A New Classification of Positive Integers
2025
Read A New Classification of Positive Integers (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-08844.txt.
Introduction
Brahim Mittou defines new divisor functions and a new class of integers in a 2025 paper. Divisors of a number can be listed in increasing order. The new functions sum powers of divisors at even positions and at odd positions separately.
Properties of both functions are studied for real or complex exponents. A new type called k-index ratio numbers is also introduced. Three conjectures related to those numbers are stated but not proved. The work is number theory at a basic, definitional level. It extends the classic divisor sum by splitting it according to position in the divisor list.
Catalog record
| YEAR | 2025 |
|---|---|
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 14 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-08844
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A New Classification of Positive Integers Abstract Let d 1 = 1 < d 2 < d 3 < ⋯ < d τ ( n ) = n d_{1}=1<d_{2}<d_{3}<\cdots<d_{\tau(n)}=n denote the increasing sequence of the divisors of a positive integer n n . In this paper, for real or complex values of α \alpha , we define and study some properties of two new divisor functions σ e , α \sigma_{e,\alpha} and σ o , α \sigma_{o,\alpha} . The first computes the sum of the α \alpha -th powers of the divisors of n n with even indices, and the second computes the sum of the α \alpha -th powers of the divisors of n n with odd indices. We also introduce a new type of positive integers, namely, k k -index ratio numbers and state three conjectures related to them. Via New Divisor Functions Brahim Mittou Laboratory of Applied Mathematics University Kasdi Merbah Ouargla Algeria [email protected] 2020 Mathematics Subject Classification: Primary 11A25; Secondary 11A99. Keywords: divisor function, perfect square, classification of integers. 1 Introduction In number theory, arithmetic functions are essential tools for exploring and understanding the deep properties of integers. One of the most important among them is the divisor function, defined for real or complex values of \alpha by \sigma_{\alpha}(n)=\sum_{d\mid n}d^{\alpha}, where the sum is taken over all positive divisors d of n . When \alpha=0 , the function \sigma_{0}(n) counts the number of divisors of n ; this is often denoted by \tau(n) or d(n) , and when \alpha=1 , the function \sigma_{1}(n) gives the sum of the divisors of n ; this is often denoted by \sigma(n) , (see, e.g., [ 1 ] ). In addition to their appearance in many formulas in analytic number theory (see, e.g., [ 4 ] ), the divisors functions have played an important role, especially in the classification of integers. For example, prime numbers satisfy \tau(n)=2 , perfect numbers are characterized by the identity \sigma(n)=2n (see for instance the book of Hardy and Wright [ 2 ] ). Another interesting class is that of Zumkeller numbers which are positive integers whose divisors can be partitioned into two disjoint subsets with equal sums (see, e.g., [ 3 ] ). Additional related integer sequences and classifications can be found in [ 4 ] . An interesting variant of divisor functions arises when one partitions the divisors of n into two groups based on their ordering. In this work, we focus on two such functions. Throughout this paper, we let d_{1}=1<d_{2}<d_{3}<\cdots<d_{\tau(n)}=n denote the increasing sequence of the divisors of a positive integer n . For real or complex \alpha , we define \sigma_{e,\alpha} and \sigma_{o,\alpha} to be arithmetic functions such that: \sigma_{e,\alpha}(n)=\sum_{i=1,\,2\mid i}^{\tau(n)}d_{i}^{\alpha}\text{ and }\sigma_{o,\alpha}(n)=\sum_{i=1,\,2\nmid i}^{\tau(n)}d_{i}^{\alpha}. Easily one can show, for any prime p and odd positive integer l , that: \sigma_{e,\alpha}(p^{l})=p^{\alpha}\dfrac{p^{\alpha(l+1)}-1}{p^{2\alpha}-1}\text{ and }\sigma_{o,\alpha}(p^{l})=\dfrac{p^{\alpha(l+1)}-1}{p^{2\alpha}-1}, and if l is even, we have \sigma_{e,\alpha}(p^{l})=p^{\alpha}\dfrac{p^{\alpha l}-1}{p^{2\alpha}-1}\text{ and }\sigma_{o,\alpha}(p^{l})=\dfrac{p^{\alpha(l+2)}-1}{p^{2\alpha}-1}. Now, we let \sigma_{e,0}(n)=\tau_{e}(n) , \sigma_{o,0}(n)=\tau_{o}(n) , \sigma_{e,1}(n)=\sigma_{e}(n) , and \sigma_{o,1}(n)=\sigma_{o}(n) . Then \tau_{e}(n)=\frac{\tau(n)-1}{2}=\tau_{o}(n)-1 and \sigma_{e}(n)<\sigma_{o}(n) if n is a perfect square, and \tau_{e}(n)=\frac{\tau(n)}{2}=\tau_{o}(n) and \sigma_{e}(n)>\sigma_{o}(n) if n is not a perfect square. The following definition introduces a new classification of positive integers based on the functions \sigma_{e} and \sigma_{o} . Definition 1 . For any positive integer n , define k=k(n):=\dfrac{\sigma_{e}(n)}{\sigma_{o}(n)}, and we call n a k - index ratio number . For each positive rational number k , let G_{k} denote the set of all k -index ratio numbers. For example, see Table 1 . In particular, we call an index ratio number any positive integer n such that \sigma_{o}(n) divides \sigma_{e}(n) . The first few index ratio numbers are 1,2,3,5,6,7,8,10,11,13,14,15,17,18,19,21,22,23,26,27,29,31,32,\ldots k The first few k -index ratio numbers \leq 10^{5} 2 2,6,8,10,14,18,22,26,\ldots,99982,99986,99992,99998 2/5 4 3 3,15,21,27,33,39,51,\ldots,99951,99969,99987,99993 3/10 9 9/5 12,20,156,204,228,276,\ldots,99860,99948,99960,99996 10/21 16 47/25 30,646,930,1110,1230,\ldots,99570,99690,99870,99930 33/58 36 19/7 45,117,2115,2385,2655,\ldots,99405,99585,99801,99945 7109/15862 11025 5 5,35,55,65,85,95,115,\ldots,99955,99965,99985,99995 Table 1: The set G_{k} for some values of k Section 2 of this paper is divided into two subsections. In the first, we study the k -index ratio numbers n in the case where \tau(n) is even, and we also state two related conjectures. In the second subsection, we consider the complementary case, where \tau(n) is odd. 2 Main Results 2.1 \tau(n) is Even In this subsection, we suppose that \tau(n) is even; that is n is not a perfect square. Let us begin with the following results, which provide a method for generating new k -index ratio numbers from known ones. Theorem 2 . Let n be a k -index ratio number and p be a prime number such that p>n . Then np^{a} for any a\geq 1 is also a k -index ratio number. Proof. Let A(n)=\{d_{2},d_{4},\ldots,d_{\tau(n)}\} and B(n)=\{d_{1},d_{3},\ldots,d_{\tau(n)-1}\} . For an integer x and a set S of integers, let xS=\{xy:y\in S\} , and by \sum S we mean \sum_{x\in S}x . Then, A(np^{a})=A(n)\cup pA(n)\cup\cdots\cup p^{a}A(n), and B(np^{a})=B(n)\cup pB(n)\cup\cdots\cup p^{a}B(n), since p>n . Thus, \displaystyle k(np^{a}) \displaystyle=\dfrac{\sum A(n)+p\sum A(n)+\cdots+p^{a}\sum A(n)}{\sum B(n)+p\sum B(n)+\cdots+p^{a}\sum B(n)} \displaystyle=\dfrac{\sum A(n)(1+p+\cdots+p^{a})}{\sum B(n)(1+p+\cdots+p^{a})} \displaystyle=\dfrac{\sum A(n)}{\sum B(n)}=k(n), from which it follows that np^{a} is a k -index ratio number. ∎ Corollary 3 . Let n be a k -index ratio number, and let \prod_{i=1}^{r}p_{i}^{m_{i}} be the prime factorization of an integer m>1 such that n<p_{1} , and n\prod_{j=1}^{i-1}p_{j}^{m_{j}}<p_{i} for all 2\leq i\leq r . Then nm is also a k -index ratio number. Proof. The proof follows immediately by induction on r and Theorem 2 . ∎ As a consequence of the above results, we have the following corollary. Corollary 4 . Let k\geq 1 be a rational number. Then G_{k} is either empty or has infinitely many elements. It is well known that if d is a non-trivial positive divisor of n (i.e., d\neq 1 and d\neq n ), then d_{2}\leq d\leq\frac{n}{2} . Hence, \tau(n)-2+n\leq\sigma_{e}(n)\leq\frac{\tau(n)+2}{4}\cdot n, and \frac{4n}{(\tau(n)-2)n+4}\leq\frac{1}{\sigma_{e,-1}(n)}\leq\frac{n}{\tau(n)-1}. Combining these inequalities yields the following bounds: 4\cdot\frac{\tau(n)-2+n}{(\tau(n)-2)n+4}\leq k=\frac{1}{n}\cdot\frac{\sigma_{e}(n)}{\sigma_{e,-1}(n)}\leq\frac{n}{4}\cdot\frac{\tau(n)+2}{\tau(n)-1}. In particular: • If n is prime (i.e., \tau(n)=2 ), then k=n . • If \tau(n)=4 , then 2\leq k\leq\frac{n}{4} . • If \tau(n)=6 , then \frac{n+4}{n+1}\leq k\leq\frac{2n}{5} . The following theorem provides an improved upper bound, which we will later show to be optimal. Theorem 5 . If n is k -index ratio number, then k<d_{2}+\frac{1}{d_{2}} . Proof. Suppose first that n is a power of a prime number, i.e., n=p^{l} with l odd. In this case, we have k=p=d_{2} . Therefore, for the remainder of the proof, we assume that n is not a power of a prime number. \bullet If \tau(n)=4 , then n=pq where p<q are primes, and it follows that k=\frac{p+pq}{1+q}=p=d_{2} . \bullet If \tau(n)=6 , then k=\frac{d_{2}+d_{4}+d_{6}}{d_{1}+d_{3}+d_{5}} . We proceed as follows: \displaystyle k \displaystyle<d_{2}+\frac{1}{d_{2}} \displaystyle\Leftrightarrow d_{4}<d_{2}d_{3}+\frac{1}{d_{2}}(d_{1}+d_{3}+d_{5}) \displaystyle\Leftrightarrow\frac{n}{d_{2}^{2}d_{3}}\left(d_{2}^{2}-d_{3}\right)<d_{2}d_{3}+\frac{1}{d_{2}}(d_{1}+d_{3}). Clearly, the above inequality holds when d_{2}^{2}=d_{3} or d_{2}^{2}<d_{3} . If d_{2}^{2}>d_{3} , then either d_{2}d_{3}=d_{4} or d_{2}d_{3}=d_{5} . In this case, we have \frac{n}{d_{2}^{2}d_{3}}\left(d_{2}^{2}-d_{3}\right)<\frac{n}{d_{2}^{2}d_{3}}d_{2}\left(d_{2}-1\right)=\frac{n}{d_{2}d_{3}}\left(d_{2}-1\right)<d_{2}d_{3}. This concludes the proof for the case \tau(n)=6 . \bullet If \tau(n)\geq 8 , then k=\frac{d_{2}+d_{4}+\cdots+d_{\tau(n)}}{d_{1}+d_{3}+\cdots+d_{\tau(n)-1}} . The inequality k<d_{2}+\frac{1}{d_{2}} is equivalent to (d_{4}-d_{5})+\cdots+(d_{\tau(n)-4}-d_{\tau(n)-3})+d_{\tau(n)-2}-\dfrac{d_{\tau(n)-1}}{d_{2}}< (d_{2}-1)(d_{5}+\cdots+d_{\tau(n)-3})+d_{2}d_{3}+\frac{1}{d_{2}}(d_{1}+d_{3}+\cdots+d_{\tau(n)-3}). (1) Let LHS and RHS denote the left-hand side and right-hand side of ( 1 ), respectively. Then we have LHS<d_{\tau(n)-2}-\frac{d_{\tau(n)-1}}{d_{2}}=\frac{n}{d_{2}^{2}d_{3}}\left(d_{2}^{2}-d_{3}\right). It is obvious that ( 1 ) holds when d_{2}^{2}=d_{3} or d_{2}^{2}<d_{3} . Now, it follows by assuming d_{2}^{2}>d_{3} , that d_{2}d_{3}\mid