My Math Forum  

Go Back   My Math Forum > College Math Forum > Number Theory

Number Theory Number Theory Math Forum


Thanks Tree2Thanks
Reply
 
LinkBack Thread Tools Display Modes
November 30th, 2016, 07:29 PM   #11
Math Team
 
topsquark's Avatar
 
Joined: May 2013
From: The Astral plane

Posts: 1,659
Thanks: 652

Math Focus: Wibbly wobbly timey-wimey stuff.
Quote:
Originally Posted by Compendium View Post
You need only one counterexample to disprove it, though.
True, but I somehow doubt that's what he's trying to do.

-Dan
topsquark is offline  
 
December 10th, 2016, 01:30 PM   #12
Newbie
 
Joined: Aug 2016
From: Auckland

Posts: 10
Thanks: 1

Hello all,

I've come back!!!

Sorry for my being invisible few days ago. I focused in programming, and with the help of programming community, now I have proved my theory with the first 10,000,000 (ten millions) whole number from 1 to 10,000,000. My program was written in Java.

I can test more with greater number, but currently ten millions is enough.

I put my data in a microsoft access file Data.accdb, share it in Google drive with link

https://drive.google.com/drive/u/0/f...Uc2ZnkzV2xaTkE

I shared full access for everyone. So please com there, open this file and see table data1. They're there.

Look at column four 'Sum or Difference' in this table. For cases which belong to the second rule Integer = Prime - Fibonacci, I put comment 'Difference'. Otherwise it's NULL.

Please enjoy.

Last edited by skipjack; December 16th, 2016 at 05:47 AM.
thinhnghiem is offline  
December 11th, 2016, 07:54 PM   #13
Newbie
 
Joined: Aug 2016
From: Auckland

Posts: 10
Thanks: 1

Hello all,

Please ignore my output file. It has many incorrect calculations. My program has a bug, and I have to fix it first.

Last edited by skipjack; December 16th, 2016 at 05:45 AM.
thinhnghiem is offline  
December 16th, 2016, 02:34 AM   #14
Newbie
 
Joined: Aug 2016
From: Auckland

Posts: 10
Thanks: 1

Hello all,

I checked my data. All of the incorrect records belong to Prime numbers which are greater than 2.1 billions. it results from the limit of Java in huge number processing.

I use the same algorithm to rewrite the program in C. It solve a number of missing records, but not all of them

For example 1651 has been expressed successfully as
18446744073709550683 2584 1651.That means the prime number found is over 18 billions billions (Terrible !!!)

Finally, there are 9,335,421 numbers from 1 to 10,000,000 are analyzed successfuly. The number of failure is 664,579 records ~ 6.6%.

There may be two cases:

1. Either the calculation is out of C language capability so it had to break out.

2. Or my conjecture fails for these records. That means there is no pair of Prime/Fibonacci to fulfill the rules stated by me.

New data has been updated into the same sharing location in Google drive (Data.txt and Data.accdb)

Thank you.

Last edited by skipjack; December 16th, 2016 at 05:45 AM.
thinhnghiem is offline  
December 16th, 2016, 05:27 AM   #15
Senior Member
 
Compendium's Avatar
 
Joined: Sep 2007
From: USA

Posts: 349
Thanks: 67

Math Focus: Calculus
I recommend using Sage for something like this. It is a computer algebra system designed to work with very large numbers, has a large amount of functions including number theory, and you can create your own algorithms on the fly without much trouble. It would require learning some Python to create the algorithm to find a counterexample, but the language is relatively easy to learn.

Instead of generating large lists, I recommend only searching for a counterexample. It would run substantially faster. I would program it to search only a specific range at a given time so you can keep track of the progress. Since Fibonacci numbers are easy to generate, I would generate one, find the difference between your current integer and the Fibonacci number, and ask Sage if it is prime. If not, move on to the next Fibonacci number.

Given that the integer can be the difference between a prime and Fibonacci number, it will be impossible to be 100% certain you found a counterexample since the prime numbers and Fibonacci numbers are unbounded.

Last edited by Compendium; December 16th, 2016 at 06:05 AM.
Compendium is offline  
December 18th, 2016, 02:31 PM   #16
Newbie
 
Joined: Aug 2016
From: Auckland

Posts: 10
Thanks: 1

Great, thank you Compedium. I will try to use Sage
thinhnghiem is offline  
Reply

  My Math Forum > College Math Forum > Number Theory

Tags
fibonacci, number, prime, relationships



Thread Tools
Display Modes


Similar Threads
Thread Thread Starter Forum Replies Last Post
shorter explicit Fibonacci number formula Math Message Board tutor Number Theory 1 September 9th, 2015 12:52 PM
Conjecture on period Fibonacci number mod 2^k MadSoulz Number Theory 1 May 2nd, 2015 08:56 PM
Fibonacci-Type series and the number e hfsierra Number Theory 1 April 24th, 2015 09:02 AM
Index of a fibonacci number abhijith Algebra 2 April 10th, 2010 02:17 PM
fibonacci and prime number correlation soandos Number Theory 44 November 19th, 2007 01:07 PM





Copyright © 2017 My Math Forum. All rights reserved.