuit for inputs A(n/2 -bit) and B( n - bit ) and
output C (n-bit)

2) Put a gaint AND gate D checking C such that the output of the gate is 1,
when A and B are factors of C.

For example, If the number to be factored is 5, and C is 4-bits long.

D = C_0 AND C_1' AND C_2 AND C_3'

D becomes 1, iff if C_0 = 1, C_1 = 0 , C_2 = 1, and C_3 = 0

If the number to be factor is 8, and C is 4-bits long

D = C_0' AND C_1' AND C_2' AND C_3

D becomes a 1, iff and only if C_0 = 0 , C_1 = 0, C_2 = 0, C_3 = 1

3) Find binary vectors A, B such that D = 1

-----------

What the project currently implements

1) Program for simulating binary circuits
2) multiplier generators
3) Partly implemented PREDICT Solver

-------------
What the goal of the project is

The project goal is to use statistical techniques, for finding A and B.

See
http://sourceforge.net/project/showfiles.php?group_id=61183&package_id=106530
for some methods


The idea is to find a method that is able to calculate the probability of D
being a 1, as accurately as possible. If this could be accomplished most of
the time, for weightsets A , B, it would be equivalent to factoring C.

-suresh