On a nonnegativity conjecture of Andrews

Yazan Alamoudi · 2025

Read On a nonnegativity conjecture of Andrews by Yazan Alamoudi (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-2508-10871.txt.

Introduction

Yazan Alamoudi, writing alone, settles a nonnegativity conjecture posed by Andrews about a family of polynomials, in a 2025 mathematics paper. The polynomials, linked to the Alladi-Schur family, are built from partition-like sums. The conjecture held that certain of their coefficients never go negative, a property that usually signals hidden combinatorial structure.

Alamoudi proves the conjecture outright. Beyond the settlement, the paper develops further relations among two families of polynomials tied to the Alladi-Schur ones, and draws out implications of the main result. Nonnegativity results of this kind often point to counting interpretations waiting to be uncovered. The paper adds the Andrews conjecture to the list of such statements now on firm ground.

Catalog record

CREATORYazan Alamoudi
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE16 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-10871

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

On a nonnegativity conjecture of Andrews Yazan Alamoudi Abstract . I settle a conjecture of Andrews related to the Alladi-Schur polynomials. In addition, I give further relations and implications to two families of polynomials related to the Alladi-Schur polynomials. 2000 Mathematics Subject Classification 11P81, 11P83, 05A17 1. Introduction The conjecture in question relates to polynomials arising in an investigation of integer partitions. That investigation [ 3 ] is what led to Andrews’ refinement 1 1 1 The Alladi-Schur theorem is the statement that the number of partitions of n into odd parts, each occurring at most twice, is equal to the number of Schur partitions of n . This was first observed by K. Alladi and communicated to G. Andrews (see [ 3 ] ). of the Alladi-Schur theorem, which is the equation |C(m,n)|=|\mathcal{D}(m,n)|, where C(m,n) is the set of partitions of n into m odd parts, each occurring at most twice, and \mathcal{D}(m,n) is the set of Schur partitions \pi of n where the number of parts plus the number of even parts of \pi is m . Recall that Schur partitions are partitions into parts that differ by at least 3 with no consecutive multiples of 3 . Of relevance is that Andrews proved (1.1) using some recursive relations (see [ 3 ] and [ 2 ] ) of the Alladi-Schur polynomials d_{N}(x)=\sum_{m,n\geq 0}D_{N}(m,n)x^{m}q^{n}, where D_{N}(m,n)=|\mathcal{D}_{N}(m,n)| and \mathcal{D}_{N}(m,n) is the set of partitions in \mathcal{D}(m,n) with parts \leq N . More relevant are those relations Andrews gives in [ 2 ] , such as p_{n}(x)\mid d_{6n-1}(x) and its explicit form \frac{d_{6n-1}(x)}{p_{n}(x)}=\sum_{i=0}^{n}c(n,i)x^{i}, where p_{n}(x)=\prod_{i=1}^{n}(1+xq^{2i-1}+x^{2}q^{4i-2}) . Indeed, Andrews’ conjecture can now be presented. Conjecture (Andrews’ conjecture stated in [ 2 ] ) . For all n and j , c(n,j) has nonnegative coefficients. I resolve the above conjecture by a very similar approach to the one Andrews used to prove his factorization theorem for d_{n}(x) . The aforementioned theorem [ 2 , Thm 1.2] states that p_{\lfloor\frac{n+4}{6}\rfloor}(x)\mid d_{n}(x) for n\not\equiv 3\pmod{6} , and p_{\lfloor\frac{n-2}{6}\rfloor}(x)\mid d_{n}(x) otherwise. It is easy to see that this result can be written as p_{\lceil\frac{n+3\chi_{o}(n)}{6}\rceil-\chi_{o}(n)}(x)\mid d_{n}(x), where \chi_{o} is the indicator function for odd integers. To this end, define 2 2 2 In [ 2 ] Andrews has instead written \Delta(i,n,x,q) , with i\in\{-1,0,1,2,3,4\} , in place of what I write as \mathscr{d}_{6n-i}(x) . the polynomials \mathscr{d}_{n} as \mathscr{d}_{n}(x)=\frac{d_{n}(x)}{p_{\lceil\frac{n+3\chi_{o}(n)}{6}\rceil-\chi_{o}(n)}(x)}. In due course, I establish the validity of Andrews’ conjecture by proving the following stronger claim. Theorem 1 . For n\geq 1 , \mathscr{d}_{n} is a polynomial in x and q with nonnegative integer coefficients. Hence, Andrews’ conjecture is true. After establishing Andrews’ conjecture, this paper concludes its technical matter with a section dedicated to extending some of the ideas contributing to the proof of Theorem 1 to an appropriate level of generality. This includes recursive relations for \mathscr{d}_{n}(x) , which can be thought of as the corresponding relations to Andrews’ recursive relations of the Alladi-Schur polynomials d_{n}(x) . In addition, explicit formulas for \mathscr{d}_{n}(x) in terms of x,q, and c(n,j) are given. 3 3 3 Such formulas were alluded to by Andrews in [ 2 ] , but they were not explicitly stated. Lastly, this final chapter also mentions some implications for c(n,j) that show, for example, why Theorem 1 is stronger than the statement of Andrews’ conjecture. 2. Proof of Theorem 1 We begin with the following useful lemma, which already gives one recursive relation for the polynomials \mathscr{d}_{n}(x) . Lemma 1 . For N\geq 3 , we have the following. \mathscr{d}_{2N}(x)=\mathscr{d}_{2N-3}(xq^{2}) Proof. Recall that in [ 2 ] , Andrews has shown that, for N\geq 3 , d_{2N}(x)=\lambda(x)d_{2N-3}(x) with \lambda(x)=1+xq+xq^{2} . It follows that, \mathscr{d}_{2N}(x)=\frac{d_{2N}(x)}{p_{\lceil\frac{2N}{6}\rceil}(x)}=\frac{\lambda(x)d_{2N-3}(xq^{2})}{p_{\lceil\frac{(2N-3)+3}{6}\rceil}(x)}=\frac{d_{2N-3}(xq^{2})}{p_{\lceil\frac{(2N-3)+3}{6}\rceil-1}(xq^{2})}=\mathscr{d}_{2N-3}(xq^{2}). This establishes the lemma. ∎ To settle Andrews’ conjecture, we need one more lemma. First, observe that, by the standard technique of removing the largest part (see [ 1 ] and also [ 3 , 2 ] ), it is easy to see that d_{N}(x)=d_{N-1}(x)+x^{1+\chi_{2}(N)}q^{N}d_{N-3-\chi_{3}(N)}(x), where \chi_{n} denotes the indicator function for the set of multiples of n . The next lemma gives the analogous property for only the odd-indexed \mathscr{d}_{n}(x) , which is the case that will be used in the proof of Theorem 1, with the even-indexed case being presented in the next section. Lemma 2 . For 2N-1\geq 5 with 2N-1\not\equiv 0\pmod{3} we have \mathscr{d}_{2N-1}(x)=\mathscr{d}_{2N-2}(x)+xq^{2N-1}\mathscr{d}_{2N-4}(x). Otherwise, for N\geq 0 , with initial 4 4 4 This can be arrived at from Andrews’ convention that d_{-1}(x)=1 (see [ 3 , 2 ] ) and the convention that an empty product =1 . condition \mathscr{d}_{-1}(x)=1 , we have \mathscr{d}_{6N+3}(x)=\lambda_{N+1}(x)\mathscr{d}_{6N+2}(x)+xq^{6N+3}\mathscr{d}_{6N-1}(x), where \lambda_{N}(x)=(1+xq^{2N-1}+x^{2}q^{4N-2}) . Proof. For odd indices not divisible by 3 , either 5 5 5 For ease, one may use \lceil\frac{n+3\chi_{o}(n)}{6}\rceil-\chi_{o}(n)=\lceil\frac{n-3\chi_{o}(n)}{6}\rceil throughout, but I opted to avoid any confusion from \lceil\frac{-2}{6}\rceil=0 . \mathscr{d}_{6N+1}(x)=\frac{d_{6N+1}(x)}{p_{N}(x)}=\frac{d_{6N}(x)+xq^{6N+1}d_{6N-2}(x)}{p_{N}(x)}=\mathscr{d}_{6N}(x)+xq^{6N+1}\mathscr{d}_{6N-2}(x), or \mathscr{d}_{6N-1}(x)=\frac{d_{6N-1}(x)}{p_{N}(x)}=\frac{d_{6N-2}(x)+xq^{6N-1}d_{6N-4}(x)}{p_{N}(x)}=\mathscr{d}_{6N-2}(x)+xq^{6N-1}\mathscr{d}_{6N-4}(x). Otherwise, if the index is divisible by three, then \mathscr{d}_{6N+3}(x)=\frac{d_{6N+3}(x)}{p_{N}(x)}=\frac{d_{6N+2}(x)+xq^{6N+3}d_{6N-1}(x)}{p_{N}(x)}=\lambda_{N+1}(x)\mathscr{d}_{6N+2}(x)+xq^{6N+3}\mathscr{d}_{6N-1}(x). ∎ We are now ready to prove Theorem 1 and settle Andrews’ conjecture. The proof below closely follows Andrews’ proof of his factorization theorem in [ 2 ] . Proof of Theorem 1. We proceed by strong induction. For the base case, observe the following. \mathscr{d}_{1}(x)=1+xq \mathscr{d}_{2}(x)=1 \mathscr{d}_{3}(x)=1+x(q+q^{3})+x^{2}q^{2} \mathscr{d}_{4}(x)=1+xq^{3} \mathscr{d}_{5}(x)=1+x(q^{3}+q^{5}) \mathscr{d}_{6}(x)=1+x(q^{3}+q^{5})+x^{2}q^{6} Suppose this is true for every 6\leq n<N^{\prime} . For the inductive step, we first handle the case of even indices. If N^{\prime}=2N\geq 6 , then by Lemma 1 \mathscr{d}_{2N}(x)=\mathscr{d}_{2N-3}(xq^{2}), with \mathscr{d}_{2N-3}(xq^{2}) a polynomial in x and q with nonnegative integer coefficients by the induction hypothesis. It follows that \mathscr{d}_{2N}(x) is a polynomial in x and q with nonnegative integer coefficients. For the case of odd indices, we handle two instances: one where the index is not a multiple of three and another where it is. If N^{\prime}=2N-1\not\equiv 0\pmod{3} , then by (2.3) \mathscr{d}_{2N-1}(x)=\mathscr{d}_{2N-2}(x)+xq^{2N-1}\mathscr{d}_{2N-4}(x). We see that the right-hand side consists of polynomials in x and q with nonnegative integer coefficients. It follows that the same is true for the left-hand side. Lastly, for odd indices that are divisible by 3 , write N^{\prime}=6N+3 , we have \mathscr{d}_{6N+3}(x)=\lambda_{N+1}(x)\mathscr{d}_{6N+2}(x)+xq^{6N+3}\mathscr{d}_{6N-1}(x). Again, the right-hand side consists of polynomials in x and q with non-negative integer coefficients, and so the same must be true for the left-hand side. Now, Andrews’ conjecture immediately follows as what was shown here, along with the initial conditions for c(n,i) (e.g., c(0,0)=1 ) presented in [ 2 ] , implies that, in particular, \mathscr{d}_{6n-1}(x)=\sum_{i=0}^{n}c(n,i)x^{i} is a polynomial in x and q with nonnegative integer coefficients and c(n,i) is a polynomial in q with nonnegative integer coefficients for any i,n\in\mathbb{N} . ∎ To better grasp the significance of the result, the reader is invited to examine Andrews’ explicit formula for c(n,j) presented in [ 2 , Thm 1.3] . From that perspective, the number of years the conjecture remained open can be better appreciated. 3. Further relations and remarks related to the quotients \mathscr{d}_{n}(x) . We begin by completing the analogy of equation (2.2) for the even cases. Lemma 3 . For N\geq 1 , if N\not\equiv 0\pmod{3} then \lambda_{\lceil\frac{2N}{6}\rceil}(x)\mathscr{d}_{2N}(x)=\mathscr{d}_{2N-1}(x)+x^{2}q^{2N}\mathscr{d}_{2N-3}(x). Otherwise, \mathscr{d}_{6N}(x)=\mathscr{d}_{6N-1}(x)+x^{2}q^{6N}\mathscr{d}_{6N-4}(x). Proof. For the case where the index is not divisible by three, either \lambda_{N}(x)\mathscr{d}_{6N-2}(x)=\frac{\lambda_{N}(x)d_{6N-2}(x)}{p_{N}(x)}=\frac{d_{6N-2}(x)}{p_{N-1}(x)} =\frac{d_{6N-3}(x)+x^{2}q^{6N-2}d_{6N-5}(x)}{p_{N-1}(x)}=\mathscr{d}_{6N-3}(x)+x^{2}q^{6N-2}\mathscr{d}_{6N-5}(x), or \lambda_{N}(x)\mathscr{d}_{6N-4}(x)=\frac{\lambda_{N}(x)d_{6N-4}(x)}{p_{N}(x)}=\frac{d_{6N-4}(x)}{p_{N-1}(x)} =\frac{d_{6N-5}(x)+x^{2}q^{6N-4}d_{6N-7}(x)}{p_{N-1}(x)}=\mathscr{d}_{6N-5}(x)+x^{2}q^{6N-4}\mathscr{d}_{6N-7}(x). This proves equation (3.1). Otherwise, \mathscr{d}_{6N}(x)=\frac{d_{6N}(x)}{p_{N}(x)}=\frac{d_{6N-1}(x)+x^{2}q^{6N}d_{6N-4}(x)}{p_{N}(x)}=\mathscr{d}_{6N-1}(x)+x^{2}q^{6N}\mathscr{d}_{6N-4}(x), which proves equation (3.2). ∎ Theorem 2 . For N\geq 3 , we have the following.

Read the complete text