See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/325389892

Solution to Collatz's Conjecture

Thesis · May 2018

CITATION READS
1 12,104

1 author:

Jose William Porras

Escuela Naval Almirante Padilla Colombia
77 PUBLICATIONS   81 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Oceanografía Dinámica de la Cuenca de Colombia View project

Galeon San Jose View project

All content following this page was uploaded by Jose William Porras on 07 December 2019.

The user has requested enhancement of the downloaded file.

 Solution to Collatz´s Conjecture

José William Porras Ferreira

Centro de Investigaciones Científicas, Escuela Naval de Cadetes

“Almirante Padilla”,
Isla Manzanillo, Cartagena de Indias, Colombia

Abstract

Collatz´s conjecture, also known as conjecture 3𝑛 + 1, conjecture of Ulam or the

problem of Syracuse, is a conjecture of number theory established by Lothar
Collatz in 1937 and says the following: If 𝑛 is an even number, divide it by 2 until
you reach an odd number or 1, if 𝑛 is an odd number different from 1, multiply it by
3 and add 1 and start the process of dividing by 2, the final result will always be 1.
3𝑛+1
The above means that this process always ends when 2 = 2𝑘 being 𝑘 ≥ 1,
because the successive division k times by 2 of 2𝑘 is 1. Formally, this is the same
to a function 𝑓: ℕ ↦ ℕ, such that:
𝑛
, for even 𝑛
2
3𝑛+1
, for odd 𝑛
𝑓(𝑛) = 2
3𝑛+1
= 2𝑘 or
( 2 ) for 𝑘 ≥ 1, the process ends
{ 𝑛=2 𝑘

Here an analysis of this conjecture is presented, which allows to find a solution.

Keyword: Collatz´s conjecture, Ulam conjecture, the Syracuse problem.

1. Introducción

Collatz´s Conjecture, Ulam Conjecture or the Syracuse problem, also known as

conjecture 3𝑛 + 1, [1], was enunciated by the German mathematician Lothar
Collatz (1910-1990) in 1937 and its solution is not known. This conjecture is
expressed as follows: If n is an even number, divide by 2 until you reach an odd
number or 1, if 𝑛 is an odd number different from 1, multiply by 3 and add 1 and
start the process of dividing by 2, the final result will always be 1. The above
3𝑛+1
means that this process always ends when = 2𝑘 , or 𝑛 = 2𝑘 being 𝑘 ≥ 1,
2
𝑘
because the successive division 𝑘 times by 2 of 2 is 1. Formally, this is equals to
𝑛
2
, for even 𝑛
3𝑛+1
, for odd 𝑛
A function 𝑓: ℕ ↦ ℕ, such that 𝑓(𝑛) = 2
3𝑛+1
= 2𝑘 or
( 2 ) for 𝑘 ≥ 1, the process ends
{ 𝑛 = 2𝑘

1
The importance of this conjecture in the numerical analysis, lies in the fact that the
function 𝑓(𝑛), always arrive to 1 for any 𝑛 ∈ ℕ, that is, within the numerical analysis
there is a function that arrive to the unit regardless the value of 𝑛.

The 3𝑛 + 1 conjecture is simple to enunciate and apparently difficult to solve. Paul

Erdös (1913-1996), known as the man who only loved numbers, commented on
the intractability of problem 3𝑛 + 1, [2]: "Mathematics is not ready for those
problems yet". This conjecture has interesting connections with the Diophantine
approximation of the binary logarithm of 3 and the mod 1 distribution of the
3
sequence {(2)𝑘 : 𝑘 = 1, 2, … }, with questions from the ergodic theory about the 2-
adic integers1, for example, the "2-adic" or binary expansion of 39 is 1 ∙ 25 + 0 ∙
24 + 0 ∙ 23 + 1 ∙ 22 + 1 ∙ 21 + 1 ∙ 20 , written in binary notation is: 1001112, [3] and
with the theory of computability: it has been shown that a generalization of the 3𝑛 +
1 problem is a computationally insoluble problem, [2].

This conjecture has even allowed to draw a fractal map about the real numbers, as
shown in figure 1, [4].

Figure 1. Fractal map on the real numbers of the Collatz´s conjecture. in /

Pokipsy76 / Wikimedia

1
If we set a prime number 𝑝, then any integer can be written as a p-adic expansion (which is usually said to
write the number in "base 𝑝") in the form: ± ∑𝑛𝑖=−∞ 𝑎𝑖 𝑝𝑖

2
Any counterexample to the Collatz´s conjecture should consist of a closed cycle
different from the trivial cycle (4; 2; 1) or an infinitely divergent trajectory (infinite
odd numbers that do not meet the Collatz´s conjecture). Therefore, if it could be
proved that none of these types of counterexamples exists, then all positive
integers would have a trajectory that would reach the trivial cycle, [2]-[5].

In the case of the Collatz´s conjecture, the type of cycle can be defined with
reference to the definition of "direct access" of the Collatz´s map, 𝑓(𝑛) = (3𝑛 +
1)/2 for odd 𝑛 and 𝑓(𝑛) = 𝑛/2 for even 𝑛. By definition, a cycle is a sequence
(𝑎0 ; 𝑎1 ; … ; 𝑎𝑛 ) where 𝑓(𝑎0 ) = 𝑎1 , 𝑓(𝑎1 ) = 𝑎2 and so on up to 𝑓(𝑎𝑛 ) = 𝑎0 closing
the cycle when returning to the initial value of this cycle. For this definition of direct
access, the only known cycle is (1; 2). Although 4 is part of the only known cycle
for the original Collatz´s map, it is not part of the cycle for the direct access map,
[5].

A cycle k is a cycle that can be divided into 2k contiguous subsequences: k

increasing sequences of odd numbers alternating with k decreasing sequences of
even numbers. For example, if the cycle consists of a single incremental
subsequence of odd numbers followed by a decreasing subsequence of even
numbers, it is called a cycle of 1. [6], [7]-[8].

Steiner (1977). [6], showed that there is no cycle of 1 that is not trivial (1; 2).
Simons (2004), [9], used Steiner's method to demonstrate that there is no cycle of
two. Simons & de Weger (2005) extended this test up to 68 cycles: there is no k-
cycle up to k = 68. [5]. Beyond 68, this method gives higher limits for the elements
in that cycle: for example, if there is a cycle of 75, then at least one element of the
cycle is less than 2385 × 250. [6]. Therefore, as the exhaustive searches of the
computer continue, larger cycles can be discarded. To express the argument in a
more intuitive way: it is not necessary to look for cycles that have a maximum of 68
trajectories, where each trajectory consists of consecutive ascents followed by
consecutive descents, [6].

Also B. G, Seifert (1998), [10] studied in detail the arithmetic cycles of the Collatz’s
conjecture.

In the proposed test, it can be determined that there can be neither infinite
divergent trajectories nor closed cycles that do not lead to a final value of 1, as
expressed by the Collatz´s conjecture.

2. Analysis of the Collatz´s Conjecture

The conjecture says: If 𝑛 is an even number, divide it by 2 until it reaches an odd

number or 1, if 𝑛 is an odd number different from 1, multiply it by 3 and add 1 and

3
start the process of dividing by 2, the final result always must reach 1, that is, if 𝑛 is
an even number of the form 𝑛 = 2𝑘 for 𝑘 ≥ 1 the process ends, because the
successive division k times by 2 will arrive until 𝑛 = 1, but if the result is an odd
3𝑛+1
number 𝑛 ≥ 3, applies 2 until it reaches a number (3𝑛 + 1)/2 = 2𝑘 . For example
6 3𝑛 +1 3∗3+1 3𝑛 +1 3∗5+1
𝑛 = 6, 𝑛1 = 2 = 3, 𝑛2 = 12 = 2 = 5, 𝑛3 = 22 = 2 = 8 = 23 . The process
ends because the successive division 3 times by 2 of 23 ends in 1.
2.1 Proof of the Collatz´s Conjecture

1. There are direct solutions with even numbers 𝑛 = 2𝑘 for 𝑘 ≥ 1. The successive
division 𝑘 times by 2, has a final result of 1.
2. The odd numbers 𝑛 = 1 + 2𝑡 for 𝑡 ≥ 1 would have a direct solution when (3𝑛 +
1)/2 = 2𝑘 where 𝑡 = (2𝑘 − 2)/3 for odd 𝑘 ≥ 3, taking note that always 2𝑘 − 2 is
divisible by 3 when 𝑘 ≥ 3 is odd..
Proof:

a. Let 𝑛 = 1 + 2𝑡 for 𝑡 ≥ 1
3𝑛+1
b. According to Collatz´s conjecture 2 = 2𝑘 so that the final result of the
successive division by 2 of 2𝑘 will be 1 ……..………….……… According to 1.
3𝑛+1
c. 2 = 2 + 3𝑡 = 2𝑘 ……………………………..…..…………… According to 2b.
d. 𝑡 = (2𝑘 − 2)/3 …………………………………..………………. According to 2c.
e. As (2𝑘 − 2) is always divisible by 3 when 𝑘 ≥ 3 is odd, , therefore 𝑡 would be
an integer number of the form 𝑡 = (2𝑘 − 2)/3 and 𝑛 = 1 + 2(2𝑘 − 2)/3, which
3𝑛+1
leads to a direct solution of 2 = 2𝑘 ⇒ 𝑘 ≥ 3 odd, that by successive
division k times by 2 of 2𝑘 ends in 1.
In this way the proof of point 2 is completed.
Table 1 gives examples of the above.
3𝑛+1
Table 1. Values of 𝑘, 𝑡, 𝑛 and 2 = 2𝑘 where the Collatz´s conjecture is
directly met, by successive divisions of 𝑘 times by 2 of 2𝑘
3𝑛 + 1
𝑘 𝑡 = (2𝑘 − 2)/3 𝑛 = 1 + 2𝑡 = 2𝑘
2
3 2 5 8
5 10 21 32
7 42 85 128
9 170 341 512
11 682 1365 2048
13 2730 5461 8192
15 10922 21845 32768
17 43690 87381 131072
19 174762 349525 524288

4
 3𝑛 + 1
𝑘 𝑡 = (2𝑘 − 2)/3 𝑛 = 1 + 2𝑡 = 2𝑘
2
21 699050 1398101 2097152
23 2796202 5592405 8388608
3𝑛 +1
3. Now, each direct solution 12 = 2𝑘 , generates another solution of the type 𝑛2 =
(𝑛1 2𝑘1 − 1)/3, which in turn generates another solution 𝑛3 = (𝑛2 2𝑘2 − 1)/3,
where [𝑘1 , 𝑘2 , … , 𝑘𝑚−1 ] ≥ 1 are the exponent that allows the factors [(𝑛1 2𝑘1 −
1), 𝑛2 (2𝑘2 − 1), . . . , 𝑛𝑚−1 (2𝑘𝑚−1 − 1), divisible by 3. The previous process ends
when the factor 𝑛𝑚 is divisible by 3, because (𝑛𝑚 2𝑘𝑚 − 1) would no longer be
divisible by 3. Each solution 𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 can generate infinite solutions of
odd numbers 𝑛 where 𝑛 = ([𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ]2𝑝 − 1 )/3 only varying the
exponent 𝑝 ≥ 1 for each case and allowing 3|([𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ]2𝑝 − 1 ).
Proof:
3𝑛 +1
a. Let 𝑛1 = (2𝑘+1 − 1)/3, for 𝑘 ≥ 1 + 2𝑡, for 𝑡 ≥ 0, therefore 12 = 2𝑘 , leads to
a direct solution by successive division k times by 2 of 2𝑘 .
b. Let 𝑛2 = (𝑛1 2𝑘1 − 1)/3, where 𝑘1 ≥ 1 is the exponent that allows (𝑛1 2𝑘1 − 1
to be divisible by 3.
3𝑛2 +1
c. = 𝑛1 2𝑘1 −1, according to 2b. and the successive division (𝑘1 − 1) times
2
by 2 of 2𝑘1 −1 leads to 𝑛1 , which in turn is a solution of 3a. and therefore in
total, there are a successive divisions (𝑘 + 𝑘1 − 1) times by 2 until it reaches
1, fulfilling the Collatz´s Conjecture.
d. In the same way 𝑛2 , … , 𝑛𝑚 would fulfill Collatz´s Conjecture up to 𝑛𝑚 odd
number where 𝑛𝑚 = (𝑛𝑚−1 2𝑘𝑚−1 − 1)/3 according to 3c., 3b. and 3a.
e. The above process ends when 𝑛𝑚 is divisible by 3, because (𝑛𝑚 2𝑘𝑚 − 1)
would no longer be divisible by 3. A multiple number of 3 minus 1 is not
divisible by 3, that is 3|𝑛𝑚 2𝑘𝑚 but 3 ∤ 1.
f. Each solution 𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 can generate infinite solutions of odd
numbers 𝑛 where 𝑛 = ([𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ]2𝑝 − 1 )/3 holding fixed
[𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ] and varying only the exponent 𝑝 ≥ 1 for each case and
3𝑛+1
allowing that 3|([𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ]2𝑝 − 1 ), because we have that 2 =
[𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ]2𝑝−1 would have a successive division (p-1) times by 2 of
3([𝑛1 ,𝑛2 ,𝑛3 ,…,𝑛𝑚−1 ])+1
2𝑝−1 until reaching [𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ], and every =
2
2[𝑘,𝑘1 −1,𝑘2 −1,…,𝑘𝑚−1 −1] respectively has a successive division [𝑘, 𝑘1 − 1, 𝑘2 −
1, … , 𝑘𝑚−1 − 1] times by 2 of 2[𝑘,𝑘1 −1,𝑘2 −1,…,𝑘𝑚−1 −1] until reaching 1, according
to 3a., 3b. 3c. and 3d, fulfilling the Collatz´s Conjecture.
Proof:

a. Every odd number [𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ] > 1 no multiple of three is equal to a

number b multiple of 3 plus 1 or plus 2, that is [𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ] = 𝑏 + 1 or
[𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ] = 𝑏 + 2.
b. First case [𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ] = 𝑏 + 1, therefore:

5
 [𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ]2𝑘 − 1 = (𝑏 + 1)2𝑘 − 1
. = 𝑏2𝑘 + 2𝑘 − 1

As 3|𝑏2𝑘 because 𝑏 is a multiple of 3, but it is required that 3|(2𝑘 − 1),

therefore 𝑘 = 2 + 2𝑡 for 𝑡 ≥ 0, then 2𝑘 − 1 = 22+2𝑡 − 1 = 41+𝑡 − 1, but the
decomposition of 𝑥 𝑛 − 1 in factors having 𝑥 − 1 as one of them is:
𝑛

𝑥 𝑛 − 1 = (𝑥 − 1) (∑ 𝑥 𝑛−𝑘 )
𝑘=1

Then
1+𝑡 1+𝑡
1+𝑡 1+𝑡−𝑘
4 − 1 = (4 − 1) (∑ 4 ) = 3 (∑ 41+𝑡−𝑘 )
𝑘=1 𝑘=1

Therefore 3|(22+2𝑡 − 1). Example of table 3 for 𝑛1 = 13 = 12 + 1, we have to

3|(13 ∙ 22+2𝑡 − 1) for 𝑡 ≥ 0, and it would have infinite odd numbers 𝑛 =
13∙22+2𝑡 −1 3𝑛+1
when t → ∞ who fulfill the Collatz conjecture, being = 13 ∙ 21+𝑡 .
3 2

c. Second case [𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ] = 𝑏 + 2, therefore

[𝑛1 , 𝑛2 , 𝑛3 , … , 𝑛𝑚−1 ]2𝑘 − 1 = (𝑏 + 2)2𝑘 − 1

. = 𝑏2𝑘 + 2 ∙ 2𝑘 − 1
As 3|𝑏2𝑘 because 𝑏 is a multiple of 3, but it is required that 3|(2𝑘+1 − 1),
therefore 𝑘 = 1 + 2𝑡 for 𝑡 ≥ 0, so that 3|(2𝑘+1 − 1), being a case similar to
what was demonstrated in b., because 2𝑘+1 = 22+2𝑡 = 41+𝑡 Example of table
2 we have to 𝑛1 = 5 = 3 + 2, therefore 3|(5 ∙ 21+2𝑡 − 1) for 𝑡 ≥ 0 and there
5∙22+2𝑡 −1
would be infinite odd numbers 𝑛 = when 𝑡 → ∞ who fulfill the Collatz
3
conjecture.
In this way the proof of point 3 is completed.
Tables 2 and 3 give examples of the above, for two different values of 𝑛1 and
ending when a 𝑛𝑚 is reached that is divisible by 3.

Table 2.Examples with 𝑛1 = 5 and ending when 𝑛𝑚 is divisible by 3.

𝑘1 𝑛1 𝑛2 𝑘2 𝑛3 𝑘3 𝑛4 𝑘4 𝑛5 𝑘5 𝑛6
3 13 2 17 1 11 1 7 2 9
5 53 1 35 1 23 1 15
5
7 213
9 853 2 1137

Table 3.Examples with 𝑛1 = 13 and ending when 𝑛𝑚 is divisible by 3..

6
 𝑘1 𝑛1 𝑛2 𝑘2 𝑛3 𝑘3 𝑛4 𝑘4 𝑛5 𝑘5 𝑛6 𝑘 𝑛7 𝑘7 𝑛8
2 17 1 11 1 7 2 9
4 69
13
6 277 2 369
8 1109 1 739 2 985 2 1313 1 875 1 583 2 777

Note: the demonstration that 3|(2𝑘 − 2) for 𝑘 = 1 + 2𝑡 ⇒ 𝑡 ≥ 1 is similar to the

above, just do 2𝑘 − 2 = 2(2𝑘−1 − 1) = 2(4𝑡 − 1) therefore 3|2(4𝑡 − 1).
4. As a corollary to the demonstrations of points 1., 2. and 3. We have that any
odd number 𝑛 ≥ 1, can be equal to 𝑛 = (𝑎2𝑘 − 1)/3 with 𝑎 ≥ 1 an odd
number and not a multiple of 3 for some value of 𝑘 that makes possible that
3𝑎−1
3|(𝑎2𝑘 − 1). When 𝑎 is 1, or a solution that leads to 2 = 2𝑘 , there would
be a solution to Collatz´s conjecture. If 𝑘 > 1, then 𝑛 > 𝑎. Likewise, there
are infinite odd numbers 𝑛 for the same 𝑎, it is only required to go changing
𝑘 so that 3|(𝑎2𝑘 − 1). It is clear that the only case in which. 𝑛 = 𝑎 is when
𝑛 = 1 = 𝑎, in the other cases when 𝑛 ≥ 3, we have to 𝑛 ≠ 𝑎, which is
important to take into account for the solution of the Collatz conjecture
raised in points 5 and 6 below. Table 4 gives examples of odd numbers 1 ≤
𝑛 ≤ 43, where 𝑛 = (𝑎2𝑘 − 1)/3.

Table 4. Odd numbers 1 ≤ 𝑛 ≤ 43, where 𝑛 = (𝑎2𝑘 − 1)/3

𝑎 𝑘 𝑛 = (𝑎2𝑘 − 1)/3 𝑎 𝑘 𝑛 = (𝑎2𝑘 − 1)/3

1 1 1 35 1 23
5 1 3 19 2 25
1 4 5 41 1 27
11 1 7 11 3 29
7 2 9 47 1 31
17 1 11 25 2 33
5 3 13 29 2 35
23 1 15 7 4 37
13 2 17 59 1 39
29 1 19 31 2 41
1 6 21 65 1 43

5. In the Collatz´s Conjecture, there can not be closed cycles in a sequence

(𝑎0 ; 𝑎1 ; … ; 𝑎𝑛 ) where 𝑓(𝑎0 ) = 𝑎1 , 𝑓(𝑎1 ) = 𝑎2 and so on up to 𝑓(𝑎𝑚 ) = 𝑎0 closing
the cycle when returning to the initial value of that cycle, where (𝑛0 ; 𝑛1 ; … ; 𝑛𝑚 )
are odd numbers, in:

[𝒇(𝒂𝟎 ) = (𝟑𝒏𝟎 + 𝟏)⁄𝟐; 𝒇(𝒂𝟏 ) = [𝟑𝒏𝟏 + 𝟏]⁄𝟐; … ; 𝒇(𝒂𝒎 ) = [𝟑𝒏𝒎 + 𝟏]⁄𝟐] = 𝒂𝟎

7
 Proof:

a. Assuming that there is a closed cycle (𝑎0 ; 𝑎1 ; … ; 𝑎𝑛 ) where 𝑓(𝑎0 ) = 𝑎1 ,

𝑓(𝑎1 ) = 𝑎2 and so on up to 𝑓(𝑎𝑚 ) = 𝑎0 closing the cycle when returning to
the initial value of that cycle , where (𝑛0 ; 𝑛1 ; … ; 𝑛𝑚 ) odd numbers, in:

[𝒇(𝒂𝟎 ) = (𝟑𝒏𝟎 + 𝟏)⁄𝟐; 𝒇(𝒂𝟏 ) = [𝟑𝒏𝟏 + 𝟏]⁄𝟐; … ; 𝒇(𝒂𝒎 ) = [𝟑𝒏𝒎 + 𝟏]⁄𝟐] = 𝒂𝟎

(𝑛0 ; 𝑛1 ; … ; 𝑛𝑚 ) would not fulfill Collatz´s Conjecture

b. According to the corollary in point 4. odd numbers (𝑛0 ; 𝑛1 ; … ; 𝑛𝑚 ), in their

𝑏 2𝑘𝑖 −1
respective order are equal to 𝑖 3 where 𝑖 = (0; 1; … ; 𝑚) and 𝑏𝑖 ≥ 5 an
odd number not multiple of 3 and 𝑘𝑖 ≥ 1 values that allow 3|(𝑏𝑖 2𝑘𝑖 − 1),,
𝑏 2𝑘0 −1 𝑏 2𝑘1 −1 𝑏 2𝑘𝑚 −1
that is, 𝑛0 = 0 3 ; 𝑛1 = 1 3 ; …; 𝑛𝑚 = 𝑚 3 and where 𝑛𝑖 ≠ 𝑏𝑖 ,
therefore all 𝑏𝑖 are outside the closed loop where it is assumed there is no
solution to the Collatatz conjecture.

c. Each 𝑏𝑖 originates infinite odd numbers different from (𝑛0 ; 𝑛1 ; … ; 𝑛𝑚 ), with

𝑘 ≠ 𝑘𝑖 such that 3|(𝑏𝑖 2𝑘 − 1), and these can be reduced to 𝑏𝑖 with solution in
Collatz´s Conjecture, because each one is outside the closed cycle that has
been assumed has no solution in Collatz´s Conjecture, therefore there would
be a contradiction to what was shown in 3., here there are for the same 𝑏𝑖
odd numbers with solutions in Collatz´s Conjecture and odd numbers
(𝑛0 ; 𝑛1 ; … ; 𝑛𝑚 ) without solutions in Collatz´s Conjecture, which is
contradictory and impossible.

d. According to c. what it is assumed in a. is false and there are no closed

cycles in Collatz´s Conjecture that not fulfill this conjecture

6. In Collatz´s Conjecture there can not be a divergent (infinite) sequence of odd

numbers that do not meet this conjecture.

Proof:

a. Assuming that there is an infinite sequence of odd numbers

(𝑛0 ; 𝑛1 ; … ; 𝑛𝑚→∞ ), such that:

[𝒇(𝒂𝟎 ) = (𝟑𝒏𝟎 + 𝟏)⁄𝟐; 𝒇(𝒂𝟏 ) = [𝟑𝒏𝟏 + 𝟏]⁄𝟐; … ; 𝒇(𝒂𝒎→∞ ) = [𝟑𝒏𝒎→∞ + 𝟏]⁄𝟐]

These do not fulfill Collatz´s conjecture.

b. According to the corollary in point 4., odd numbers (𝑛0 ; 𝑛1 ; … ; 𝑛𝑚→∞ ), in

𝑏 2𝑘𝑖 −1
their respective order are equal to 𝑖 3 where 𝑖 = (0; 1; … ; 𝑚 → ∞) and
and 𝑏𝑖 ≥ 5 an odd number not multiple of 3 and 𝑘𝑖 ≥ 1 values that allow

8
 𝑏 2𝑘0 −1 𝑏 2𝑘1 −1 𝑏 2𝑘𝑚→∞ −1
3|(𝑏𝑖 2𝑘𝑖 − 1),, that is, 𝑛0 = 0 3 ; 𝑛1 = 1 3 ; …; 𝑛𝑚→∞ = 𝑚→∞ 3 and
where 𝑛𝑖 ≠ 𝑏𝑖 , therefore all 𝑏𝑖 are outside the infinite sequence, where it is
assumed there is no solution to the Collatatz conjecture.

e. Each 𝑏𝑖 originates infinite odd numbers different from (𝑛0 ; 𝑛1 ; … ; 𝑛𝑚→∞ ), with
𝑘 ≠ 𝑘𝑖 such that 3|(𝑏𝑖 2𝑘 − 1), and these can be reduced to 𝑏𝑖 with solution in
Collatz´s Conjecture, because each one is outside the divergent sequence
that has been assumed has no solution in Collatz´s Conjecture, therefore
there would be a contradiction to what was shown in 3.. Here there are for
the same 𝑏𝑖 odd numbers with solutions in Collatz´s Conjecture and odd
numbers (𝑛0 ; 𝑛1 ; … ; 𝑛𝑚→∞ ) without solutions in Collatz´s Conjecture, which
is contradictory and impossible.

c. According to c. what it is assumed in a. is false and there is no infinite

sequence of odd numbers that does not meet Collatz´s Conjecture.
This allow to conclude that within the process of multiplying by 3 any odd number,
adding 1 and dividing by 2, there are no closed cycles or infinite sequences of odd
numbers that do not meet Collatz´s Conjecture, then, this conjecture has being
demonstrated.
Conclusion
It was shown that there are direct solutions of Collatz´s conjecture for even
3𝑛+1
numbers 𝑛 = 2𝑘 , odd numbers 𝑛 where 2 = 2𝑘 and odd numbers 𝑛1 where 𝑛1 =
(𝑛2𝑚 − 1)/3. Likewise any odd number 𝑛 ≥ 1 can be equal to (𝑎2𝑘 − 1)/3 with
𝑎 ≥ 1 being an odd number and not a multiple of 3 for some value of 𝑘 that makes
3𝑛−1
it possible that 3|(𝑎2𝑘 − 1). When 𝑎 is 1, or a solution that leads to 𝑎 = 2 = 2𝑘 ,
there would be a solution to the Collatz´s conjecture, equally there are infinite odd
numbers for the same odd 𝑎 not multiple of 3. Because to the above, it was
possible to demonstrate by contradiction that there are no closed cycles or
divergent sequences of odd numbers that do not fulfill Collatz´s conjecture,
therefore the process of multiplying by three any odd number, adding 1 and the
result dividing it by 2, can not contain closed cycles or the process is infinite, but it
is bounded by the proofs of points 1, 2, 3 and 4 of the test of Collatz´s conjecture,
and therefore this conjecture is true.
About the authors:
Jose William Porras-Ferreira, wporras@enap.edu.co or jwporras@balzola.org
Research Ex-professor at the Colombian Naval Academy, Member of the US. Etta
Kappa Nu Society, Member of the Spanish Foundation Charles III. Electronic
engineer, Master and EE from the Naval Postgraduate, Monterey California USA.

Acknowledgments:

9
 I am grateful to many mathematicians and colleagues for their numerous
comments along many years of research and my professors at the Colombian
Naval Academy and at the Naval Postgraduate School, Monterey California, USA.

References:

[1]. Maddux, Cleborne D.; Johnson, D. Lamont (1997). Logo: A Retrospective.

New York: Haworth Press. p. 160. ISBN 0-7890-0374-0. The problem is also
known by several other names, including: Ulam's conjecture, the Hailstone
problem, the Syracuse problem, Kakutani's problem, Hasse's algorithm, and
the Collatz´s problem.
[2]. Jeffrey C. Lagarias. (1985). The 3x + 1 Problem and Its
Generalizations. American Mathematical Monthly; Vol. 92, No. 1, pp 3-23,
DOI: 10.2307/2322189, JSTOR 2322189.
[3]. J. Blazewiez and A. Pettorossi. Collatz’s Conjecture and Binary sequence.
Institute of Control Engineering. Technical University of Poznan. Poznan,
Poland (preprint) March 1983.
[4]. Página web: http://www.microsiervos.com/archivo/ciencia/conjetura-de-
Collatz´s.html, visitada el 22-05-2018.
[5]. Simons, J.; de Weger, B. (2003). "Theoretical and computational bounds
form-cycles of the 3n + 1 problem" (PDF). Acta Arithmetica. 117 (1): 51–
70. DOI:10.4064/aa117-1-3.
[6]. Steiner, R. P. (1977). "A theorem on the Syracuse problem". Proceedings of
the 7th Manitoba Conference on Numerical Mathematics. pp. 553–9.
[7]. C. Böhm and G. Sontacchi. On the existence of cycles of given length in
integer sequences like 𝑛𝑛+1 = 𝑥𝑛 /2 f 𝑥𝑛 even, and 𝑥𝑛+1 = 3𝑥 + 1 otherwise.
Atti Accad. Naz. Lincei Rent. Cl. Sci. Fis. Mat. Natur., 64 (1978) 260-264, MR
83h.10030.
[8]. Jeffrey C. Lagarias (Murray Hill, N.J.). (1990). The set of rational cycles for
the 3x+1 problem. Acta Arithmetica. LVI pp 33-53.
[9]. Simons, John L. (2005). "On the nonexistence of 2-cycles for the 3x+1
problem". Math. Comp. 74: 1565–72. DOI: 10.1090/s0025-5718-04-01728-
4. MR 2137019
[10]. B. G, Seifert. (1998) On the arithmetic of cycles for the Collatz Hasse
(Syracuse) conjectures. Discrete Math 68, pp 293-298

Conflict of interest

I don't have any conflict of interest, since this work is the product of the research
done by us rely on the references described in the manuscript and my
mathematical knowledge on the subject of the article

10

View publication stats