Preprint
Article

This version is not peer-reviewed.

Does the Collatz Sequence Eventually Reach 1 for All Positive Integer Initial Values?

Submitted:

02 May 2023

Posted:

02 May 2023

You are already at the latest version

Abstract
This study focuses on one of the most famous open problems in mathematics, namely the Collatz conjecture. The Collatz conjecture or 3x + 1 Problem is perhaps today's most enigmatic unsolved mathematical problem. It is named after Lothar Collatz, who rst proposed it in 1937. It may be stated as as follow: Take any positive integer n. If n is even then divide it by 2, else do \triple plus one" and get 3n + 1. The conjecture is that this process will eventually reach the number 1, regardless of which positive integer is chosen initially. In this paper, we present a simple proof for the Collatz conjecture.
Keywords: 
;  ;  

1. Introduction

The Collatz conjecture is an unsolved conjecture in mathematics. It is named after Lothar Collatz, who first proposed it in 1937. The conjecture is also known as the 3 n + 1 conjecture, the Ulam conjecture (after Stanislaw Ulam), the Syracuse problem, as the hailstone sequence or hailstone numbers, or as Wondrous numbers per Gödel, Escher, Bach.
The Collatz problem is one of the most famous unsolved issues in mathematics. Paul Erdös was correct when he stated, “Mathematics is not ready for such problems”. Lagarias [7] stated that the Collatz conjecture “is an extraordinarily difficult problem, completely out of reach of present day mathematics”. Possibly, the interest to this problem is related to the fact that the question is very easy to state, but very hard to solve. In fact, it is even complicated to give partial answers. The problem addresses the following situation: Consider an iterative method over the set of positive integers defined in the following way. If n is even, then we consider the positive integer n 2 for the next step. On the other hand, if n is odd, then we consider the positive integer 3 n + 1 for the next step. The Collatz conjecture states that, independently of the chosen initial value for n, the number 1 is reached eventually.
In [7] discussion concerning the origin of the problem was given. In the last 50 years, the mathematical community has tried different approaches to the Collatz conjecture, but none of them is believed to provide a definitive path that would allow to solve the problem [8]. On one hand, there are theoretical arguments that allow proving statements which are similar to the conjecture, but a bit weaker [5,10]. On the other hand, there are numerical experiments that show that the conjecture holds for numbers which are smaller than a certain threshold N, or that the Collatz functions does not have non-trivial cycles with length less or equal to m. The values of m and N have been continuously improved, and, nowadays, we can ensure that the conjecture holds for N < 5 . 78 × 10 18 [7] or that the length of a non-trivial cycle is, at least, 1 . 7 × 10 7 [3]. These computational arguments have been feeding continuously the opinion that the Collatz conjecture is true, and that there is no non-trivial cycle.
Besides, some authors have devoted efforts to rewrite the conjecture in other terms such as algebraic and boolean fractals [2]. Furthermore, some study has been developed concerning the representation and study of the Collatz conjecture in terms of graphs [1,4,6,9].
In this paper the authors propose a proof for the Collatz conjecture. For this purpose, we use the base-2 numeral system or binary numeral system representation. The reasons for using binary system are: a) The multiplication property of 11 (recall that for any positive integer r = a b c d we have r × 11 = ( a b c d ) × 11 = a ( a + b ) ( b + c ) ( c + d ) d and r × 11 = ( a b c d ) × 11 = e f g h d , where e = a + 1 if a + b 10 ; f= ones digit of ( a + b ) + 1 if b + c 10 ; g = ones digit of ( b + c ) + 1 if c + d 10 ; h= ones digit of ( c + d ) if c + d 10 ) and note that since 3 = ( 11 ) 2 we can apply the above property in binary mode ; and b) In binary representation of any positive integer, the number of zeros appearing to the right determines that we can divide the number by which power of 2. For instance, if P = ( 11000 ) 2 , then the number of zeros is 3, thus P is dividable by 2 3 .

2. Preliminaries

In this section we first give some lemmas and then using these lemmas the main theorem is proved in the next section.
Lemma 2.1. If n = ( 1010 01 ) 2 , then the next step of sequence ends to 1.
Proof. 
Assume that n is multiplied by 3 = ( 11 ) 2 and then the result is added to one. Hence, 3 n + 1 = ( 1 000 0 k t i m e s ) 2 = 2 k . Clearly, in this case after k times dividing by 2 we reach to 1. □
Corollary 1. If n is shown as n = ( a 11 ¯ ) 2 , n = ( a 001 ¯ ) 2 or n = ( a 1101 ¯ ) 2 , where a shows a consecutive 0-digits and 1-digits, then in the next step 3 n + 1 is dividable by 2 1 , 2 2 and 2 3 , respectively.
Remark 1. In what follows we prove some lemmas which show that the generated numbers are ultimately converted to the given number in Lemma 2.1.
Lemma 2.2. Let n = ( 0 11 1 k 1 t i m e s ) 2 , that is, the first zero appears in k-th place. Then, the place of this zero is reduced to one unit at each iteration.
Proof. 
If n = ( 0 11 1 k 1 t i m e s ) 2 , then 3 n + 1 = ( 0 11 1 k 2 t i m e s 0 ) 2 . Thus, using corollary 1 in the next iteration 3 n + 1 is dividable to 2 1 and as a result n = ( 0 11 1 k 1 t i m e s ) 2 3 n + 1 2 = ( 0 11 1 k 2 t i m e s ) 2 3 n + 1 2 = ( 01 ) 2
Lemma 2.3. If n = ( a 01 ¯ ) 2 , then instead of computing 3 n + 1 it is enough to compute 3 a + 1 .
Proof. 
In this case by corollary 1 it is clear that 3 n + 1 is dividable by 4 = 2 2 . Suppose n = ( a 4 a 3 01 ) 2 , so in the next step 3 n + 1 = ( ( a 4 + a 3 ) ( a 3 + 1 ) 00 ) 2 . On the other hand
3 a + 1 = 3 ( a 4 a 3 ) 2 + 1 = ( 11 ) 2 ( a 4 a 3 ) + 1 = ( ( a 4 + a 3 ) ( a 3 + 1 ) ) 2 = 3 n + 1

3. Main Result

Now, we give a simple proof for Collatz conjecture.
Theorem The Collatz conjecture is correct.
Proof. 
Let n = ( a k a 2 a 1 ) 2 . If a 1 = 0 , then we remove it (dividing by 2). Without loss of generality assume that a k = a 1 = 1 . If a 2 = 1 ,then using Lemma 2.2 instead of working on n, after some iterations, we consider n = ( 01 ) 2 and finally by Lemma 2.3 we use n = ( 101 ) 2 , instead of n , and Lemma 2.1 shows that the sequence with n is convergence to 1. □
In the following, we show that obtaining the sequence in Lemma 2.1 occurs much faster.
Lemma 3.1. Let n = ( 0 11 1 k t i m e s 0 ) 2 , k > 1 , then in the next step 3 n + 1 = ( 0 11 1 k 1 t i m e s 0 ) 2 or 3 n + 1 = ( 0 11 1 k 2 t i m e s 01 ) 2 .
Proof. 
Let n = ( 0 11 1 k t i m e s 0 ) 2 , then using the multiplication property of 3 = ( 11 ) 2 we have 3 n + 1 = ( * 0 1 1 k 1 t i m e s 0 * ) 2 depends on the previous sums give us a 1 or 3 n + 1 = ( * 0 1 1 k 2 t i m e s 01 * ) 2 depends on we do not have 1, where * stands for a consecutive 0-digits and 1-digits. Therefore in both cases, we have a sequence as 1 1 q t i m e s , where q is k 1 or k 2 , and after at most k 1 times this sequence is converted to ( * 010 * ) 2 . □
Indeed, the above lemma shows that at each iteration at least one of 1-digits is reduced and ultimately the given number approached to mentioned number in Lemma 2.1.
Lemma 3.2. Let n = ( 1 0 0 k t i m e s 1 ) 2 . Then, after some iterations we have 3 n + 1 = ( 101 ) 2 .
Proof. 
By corollary 1, in the next step 2 2 is a factor of 3 n + 1 and 3 n + 1 2 2 = ( 1 0 0 k 2 t i m e s 1 ) 2 . Hence, if k is an odd number and k = 2 l + 1 , then after l iteration we reach to a number in which its two right digits are 01. If k is even the sequence leads to a number whose two right digits are 11. In this case by employing Lemma 3.1 we reach to a number with 01 in its right digits. □
In a similar way, we get ( * 101 * ) 2 for the case where there is a sequence of zeros between two ones.
Lemma 3.3. (The worst case) Let n = ( 11 1 ) 2 then after four iteration we reach to a sequence that its three left digits is 111 or 10101, periodically. That is, in this worst case, the Collatz’s sequence goes to 5 and then to 1.
Proof. 
Since n = ( 11 1 k t i m e s ) 2 , thus 3 n + 1 = ( 1011 ) 2 3 n + 1 = ( 100011 ) 2 3 n + 1 = ( 110101 ) 2 3 n + 1 = ( 10100001 ) 2 and after this step three digits on the left have the following periodic behavior: 111 10101 111

References

  1. P.J. Andaloro, The 3x + 1 Problem and Directed Graphs. Fibonacci Q. 40 (2002), 43–54.
  2. C. Böhm, G. Sontacchi, On the existence of cycles of given length in integer sequences like xn+1 = x n 2 if xn even, and xn+1= 3xn + 1 otherwise, Atti Accad. Naz. Lincei VIII. Ser. Rend. Cl. Sci. Fis. Mat. Nat. 64 (1978), 260–264.
  3. S. Eliahou, The 3x + 1 problem: New lower bounds on nontrivial cycle lengths, Discret. Math., 118 (1993), 45–56. [CrossRef]
  4. F. Emmert-Streib, Structural properties and complexity of a new network class: Collatz step graphs, PLoS ONE., 8 (2013), e56461. [CrossRef]
  5. J.C.L. Ilia Krasikov, Bounds for the 3x + 1 problem using difference, Acta Arith., 109 (2003), 237–258. [CrossRef]
  6. T. Laarhoven, B. T. Laarhoven, B. de Weger, The Collatz conjecture and De Bruijn graphs, Indag. Math., 24 (2013), 971–983.
  7. J. Lagarias, The Ultimate Challenge: The 3x + 1 Problem, Providence, RI, USA, (2010). [CrossRef]
  8. J.A.T. Machado, A. Galhano, D. Cao Labora, A clustering perspective of the Collatz conjecture, Math., 9(4) (2021), 314. [CrossRef]
  9. B. Snapp, M. Tracy, The Collatz Problem and Analogues, J. Integer Seq., 11 (2008), 1–10.
  10. T. Tao, Almost all orbits of the Collatz map attain almost bounded values, arXiv, arXiv:1909.03562v3 (2019). [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.
Prerpints.org logo

Preprints.org is a free preprint server supported by MDPI in Basel, Switzerland.

Subscribe

Disclaimer

Terms of Use

Privacy Policy

Privacy Settings

© 2025 MDPI (Basel, Switzerland) unless otherwise stated