Preprint
Article

This version is not peer-reviewed.

A Proof of Collatz Conjecture

Submitted:

21 September 2023

Posted:

22 September 2023

You are already at the latest version

Abstract
In 1937,German mathematician L.Collatz proposed the following conjeture:for any positive integer,if it is even,divide it by 2,if it is odd,multiply it by 3 and add 1 to get an even number.Continuing with the above rule, the final result will be 1.This paper gives a proof of this conjecture.
Keywords: 
;  

§1. Introduction

In 1937, German mathematician L. Collatz proposed the following conjecture that for a definite positive integer p, if p is even, divide it by 2,if p is odd, multiply it by 3 and add 1 to get an even number.Continuing with the above rule, the final result will be 1.
Collatz conjecture has been studied by many people and has long been regarded as an unsolved problem.See [1,2,3]. This paper gives a proof of this conjecture.

§2. Preliminaries

Definition 2.1: Starting from a positive integer p, the process with the rule in introduction is called a Collatz sequence.p is said to be successful if 1 is finally obtained.Otherwise,p is not successful.
Remark: Any positive integer p can be written as that p = 2k(2m-1),where m is a positive integer and k is a positive integer or 0. When k > 0, p is even,and when k = 0, p=2m-1 is odd. If any odd number is successful, then since the even number p = 2k(2m-1) is divided by 2 k times to get the odd number 2m-1, the even number p = 2k(2m-1) is also successful. In other words, to prove that Collatz conjecture holds, it is sufficient to show that any odd number is successful.
In this paper, the following discussion focuses on the Collatz sequence for odd numbers, and the even numbers in the Collatz sequence are omitted.
Example: For a positive integer p = 44, its Collatz sequence is that 44, 22, 11, 34, 17, 52, 26, 13, 40, 20, 10, 5, 16, 8, 4, 2, 1,4,2,1,.... By removing all even numbers, it becomes that 11, 17, 13, 5, 1,1....We will omit the duplicate odd number 1 later on.It becomes that 11, 17, 13, 5, 1. That is, the odd number 11 is successful. This implies that odd 17, odd 13, odd 5, odd 1 are all successful. This also implies that even 2k , even 2k × 5, even 2 k × 13, even 2 k × 17, even 2 k × 11 are all successful,where k = 1, 2, 3, ....
Definition 2.2: A sequence obtained by omitting all even numbers in a Collatz sequence starting with an odd number p is called a Collatz odd sequence.p is said to be successful if 1 is finally obtained.Otherwise,p is not successful.
Remark: Obviously,whether an odd number p is successful,and Definition 2.1 and Definition 2.2 are equivalent.
Definition 2.3: For two odd numbers p and q in a Collatz odd sequence, if 3p + 1= 2k q ,where k > 0 is a positive integer, then p is said to be a front odd number of q ,and q is said to be a back odd number of p.
In the previous example , because 11 × 3 + 1 = 17 × 2, where k = 1, so 11 is a front odd number of 17, and 17 is a back odd number of 11. Similarly, because 17 × 3 + 1 = 13 × 2 2 , where k = 2, so 17 is a front odd number of 13, 13 is a back odd number of 17, etc.
Theorem 2.1: If p is a front odd number of q,then 4p+1 is also a front odd number of q. Further if 3p+1 = 2 kq,where k > 0 is a positive integer, then 3(4p+1)+1 = 2k+2q.
Proof: Since p is a front odd number of q, there exists a positive integer k > 0,such that 3p+1 = 2kq. So 3(4p+1)+1 = 12p+4 = 4(3p+1) = 4 × 2 kq= 2k+2q. ■
Consider the sequence {an}:a1 ,a2 ,..., an ,...,where a1 is an odd number, and an = 4an-1 + 1,n = 2,3,4,....
Corollary 2.2: For the above sequence {a n}, if the odd number a1 is a front odd number of the odd number p, then each term in the sequence {an} is a front odd number of p.Further if 3a1 + 1 = 2kp, then 3a2 + 1 = 2k+2 p,... ,3an + 1 = 2k+2(n-1)p, where n = 2,3,4,....
Proof: This is a direct corollary of Theorem 2.1.■
Theorem 2.3: For any odd number p,there is a back odd number q, and q is the unique back odd number of p.
Proof: According to the rule of generating a Collatz odd sequence,it is immediate.■
Theorem 2.4: In a Collatz odd sequence,if p is a front odd number of p,(Of course,,p is also a back odd number of p.)then it is only case that p = 1.
Proof: Since p is a front odd number of p,there exists a positive integer k > 0,such that 3p+1 = 2kp. So 2kp -3p = p(2k -3) = 1,where p is odd.
Case 1: k = 1. Then p(2-3) = -p = 1, which cannot be true.
Case 2: k > 1. Then 2k -3 is a positive integer,and p(2k -3) = 1.This equation can be true ,only if p = 1 and k = 2. ■

§3. Bijections between Odd Subsets and Sequence Sets

This paragraph gives two classification methods of the odd number set ,and gives a classification of sequence sets of the odd number set,and gives a bijection between odd subsets and sequence sets.
The odd number set can be divided into three subsets {4n-1∣n = 1,2,3,...}, {4n+1∣n is 0 or even}= {8n+1∣n = 0,1,2,3,...},{4n+1∣n is odd}.They are disjoint with each other.
Definition 3.1:Write A ={4n-1∣n = 1,2,3,...},B ={4n+1∣n is 0 or even}= {8n+1∣n = 0,1,2,3,...},C ={4n+1∣n is odd}.
Definition 3.2: (1)Write S1 = {{an}∣a1∈ A,and an = 4an-1+1,n = 2,3,4,...},where {an} is a sequence with the first term a1 ∈ A,and an = 4an-1+1,n = 2,3,4,....
(2)Write S2 = {{an}∣a1∈ B,and an = 4an-1+1,n = 2,3,4,...},where {an} is a sequence with the first term a1∈ B,and an = 4an-1+1,n=2,3,4,....
Examples: The first two sequences in S1 are the following {an} and {bn},where {an }:3,13,53,..., (a1 = 3); {bn }:7,29,117,..., (b1 = 7).
The first two sequences in S2 are the following {an} and {bn},where {an }:1,5,21,85,..., (a1 = 1); {bn }:9,37,149,..., (b1 = 9).
Theorem 3.1: All terms (odd numbers) in all sequences of S1 and S2 contain all odd numbers. And each odd number must be in the only one of these sequences.
Proof: Note that all odd numbers can be divided into three subsets A,B,C. Any odd number in A can be the first term a1 of some definite sequence {an}in S1.Similarly,any odd number in B can be the first term a1 of some definite sequence {an}in S2. So it is enough to prove that all terms of all sequences in S1 and S2 contain any odd number with the form 4n+1 (n is odd) in C, and that any odd number in C must be in the only one of these sequences.
Let p = 4n1 + 1 ∈ C ,that is any definite odd number in C, where n1 is a definite odd number. But all odd numbers can be divided into three subsets A,B,C.
Case 1: If n1 ∈ A, then n1 is the first term a1 of some definite sequence {an}in S1.By the construction of the sequence {an} in S1, p = 4n1 + 1 is the second term a2 of this definite sequence {an} in S1.Similarly, if n1 ∈ B, then p = 4n1 + 1 is the second term a2 of some definite sequence {an} in S2.
Case 2: If n1 ∈ C, then there exists a definite odd number n2 , such that n1 = 4n2 + 1. If n2 ∈ A (or ∈ B), then according to Case 1, n1 = 4n2 + 1 is the second term a2 of some definite sequence {an}in S1 (or S2), and p = 4n1 + 1 is the third term a3 of this definite sequence {an}in S1 (or S2).
Case 3: If n 2 ∈ C, then we are back to the beginning of case 2. Continuing, since p = 4n1 + 1 is a definite odd number in C, and the odd number 1 ∈ B and 3 ∈ A ,Case 2 cannot occur indefinitely.So, after finitely many case 2, we can always get some nk , so that nk-1 = 4nk + 1, n1 > n2 > ... > nk, n1,n2,...,nk-1∈ C and nk ∈ A (or ∈ B), where nk is the first term a1 of some definite sequence {an}in S1 (or S2) ,nk-1 is the 2nd term a2 of the sequence ,...,n1 is the kth term ak ,and p = 4n1 + 1 is the k+1st term ak+1 of the definite sequence {an}in S1 (or S2). ■
Note that all odd numbers can again be divided into three subsets : {6n-3| n = 1,2,3,...}, {6n-1| n = 1,2,3,...}, {6n+1| n = 0,1,2,3,...}.
Definition 3.3: Write D = {6n-1| n = 1,2,3,...},E = {6n+1| n = 0,1,2,3,...},and F ={6n-3| n = 1,2,3,...}. Since 6n-3 is divisible by 3,an odd number in F is also called a triple odd number.
Lemma 3.2: The odd number 4n-1 in A is a front odd number of the odd number 6n-1 in D, or 6n-1 is the back odd number of 4n-1,where n = 1,2,3,....
Proof: Since 3(4n-1) + 1 = 2(6n-1),the result holds. ■
Note that B ={4n+1∣n is 0 or even}= {8n+1∣n = 0,1,2,3,...}.
Lemma 3.3: The odd number 8n+1 in B is a front odd number of the odd number 6n+1 in E, or 6n+1 is the back odd number of 8n+1,where n = 0,1,2,3,....
Proof: Since 3(8n+1) + 1 = 22(6n+1),the result holds. ■
Lemma 3.4: The triple odd number 6n-3 in F cannot be a back odd number of any odd number, where n = 1,2,3,....
Proof: Suppose that 6n-3 is a back odd number of some odd number p= 2m-1.Then 3(2m-1)+1 = 2k(6n-3) holds for some positive integer k. At this point, the right side 2k(6n-3) of the equation is divisible by 3, while the left side 3(2m-1)+1 of the equation is not divisible by 3. Contradiction. ■
Theorem 3.5:There exists a bijection ƒ1: ƒ1(D)= S1,such that for any 6k-1 ∈ D,ƒ1(6k-1) = {an} = {an∣a1 = 4k-1,an = 4an-1 +1 ,n = 2,3,4,...}∈ S1 ,k = 1,2,3,..., where all front numbers of 6k-1 are exactly all terms of the sequence {an},and 3an + 1 = 22n-1(6k-1) (n = 1,2,3,...);conversely, all terms of the sequence {an} have the unique back number 6k-1.
Proof: By Lemma 3.2,for a definite positive integer k,the first term 4k-1 of the sequence {an} is a front odd number of the odd number 6k-1.Then by Corollary 2.2, each term of {an} is a front odd number of 6k-1,and 3an + 1 = 22n-1(6k-1) (n = 1,2,3,...).We prove that any front odd numbers of 6k-1 is contained in this sequence {an}.
If an odd number p is not in {an},then by theorem 3.1,p is a term of some other sequence {bn}.We prove that p cannot be a front odd number of 6k-1.
Case 1:{bn}∈ S1.Then the first term of {bn} is b1,and b1 = 4s-1,where s is a definite positive integer with s ≠ k. As above, each term of {bn} is a front odd number of 6s-1.So,p is also a front odd number of 6s-1.By Theorem 2.3,6s-1 is the unique back odd number of p.Since s ≠ k,6k-1 cannot be a back odd number of p.In other words,p cannot be a front odd number of 6k-1.
Case 2:{bn}∈ S2.Then b1 = 8s+1,where s is 0 or a definite positive integer.Then by Lemma 3.3,b1 = 8s+1 is a front odd number of 6s+1.Then by Corollary 2.2, each term of {bn} is a front odd number of 6s+1.So,p is also a front odd number of 6s+1.Since D ∩ E = Ø,6k-1 ∈ D and 6s+1 ∈ E,6k-1 ≠ 6s+1.By Theorem 2.3, 6k-1 cannot be a back odd number of p.In other words,p cannot be a front odd number of 6k-1.
Therefore, any front odd numbers of 6k-1 is contained in this sequence {an}.So,for any odd number 6k-1 in D,ƒ1(6k-1) = {an∣a1 = 4k-1,an = 4an-1 +1 ,n=2,3,4,...},where k = 1,2,3,....
Conversely, by Theorem 2.3,for any sequence {an} = {an∣a1 = 4k-1,an = 4an-1 +1,n = 2,3,4,...} in S1,all terms of the sequence {an} have the unique back odd number 6k-1.So,ƒ1-1({an∣a1 = 4k-1,an = 4an-1 +1 ,n=2,3,4,...}) = 6k-1,where k = 1,2,3,.... To sum up,we obtain that ƒ1(D)= S1 is a bijection.■
Examples:Take k = 1,ƒ1(5) = {an}:3,13,53,213,.... Each term of {an} is a front odd number of 5 ∈ D.Take k = 2,ƒ1(11) = {an}:7,29,117,469,.... Each term of {an} is a front odd number of 11 ∈ D.etc.
Theorem 3.6:There exists a bijection ƒ2: ƒ2(E)= S2,such that for any 6k+1 ∈ E,ƒ2(6k+1) = {an} = {an∣a1 = 8k+1,an = 4an-1 +1 ,n=2,3,4,...}∈ S2 ,k = 1,2,3,..., where all front numbers of 6k+1 are exactly all terms of the sequence {an},and 3an + 1 = 22n(6k+1) (n = 1,2,3,...);conversely, all terms of the sequence {an} have the unique back number 6k+1.
Proof: Following the method proved in Theorem 3.5,the result holds.■
Examples: Take k = 0, ƒ2(1) = {an}:1,5,21,85,....Each term of {an} is a front odd number of 1 ∈ E.Take k = 1, ƒ2(7) = {an}:9,37,149,597,....Each term of {an} is a front odd number of 7 ∈ E.etc.
Corollary 3.7: There exists a bijection ƒ: ƒ(D ∪ E)= S1 ∪ S2 ,where if an odd number p ∈ D,then define ƒ(p) = ƒ1(p),and if p ∈ E,then define ƒ(p) = ƒ2(p).
Proof: Since D ∩ E = Ø,and S1 ∩ S2 = Ø,by Theorem 3.5 and Theorem 3.6, it is immediate.■

4. Successful Odd Sequence Set H

This paragraph gives such an odd sequence set H that an odd number p is successful,if and only if p is in a sequence of H.
Theorem 4.1: Let p be an odd number and q = 4p+1.Then
(1) if p ∈ E, then q ∈ D;
(2) if p ∈ D, then q ∈ F;
(3) if p ∈ F, then q ∈ E.
Proof: (1) Since p ∈ E,let p = 6k+1, where k is 0 or a definite positive integer. Then q = 4p+1 = 4(6k+1)+1 = 24k+5 = 6(4k+1)-1, so q ∈ D.
Since p ∈ D,let p = 6k-1,where k is a definite positive integer.Then q = 4p+1 = 4(6k-1)+1 = 24k-3 = 6(4k)-3, so q ∈ F.
(3) Since p ∈ F,let p = 6k-3, where k is a definite positive integer. Then q = 4p+1 = 4(6k-3)+1 = 24k-11= 6(4k-2)+1, so q ∈ E.■
Remarks: For a sequence {an} ∈ S1, a1 = 4k-1 ∈ A. And all odd numbers in A are {4k-1}:3, 7, 11, 15, 19, 23, ...; where 3, 15, ... ∈ F,and 7, 19, ... ∈ E, and 11, 23, . ...∈ D.
For a sequence {an}∈ S2, a1 = 8k+1 ∈ B. And all odd numbers in B are {8k+1}:1, 9, 17, 25, 33, 41,...; where 9, 33,...∈ F,and 1, 25,...∈ E, and 17, 41, ...∈ D.
That is, the first term a1 can be any one in three odd subsets D,E,F, whether the sequence {an} ∈ S1 or {an}∈ S2 .
Theorem 4.2: Let {an} ∈ S1 or {an}∈ S2, if a1 ∈ E, then a3k-2 ∈ E,a3k-1 ∈ D,a3k ∈ F,where k = 1,2,3,...;
(2) if a1 ∈ F, then a3k-2 ∈ F,a3k-1 ∈ E,a3k ∈ D,where k = 1,2,3,...;
(3) if a1 ∈ D, then a3k-2 ∈ D,a3k-1 ∈ F,a3k ∈ E,where k = 1,2,3,....
Proof: This is a direct corollary of Theorem 4.1.■
Remark: According to Theorem 4.2, any term of a sequence {an} in S1 or S2 can belong to any one in three odd subsets D,E,F.
Collatz sequences now under discussion are Collatz odd sequences, so the following definition is given.
Definition 4.1: If pk ,pk-1 ,...,p1 ,1 is a Collatz odd sequence with k odd numbers other than 1, then we say that pk is k steps successful; and it is specified that the odd number 1 is 0 step successful.
Example: 11, 17, 13, 5, 1 is a Collatz odd sequence given in §2. Then we have: odd 5 is 1 step successful, 13 is 2 steps successful, 17 is 3 steps successful, and 11 is 4 steps successful.Specially,1 is 0 step successful.
Corollary 4.3: If p is a front odd number of q ,p ≠ 1 and q is k steps successful, then p is k+1 steps successful. Specially,p = 1 is 0 step successful.
Proof: This is a direct consequence of Definition 4.1.■
Note that in the following,ƒ12 and ƒ are the three bijections given in Theorem 3.5, Theorem 3.6 and Corollary 3.7,where ƒ(D) = ƒ1(D) = S1,and ƒ(E)= ƒ2(E)= S2.For simplicity,if R is a sequence set,a sequence {an} ∈ R,and ai∈ {an} is a term of {an},then write ai ∈ R. If ai ∈ R and ai ∈ D, then write ai ∈ R ∩ D,etc.
In the following, by a recursive method, we construct such a sequence set H that an odd number p is successful, if and only if p is in a sequence of H.
By Theorem 3.6 and Corollary 3.7,since 1 ∈ E, ƒ(1) = ƒ2(1) = {an}:1,5,21,85,341,...∈ S2.
Write H1 ={ƒ(1)} = {ƒ2(1)}={{an}:1,5,21,85,341,...},where H1 is a sequence set,and there is exactly one sequence in H1.
In this sequence of H1 , since a1 = 1 ∈ E, by Theorem 4.2,a3k-2 ∈ E,a3k-1 ∈ D,a3k ∈ F,where k = 1,2,3,....By Lemma 3.4, a3k (∈ F) has no front odd number,where k = 1,2,3,.... By Theorem 3.5,for a definite a3k-1 (∈ D),there is the unique odd sequence {bn}∈ S1 such that ƒ1(a3k-1) = {bn},where k = 1,2,3,...,and each term of {bn} is a front odd number of a3k-1. By Theorem 3.6,for a definite a3k-2 (∈ E),there is the unique odd sequence {cn}∈ S2 such that ƒ2(a3k-2) = {cn},where k = 1,2,3,...,and each term of {cn} is a front odd number of a3k-2.
Write H2 = {ƒ1(p)| p ∈ H1 ∩ D} ∪ {ƒ2(p)| p ∈ H1 ∩ E and p ≠ 1 }.Note that (H1 ∩ D)∪ (H1 ∩ E) = H1 ∩ (D ∪ E).By Corollary 3.7,H2 = {ƒ(p)| p ∈ H1 ∩ (D ∪ E)and p ≠ 1}.(Note that the removal of ƒ(1) = ƒ2(1) in H2 is to prevent the sequence that appeared in H1 from reappearing in H2.)
Write H3 = {ƒ1(p)| p ∈ H2 ∩ D} ∪ {ƒ2(p)| p ∈ H2 ∩ E } = {ƒ(p)| p ∈ H2 ∩ (D ∪ E)}.
Assume that the sequence set Hn has been formed.
Write Hn+1 ={ƒ(p)| p ∈ Hn ∩ (D ∪ E)}.
Now the sequence sets H1 ,H2 ,H3 ,...,Hn ,... have been constructed.
Write H = n = 1 H n.
Remark: To get an intuitive sense of the composition of the set H, several sequences in H1,H2 and H3 are given here according to Theorem 3.5 and Theorem 3.6.
The unique sequence in H1 is ƒ(1) = ƒ2(1) = {an}:1,5,21,85,341,1365,....
Note that H2 = {ƒ1(p)| p ∈ H1 ∩ D} ∪ {ƒ2(p)| p ∈ H1 ∩ E and p ≠ 1 }.In the sequence {an} of H1, a3 = 21 ,a6 = 1365,... (∈ F) have no front odd number.Since p =5,341∈ H1 ∩ D,by Theorem 3.5,ƒ1(5) = {bn}:3,13,53,...∈ H21(341) = {bn}:227,909,3637,...∈ H2.And since p = 85 ∈ H1 ∩ E and p = 85 ≠ 1,by Theorem 3.6,ƒ2(85)= {bn}:113,453,1813,...∈ H2. And so on. And see H3.H3 = {ƒ1(p)| p ∈ H2 ∩ D} ∪ {ƒ2(p)| p ∈ H2 ∩ E }.Because ƒ1(5) = {bn}:3,13,53,...∈ H2, where 3 (∈ F) has no front odd number,13 ∈ H2 ∩ E,and 53 ∈ H2 ∩ D,so,by Theorem 3.6, ƒ2(13) = {cn}: 17,69,277,...∈ H3,and by Theorem 3.5,ƒ1(53) = {cn}: 35,141,565,...∈ H3.Because ƒ1(341) = {bn}:227,909,3637,...∈ H2, where 909 (∈ F) has no front odd number, 227 ∈ H2 ∩ D,and 3637 ∈ H2 ∩ E,so,by Theorem 3.5,ƒ1(227) = {cn}: 151,605,2401,...∈ H3, and by Theorem 3.6,ƒ2(3637) = {cn}: 4849,19397,77589,...∈ H3.Because ƒ2(85)= {bn}:113,453,1813,...∈ H2,where 453 (∈ F) has no front odd number,113 ∈ H2 ∩ D,and 1813 ∈ H2 ∩ E, so,by Theorem 3.5, ƒ1(113) = {cn}: 75,301,1205,...∈ H3,and by Theorem 3.6,ƒ2(1813) = {cn}: 2417,9669,38677,... ∈ H3.And so on.
Theorem 4.4: (1) There are no identical sequences in all sequences of H = n = 1 H n.And no two different sequences in them have the same term.
(2) All terms of any sequence in Hn are n steps successful, where n = 1, 2, 3, ...,except for the odd number 1 in H1.
(3) If an odd number p is successful, then p must be in some sequence of the sequence set H.
Proof: (1) Induction. Because H1 = {ƒ(1)} and H2 = {ƒ(p)| p ∈ H1 ∩ (D ∪ E)and p ≠ 1},so H1 ∪ H2 = {ƒ(p)| p ∈ H1 ∩ (D ∪ E)}.
Because H1 = {ƒ(1)} is a sequence set = {{an}:1,5,21,85,341,1365,...},and it has not the same term,so,there is no same odd number in H1 ∩ (D ∪ E). Since ƒ is a bijection,all sequences in H1 ∪ H2 are not identical to each other.Then since H1 ∪ H2 ⊂ S1 ∪ S2 ,by Theorem 3.1,no two different sequences in H1 ∪ H2 have the same term.
Consider H3 = {ƒ(p)| p ∈ H2 ∩ (D ∪ E)}.Then H1∪H2∪H3 ={ƒ(p)| p ∈ H1 ∩(D ∪ E)}∪{ƒ(p)| p ∈ H2 ∩ (D ∪ E)} = {ƒ(p)| p ∈ (H1∪H2 )∩(D∪E)}.Since no two different sequences in H1 ∪ H2 have the same term,there is not the same odd number in (H1∪H2 )∩(D∪E).Since ƒ is a bijection,all sequences in H1∪H2∪H3 are not identical to each other.By Theorem 3.1,no two different sequences in H1∪H2∪H3 have the same term.
Suppose all sequences in k = 1 n H k are different from each other, and no two different sequences have the same term.
Consider Hn+1 = {ƒ(p)| p ∈ Hn ∩ (D ∪ E)}.Then k = 1 n + 1 H k ={ƒ(p)| p ∈ H1 ∩(D ∪ E)}∪{ƒ(p)| p ∈ H2 ∩ (D ∪ E)} ∪...∪{ƒ(p)| p ∈ Hn ∩ (D ∪ E)} = {ƒ(p)| p ∈ ( k = 1 n H k)∩(D∪E)}.By the inductive assumption,no two different sequences in k = 1 n H k have the same term.So,there is no same odd number in ( k = 1 n H k)∩(D∪E).Since ƒ is a bijection,all sequences in k = 1 n + 1 H k are not identical to each other.By Theorem 3.1,no two different sequences in k = 1 n + 1 H k have the same term.
(2) Induction. First of all, all terms of the sequence in H1 are the front odd numbers of the odd number 1. By Definition 4.1,all terms of the sequence in H1 are 1 step successful,except for the odd number 1.
Consider any sequence {an} in H2.Then H2 = {ƒ(p)| p ∈ H1 ∩ (D ∪ E)and p ≠ 1}.So,there is a definite odd number q,such that q∈ H1 ∩ (D ∪ E),q ≠ 1, and ƒ(q) = {an}.Note that all terms of {an} are all front numbers of q,and q ∈ H1 ,q ≠ 1 is 1 step successful.By Corollary 4.3, all terms of {an} are 2 steps successful. Therefore,all terms of any sequence in H2 are 2 steps successful.
Assume that all terms of any sequence in Hn are n steps successful,n ≧ 2.
Consider any sequence {bn} in Hn+1.Then Hn+1 = {ƒ(p)| p ∈ Hn ∩ (D ∪ E)}.So,there is a definite odd number q,such that q∈ Hn ∩ (D ∪ E), and ƒ(q) = {bn}.By the inductive assumption, q ∈ Hn is n steps successful.Note that all terms of {bn} are all front numbers of q.By Corollary 4.3, all terms of {bn} are n+1 steps successful.Therefore,all terms of any sequence in Hn+1 are n+1 steps successful.
(3) Firstly,the odd number p = 1 is 0 step successful,1 ∈ H1.Now let p ≠ 1.Since the odd number p is successful, there must exist some positive integer n,such that p is n steps successful.It is enough to prove that p must be in some sequence of Hn.
Induction. Look at n = 1.Let p be 1 step successful.Then p ≠ 1.And let Collatz odd sequence of p be p,1.Then p is a front odd number of 1.Note that H1 ={ƒ(1)} = {{an}:1,5,21,85,341,...},and ƒ is a bijection,so,all front odd numbers of 1 belong to H1.So,p ∈ H1.
Look at n = 2.Let p be 2 steps successful. And let Collatz odd sequence of p be p ,p1 ,1.Then the odd number p1 is 1 step successful,and p1 ≠ 1.As above,p1 ∈ H1 and p1 ≠ 1.Because p1 has the front odd number p,so p1 ∈ D ∪ E,and p1 ∈ H1 ∩ (D ∪ E)and p1 ≠ 1. Note that H2 = {ƒ(q)| q ∈ H1 ∩ (D ∪ E) and q ≠ 1},then ƒ(p1) ∈ H2,where ƒ(p1) is a sequence of H2.And ƒ(p1) contain all front odd numbers of p1 ,and p is a front odd number of p1.So,p ∈ƒ(p1)∈ H2.
Assume that the result holds for n = k,that is, if the odd number p is k steps successful, then p ∈ Hk.
Now let n = k+1,that is, the odd number p is k+1 steps successful.And let Collatz odd sequence of the odd number p be p ,pk ,pk-1 ,...,p1 ,1. Then the odd number pk is k steps successful. By inductive assumption, pk ∈ Hk.Because pk has the front odd number p,so pk∈ D ∪ E,and pk ∈ Hk ∩ (D ∪ E). Note that Hk+1 = {ƒ(q)| q ∈ Hk ∩ (D ∪ E)},then ƒ(pk) ∈ Hk+1, where ƒ(pk) is a sequence of Hk+1.And ƒ(pk) contain all front odd numbers of pk ,and p is a front odd number of pk.So,p ∈ƒ(pk)∈ Hk+1.■

§5. Possibility of Not Successful Odd Sequence Set

Remark: From Theorem 4.4, it follows that every term of every sequence in H = n = 1 H n is successful, and every odd number that is successful must be in one of these sequences of H. Therefore, to prove that all odd numbers are successful, it is sufficient to show that every odd number is in one of these sequences of H.
Lemma 5.1: (1) If p is a front odd number of q,and the odd number q is not successful, then the odd number p is also not successful.
(2) If r is a back odd number of q,and the odd number q is not successful,then the odd number r is also not successful.
Proof: (1) First, since q is not successful, q ≠ 1. Assume that the odd number p is successful,then there exists a positive integer k,such that p is k steps successful. And by Theorem 2.3, q is the unique back odd number of p. Then, by Definition 4.1, k > 1,and q is k-1 steps successful. Contradiction.
Assume that the odd number r is successful,then there exists a positive integer m, such that r is m steps successful, so q is m+1 steps successful. Contradiction.■
By Theorem 3.1, any odd number is a term of only one sequence in S1 ∪ S2.
Lemma 5.2: If an odd number p is not successful,and p is a term of some sequence {an} in S1 ∪ S2,then all terms of {an} are not successful.
Proof: From Theorem 4.4, it follows that all odd numbers in the sequence set H are successful. Since the odd number p is not successful, p is not in H.And the odd number p is a term in some sequence {an}.Since H is the sequence set , that sequence {an} is also not in H. In the other words, all terms of {an} are not in H. By Theorem 4.4, all terms of {an} are not successful.■
Remarks:(1)In a Collatz odd sequence,if there is a situation like q1,q2,...,qk,q1,where q1,q2,...,qk are different from each other,and the odd number q1 recurring,then we call these k odd numbers to form a k-cycle.In fact,it is impossible to have such a k-cycle in a successful Collatz odd sequence.Because there must be a positive integer s for a successful odd number q1,so that q1 is s steps successful,so,there is a Collatz odd sequence q1,ps-1,ps-2,...,p1,1.
Assume q1 is in a k-cycle q1,q2,...,qk,q1.Then q1 has a Collatz odd sequence q1,q2,...,qr,pt,pt-1,...,p1,1,where q2 = ps-1,...,qr = pt+1,1 ≦ r ≦ k,r is a definite integer.Then qr has 2 back odd numbers qr+1 and pt,where if r = k,then qr+1 = qk+1 = q1,and if r = 1,then pt = ps-1.By Theorem 2.3,this is impossible.
(2)Assuming that p1 is not successful,and p1 is in a k-cycle p1,p2,...,pk,p1.Then by Theorem 2.3, Collatz odd sequence of these k odd numbers is the k-cycle,and any one of these k odd numbers can not appear in other j-cycle.And these k odd numbers are not successful.
(3)Let p be a front odd number of q.Then there exists a positive integer i,such that 3p+1 = 2iq.If i = 1,then q = (3p+1)/2 > p.If i ≧ 2,then q = (3p+1)/2i < p.Note that any odd number is a term of some sequence in S1 ∪ S2.By Theorem 3.5,for any sequence {an} = {an∣a1 = 4k-1,an = 4an-1 +1 ,n=2,3,4,...}∈ S1 ,3an + 1 = 22n-1(6k-1) (n = 1,2,3,...),where any term of {an} is a front odd number of 6k-1.And when n =1,that is,i = 2n-1 = 1,6k-1 > a1.When n ≧ 2,that is,i = 2n-1 ≧ 3,6k-1 < an.And by Theorem 3.6,for any sequence {an} = {an∣a1 = 8k+1,an = 4an-1 +1 ,n=2,3,4,...}∈ S2,3an + 1 = 22n(6k+1) (n = 1,2,3,...),where any term of {an} is a front odd number of 6k+1.When n = 1,2,3,...,that is,i = 2n ≧ 2,6k+1 < an.
Lemma 5.3:(1)There are no 2-cycles in Collatz odd sequences.
(2)There are no 3-cycles in Collatz odd sequences.
Proof: (1)Suppose that there is a 2-cycle q1,q2,q1 in Collatz odd sequences.Let q1 < q2.Because q1 is a front odd number of q2,so there is a positive integer i,such that 3q1+1 = 2iq2.By Remark(3) above,i = 1,that is,3q1+1 = 2q2.And because q2 is a front odd number of q1,so there is a positive integer j,such that 3q2+1 = 2jq1,i.e,(3q2+1)/2j = q1.By Remark(3) above,since q2 > q1,j ≧ 2.But (3q1+1)/2 = q2.When j = 2,(3q2+1)/2j =(3((3q1+1)/2)+1)/22 = (9q1+5)/8 > q1.When j ≧ 3,(3q2+1)/2j ≦ (3((3q1+1)/2)+1)/23 = (9q1+5)/16 < q1.That is to say, (3q2+1)/2j ≠ q1.So,there are no 2-cycles in Collatz odd sequences.
(2)And suppose that there is a 3-cycle q1,q2,q3,q1 in Collatz odd sequences.Let q1 < q2 and q1 < q3.The following are divided into two cases.
Case 1:q2 > q3.Since q1 < q2,and q1 is a front odd number of q2,by Remark(3) above,3q1+1 = 2q2,i.e,(3q1+1)/2 = q2.Since q2 > q3,and q2 is a front odd number of q3,3q2+1 = 2iq3,i.e,(3q2+1)/2i = q3,where i ≧ 2.Since q3 > q1,and q3 is a front odd number of q1,3q3+1 = 2jq1,i.e,(3q3+1)/2j = q1,where j ≧ 2.But (3q3+1)/2j ≦ (3q3+1)/22 = (3( (3q2+1)/2i)+1)/22 ≦(3( (3q2+1)/22)+1)/22 = (9q2+7)/16 = (9((3q1+1)/2)+7)/16 = (27q1+23)/32 < q1.(Note that because small odd numbers must be successful,and odd numbers in k-cycle are all not successful,so,we can set q1 > 99.) That is to say,(3q3+1)/2j ≠ q1.Contradiction.
Case 2:q2 < q3.Since q1 < q2,and q1 is a front odd number of q2,by Remark(3) above,3q1+1 = 2q2,i.e,(3q1+1)/2 = q2.Since q2 < q3,and q2 is a front odd number of q3,3q2+1 = 2q3,i.e,(3q2+1)/2 = q3.Since q3 > q1,and q3 is a front odd number of q1,3q3+1 = 2iq1,i.e,(3q3+1)/2i = q1,where i ≧ 2.But (3q3+1)/2i = (3((3q2+1)/2)+1)/2i = (9q2+5)/2i+1 =(9((3q1+1)/2)+5)/2i+1 = (27q1+19)/2i+2,
where i ≧ 2.And when i = 2,(3q3+1)/2i = (27q1+19)/2i+2 = (27q1+19)/16 > q1.when i ≧ 3,(3q3+1)/2i = (27q1+19)/2i+2 ≦ (27q1+19)/32 < q1,where set q1 > 99. That is to say,(3q3+1)/2i ≠ q1.
Contradiction.
So,there are no 3-cycles in Collatz odd sequences.■
Remark:The same method can be used to prove that there are no 4-cycles,5-cycles,etc.But there are too many cases to consider,and it will not go on.Later,in Theorem 5.5,we will prove that the k-cycle does not exist.
Remark:Suppose that an odd number q1 is not successful.Then a Collatz odd sequence q1,q2,...,qn,...can be obtained,where qi+1 is the unique back odd number of qi,i = 1,2,3,....By Lemma 5.1,each odd number in the Collatz odd sequence is not successful.In this point,there exist two cases for the Collatz odd sequence q1,q2,...,qn,...
Case 1:All odd numbers of the Collatz odd sequence q1,q2,...,qn,... are different from each other.
Case 2:q1,q2,...,qk,qk+1,qk+2 are different from each other,and qk+3 is an odd number in q1,q2,...,qk+2,such as,qk+3 = q3.Then q3,q4,...,qk+2,q3 is a k-cycle,k ≧ 4.(k ≧ 4 is because the non-existence of 2-cycle and 3-cycle has been proved in Lemma 5.3.)And the Collatz odd sequence q1,q2,...,qn,... is changed to q1,q2,q3,...,qk,qk+1,qk+2,q3,...,qk,qk+1,qk+2,q3,...,by Theorem 2.3,here this k-cycle keeps appearing repeatedly.For convenience,in the following, the Collatz odd sequence has been replaced with r1,r2,q1,q2,q3,...,qk,q1,q2,q3,...,qk,q1,....
Now,As with the construction of the sequence set H,from not successful Collatz odd sequence q1,q2,...,qn,..., the following series Gn of sequence sets is constructed.In the above Case 2,q1,q2,...,qn,...has been replaced with q1,q2,q3,...,qk,q1,q2,q3,...,qk,q1,....
The following construction is unified for Case 1 and Case 2.In the construction of Gn,for Case 2,when n = ik + m,let qn = qm,where i = 0,1,2,...,m = 1,2,...,k.
Firstly, we construct a sequence set G1.
Note that each odd number in the Collatz odd sequence q1,q2,...,qn,...is not successful, and q2 is the unique back odd number of q1 .
Write G11 = {ƒ(q2)} = {{an}},where {an} ∈ S1 ∪ S2,all terms of {an} are all front odd numbers of q2,and the odd number q1 is a term of {an}.
Write G12 = {ƒ(p)| p ∈ G11 ∩ (D ∪ E)}.
Write G13 = {ƒ(p)| p ∈ G12 ∩ (D ∪ E)}.
Assume that the sequence set G1n has been formed.
Write G1,n+1 = {ƒ(p)| p ∈ G1n ∩ (D ∪ E)}.
Denote G1 = n = 1 G 1n.
And construct a sequence set G2.
Note that q3 is the unique back odd number of q2.
Write G21 = {ƒ(q3)} = {{bn}},where the odd number q2 is a term of {bn}.
Write G22 = {ƒ(p)| p ∈ G21 ∩ (D ∪ E)}.
Write G23 = {ƒ(p)| p ∈ G22 ∩ (D ∪ E)}.
Assume that the sequence set G2n has been formed.
Write G2,n+1 = {ƒ(p)| p ∈ G2n ∩ (D ∪ E)}.
Denote G2 = n = 1 G 2n.
So on and so forth.
Suppose again that Gn = k = 1 G nk has been constructed.In the construction of Gn,for Case 2,when n = ik + m,let qn = qm,where i = 0,1,2,...,m = 1,2,...,k.
At this point, Gn1 has only one sequence, denoted {cn},and qn ∈ {cn}.qn+1 is the unique back odd number of qn.
Write Gn+1,1 ={ƒ(qn+2)}= {{dn}}.where the odd number qn+1 is a term of {dn}.
Write Gn+1,2 = {ƒ(p)| p ∈ Gn+1,1 ∩ (D ∪ E)}.
Write Gn+1,3 = {ƒ(p)| p ∈ Gn+1,2 ∩ (D ∪ E)}.
Assume that the sequence set Gn+1,k has been formed.
Write Gn+1,k+1 = {ƒ(p)| p ∈ Gn+1,k ∩ (D ∪ E)}.
Denote Gn+1 = k = 1 G n+1,k.
Now Gn has been constructed recursively, where n = 1,2,3,....
Remark:For r1 and r2 mentioned in Case 2,because r2 is a front odd number of q1,and r1 is a front odd number of r2,and q1 ∈ G11,so,r2 ∈ ƒ(q1) ∈ G12 and r1∈ ƒ(r2) ∈ G13.In other words, r1 and r2 are all in G1.
The following Theorem 5.4 is valid for both Case 1 and Case 2.
Theorem 5.4:(1) Any term of all sequences in the set Gn is not successful,where n = 1,2,3,....
G1 ⊂ G2 ⊂ ... ⊂ Gn ⊂ ....,where Gi is a proper subset of Gi+1,i = 1,2,3,....
Proof:(1)Induction for Gnk,where n = 1,2,3,.... Firstly,consider Gn1.Since qn is not successful, and qn ∈ {cn},(see Gn1.)by Lemma 5.2,any term of the unique sequence {cn} in Gn1 is not successful.The result holds for k = 1.
Consider Gn2 = {ƒ(p)| p ∈ Gn1 ∩ (D ∪ E)}.Let ƒ(q) = {bn}(∈ S1 ∪ S2) be any sequence in Gn2.Then q ∈ Gn1 ∩ (D ∪ E),and all term of {bn} are all front odd numbers of q.Since q ∈ Gn1,q is not successful.And by Lemma 5.1(1),all term of {bn} are not successful.So the result holds when k = 2.
Assume that the result holds for k = i, that is, any term of any sequence in Gni is not successful.
Consider Gn,i+1 = {ƒ(p)| p ∈ Gni ∩ (D ∪ E)}.Let ƒ(q) = {dn}(∈ S1 ∪ S2) be any sequence in Gn,i+1.Then q ∈ Gni ∩ (D ∪ E),and all term of {dn} are all front odd numbers of q.Since q ∈ Gni,by inductive assumption,q is not successful.And by Lemma 5.1(1),all term of {dn} are not successful.So the result holds when k = i+1.
(2) First prove that G1 ⊂ G2 . Look back at the construction of G2.Note that G11 = {ƒ(q2)}
= {{an}}, where {an} ∈ S1 ∪ S2,and the odd number q1 is a term of {an}.G21 = {ƒ(q3)} = {{bn}},where {bn} ∈ S1 ∪ S2,and the odd number q2 is a term of {bn}. G22 = {ƒ(p)| p ∈ G21 ∩ (D ∪ E)}.Since q2∈ {bn} ∈ G21,and ƒ(q2) = {an},(any term of {an} is a front odd numbers of q2.)q2 ∈ G21 ∩ (D ∪ E).Thus ƒ(q2) ∈ G22,i.e, G11 ⊂ G22.Because G12 = {ƒ(p)| p ∈ G11 ∩ (D ∪ E)},and G23 = {ƒ(p)| p ∈ G22 ∩ (D ∪ E)},and (G11 ∩ (D ∪ E)) ⊂ (G22 ∩ (D ∪ E)),so,G12 ⊂ G23.By analogy,it follows that G13 ⊂ G24 ,...,G1n ⊂ G2,n+1 , etc.
Since G1 = G11 ∪ G12 ∪ ... ∪ G1n ∪ ... G2 = G21 ∪ G22 ∪ ... ∪ G2n ∪ G2,n+1 ∪ ... So G1 ⊂ G2.Obviously,G1 is a proper subset of G2. By analogy, G2 ⊂ G3 ⊂ ... ⊂ Gn ⊂ ....■
Theorem 5.5:Case 2 cannot exist,that is,there are no k-cycles in not successful Collatz odd sequence.
Proof: Firstly,for Case 2,by Theorem 5.4,G1 ⊂ G2 ⊂ ... ⊂ Gk+1,where Gi is a proper subset of Gi+1,i = 1,2,...,k.So,G1 is a proper subset of Gk+1, On the other hand, consider the constructions of Gk+1 and G1.Note that Gk+1,1 = {ƒ(qk+2)}, where qk+2 = q2.So,Gk+1,1 = {ƒ(q2)}.But G11 = {ƒ(q2)},so,Gk+1,1 = G11.And because Gk+1,2 = {ƒ(p)| p ∈ Gk+1,1 ∩ (D ∪ E)} = {ƒ(p)| p ∈ G11 ∩ (D ∪ E)},and G12 = {ƒ(p)| p ∈ G11 ∩ (D ∪ E)}, so,Gk+1,2 = G12.By analogy,Gk+1,i = G1i,where i = 3,4,5,....Thus,G1 = n = 1 G 1n = n = 1 G k+1,n = Gk+1. Contradiction. Case 2 cannot exist.■
Next, we just need to discuss Case 1.Consider G1 first.
Theorem 5.6:If q1,q2,...,qn,...is the Collatz odd sequence in the construction of Gn,where all odd numbers of the Collatz odd sequence are different from each other,then
(1)q2 G1i,where i = 1,2,3,....
(2)There are no identical sequences in the sequence set i = 1 G 1i,and no two different sequences in i = 1 G 1i have the same term.
Proof:(1)Firstly,G11 = {ƒ(q2)} = {{an}},where all terms of {an} are all front odd numbers of q2.Because q2 is not successful,so,q2≠ 1, and by Theorem 2.4,q2 G11.Suppose that q2 is a term of some sequence {dn} in some G1i,where i ≧ 2 is a definite integer.Since G1i = {ƒ(p)| p ∈ G1,i-1 ∩ (D ∪ E)},there exists a definite odd number s ∈ G1,i-1 ∩ (D ∪ E),such that ƒ(s) = {dn},and all terms of {dn} are all front odd numbers of s.But q3 is the unique back odd number of q2 (∈{dn}),so,q3 = s ∈ G1,i-1.By analogy,q4 ∈ G1,i-2,...,qi ∈ G12,qi+1 ∈ G11.But G11 = {ƒ(q2)} = {{an}},where all terms of {an} are all front odd numbers of q2.So,q2 is a back odd number of qi+1(∈ {an}).Then qi+1 has two back odd number q2,qi+2(≠ q2),where i ≧ 2.By Theorem 2.3,this is impossible.
(2)Induction.First prove that the result holds for G11 ∪ G12.Note that G11 = {ƒ(q2)} = {{an}}= {ƒ(p)| p ∈ {q2} ∩ (D ∪ E)},and G12 = {ƒ(p)| p ∈ G11 ∩ (D ∪ E)}.So,G11 ∪ G12 = {ƒ(p)| p ∈ ({q2}∪ G11)∩ (D ∪ E)}.By (1),q2 G11.And ({q2}∪ G11)∩ (D ∪ E) has not the same odd number.Because ƒ is a bijection,so,there are no identical sequences in the sequence set G11 ∪ G12.Then since G11 ∪ G12 ⊂ S1 ∪ S2 ,by Theorem 3.1,no two different sequences in G11 ∪ G12 have the same term.
Consider G13 = {ƒ(p)| p ∈ G12 ∩ (D ∪ E)}.Then i = 1 3 G 1i = {ƒ(p)| p ∈ ({q2}∪ G11)∩ (D ∪ E)} ∪ {ƒ(p)| p ∈ G12 ∩ (D ∪ E)} = {ƒ(p)| p ∈({q2}∪ G11∪ G12 ) ∩ (D ∪ E)}.By (1),q2 G11 ∪ G12.And because there is no same odd number in G11∪ G12,so, there is no same odd number in ({q2}∪ G11∪ G12 ) ∩ (D ∪ E).Because ƒ is a bijection,so,there are no identical sequences in the sequence set i = 1 3 G 1i .Then by Theorem 3.1,no two different sequences in i = 1 3 G 1i have the same term.
Assume there are no identical sequences in the sequence set i = 1 n G 1i,and no two different sequences in i = 1 n G 1i have the same term,i,e.there is no same odd number in i = 1 n G 1i.
Note that i = 1 n + 1 G 1i = {ƒ(p)| p ∈({q2}∪( i = 1 n G 1i) ) ∩ (D ∪ E)}.By (1),q2 i = 1 n G 1i. By the inductive assumption, there is no same odd number in i = 1 n G 1i.So,there is no same odd number in ({q2}∪( i = 1 n G 1i) ) ∩ (D ∪ E).Because ƒ is a bijection,so,there are no identical sequences in the sequence set i = 1 n + 1 G 1i .Then by Theorem 3.1,no two different sequences in i = 1 n + 1 G 1i have the same term.■
The same method can be used to prove the following Theorem 5.7.
Theorem 5.7:If q1,q2,...,qn,...is the Collatz odd sequence in the construction of Gn,where all odd numbers of the Collatz odd sequence are different from each other,then for n = 2,3,4,...,
(1)qn+1 Gni,where i = 1,2,3,....
(2)There are no identical sequences in the sequence set i = 1 G ni,and no two different sequences in i = 1 G ni have the same term.■

§6. Final Results

In order to make a comparison between the sequence sets H and Gn,, another construction method of H is introduced here.
Let the odd number q1 be k steps successful. And let q1,q2,...,qk,1 (here set qk+1 = 1)be the Collatz odd sequence.
Next the following series Ti of sequence sets is constructed,where i = 1,2,...,k.
Firstly, we construct a sequence set T1.
Note that each odd number in the Collatz odd sequence q1,q2,...,qk,1 is successful,and q2 is the unique back odd number of q1 .
Write T11 = {ƒ(q2)} = {{an}},where {an} ∈ S1 ∪ S2,and the odd number q1 is a term of {an}.
Write T12 = {ƒ(p)| p ∈ T11 ∩ (D ∪ E)}.
Assume that the sequence set T1n has been formed.
Write T1,n+1 = {ƒ(p)| p ∈ T1n ∩ (D ∪ E)}.
Denote T1 = n = 1 T 1n.
Note that q3 is the unique back odd number of q2.
Write T21 = {ƒ(q3)} = {{bn}},where {bn} ∈ S1 ∪ S2,and the odd number q2 is a term of {bn}.
Write T22 = {ƒ(p)| p ∈ T21 ∩ (D ∪ E)}.
Assume that the sequence set T2n has been formed.
Write T2,n+1 = {ƒ(p)| p ∈ T2n ∩ (D ∪ E)}.
Denote T2 = n = 1 T 2n.
So on and so forth.
Suppose again that Ti = n = 1 T in has been constructed,where i = 3,4,...,k-2.
At this point, Ti1 has only one sequence, denoted {cn},and qi ∈ {cn}.qi+1 is the unique back odd number of qi.
Write Ti+1,1 = {ƒ(qi+2)}= {{dn}},where {dn} ∈ S1 ∪ S2,and the odd number qi+1 is a term of {dn}.
Write Ti+1,2 = {ƒ(p)| p ∈ Ti+1,1 ∩ (D ∪ E)}.
Assume that the sequence set Ti+1,j has been formed.
Write Ti+1,j+1 = {ƒ(p)| p ∈ Ti+1,j ∩ (D ∪ E)}.
Denote Ti+1 = j = 1 T i+1,j.
And consider i = k-1.Then qi+2 = qk+1 = 1,Ti+1,1 = Tk1 = {ƒ(qi+2)} = {ƒ(1)},where ƒ(1) is the odd sequence {en}:1,5,21,85,341,...,qi+1 = qk is a term of {en}.
Write Tk2 = {ƒ(p)| p ∈ Tk1 ∩ (D ∪ E),and p ≠ 1}.
Assume that the sequence set Tkj has been formed,where j ≧ 2.
Write Tk,j+1 = {ƒ(p)| p ∈ Tkj ∩ (D ∪ E)}.
Denote Tk = j = 1 T kj.
Remarks:(1)Note that Tk = H,that is,the sequence set H in §5 can also be constructed by the above method,and this process is the same as the generation process of Gn.
(2)See §5.Consider the constructions of G1 and G2.G11 = {ƒ(q2)},and G21 = {ƒ(q3)} = {{bn}},where the odd number q2 is just a general odd number in the sequence {bn} of G21,or in G21 ∩ (D ∪ E).G11 = {ƒ(q2)} grows in the front odd number direction (let's say) to obtain a G1. While there are infinitely many odd numbers in G21 ∩ (D ∪ E),and G2 is obtained by growing the infinitely many odd numbers of G21 ∩ (D ∪ E) in the front odd number direction.Where each odd number in G21 ∩ (D ∪ E) also generates a sequence set equivalent to G1.Thus,G2 is "infinitely many times" larger than G1 .
Similarly, for n = 3,4,...,Gn is "infinitely many times" larger than Gn-1.It can also be obtained for i = 1,2,...,k-1,Ti+1 is "infinitely many times" larger than Ti.
(3)By the constructions of Gn and Ti,the difference between sequence set Gn and the sequence set Ti is that all terms of all sequences in Ti are successful,so when it grows in the back odd direction (let's say ),it stops at odd number 1 to obtain Tk = H; while all terms of all sequences in Gn are not successful, so as n increases infinitely, Gn is endlessly expanding at a great speed. From Theorem 4.4, we know that the sequences in H are not identical to each other,and no two sequences have the same term (odd number).So H can be regarded as a set of odd numbers.Similarly,from Theorem 5.7, Gn can also be regarded as a set of odd numbers. From the above, it is obtained that as a set of odd numbers, H is a definite set of odd numbers, and as n increases infinitely, lim n G n is an indeterminate set of odd numbers.
Theorem 6.1: Any odd number is successful.
Proof: Let the set of all odd numbers be Q.Consider H as a set of odd numbers. By Theorem 4.4,the set of all successful odd numbers is H.Suppose there is an odd number that is not successful.And let the set of all not successful odd numbers be G. then G ∩ H = Ø,and G ∪ H = Q.
From Theorem 5.4 and Theorem 5.7,suppose there is an odd number that is not successful,then the sequence set Gn ,which is regarded as the set of odd numbers, can be obtained. And it is known that any odd number of Gn is not successful. And from the above remark (3), lim n G n is an indeterminate set of odd numbers. Because Gn ⊂ G,where n = 1,2,3,...,so lim n G n ⊂ G,and the set G of all odd numbers, those are not successful ,is also an indeterminate set of odd numbers.And the sets of odd numbers H and Q are both definite sets of odd numbers. So the equation G ∪ H = Q does not hold, contradiction.Thus,G = Ø, H = Q,and any odd number is successful.■
Remark:Treat each odd number in the set H as a vertex.Then connect an edge between any two odd numbers (vertices) in H that have a front and back odd number relationship.In particular,connect an edge between odd number 1 and any other odd number in the odd number set H1 except for odd number 1.At this point,H can be regarded as a tree with an odd root 1.We call it the H-tree.This H-tree contains all odd numbers.Because any triple odd number has no front odd number,each odd number in the triple odd number set F is a leaf of this H-tree.
Theorem 6.2: Any positive integer is successful,i.e.,Collatz conjecture holds.
Proof: See Remark at the top part of §2.■
Problem:For any odd number p,how to get a integer k with a formula,so that p is k-steps successful,or,p ∈ Hk in H.

Acknowledgment

Author is very grateful to the reviews for their comments on the original manuscript and some valuable suggestions,especially the suggestion to use the bijection to make the paper in its current form.

References

  1. Richrd K. Guy : Unsolved Problem In Number Theory.The 3x+1 Problem.Springer-Verlag, New York.330-336,2007.
  2. Jeffrey C.Lagarias: "The 3n+1 Problem and its generalization", Amer. Math. Monthly, 92:1 pp3-23.1985. [CrossRef]
  3. Jerffrey C.Lagarias,The 3x+1 problem:An Annotated Bibliography, II(2000-2009). arXiv:math/0608208 V6 [math NT]. [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