Abstract
In this paper, we investigate a generalization of the Bessenrodt–Ono inequality by following Gian–Carlo Rota’s advice in studying problems in combinatorics and number theory in terms of roots of polynomials. We consider the number of k-colored partitions of n as special values of polynomials \(P_n(x)\). We prove for all real numbers \(x >2 \) and \(a,b \in \mathbb {N}\) with \(a+b >2\) the inequality:
We show that \(P_n(x) < P_{n+1}(x)\) for \(x \ge 1\), which generalizes \(p(n) < p(n+1)\), where p(n) denotes the partition function. Finally, we observe for small values, the opposite can be true, since, for example: \(P_2(-3+ \sqrt{10}) = P_{3}(-3 + \sqrt{10})\).
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction and Main Results
Let p(n) be the number of partitions of n [2, 3, 17]. It is well known that this arithmetic function increases strictly:
for all \(n \in \mathbb {N}\), since every partition of n can be lifted to a partition of \(n+1\). Bessenrodt and Ono [5] discovered that (1.1) is actually an exception in the context of more general products \(p(a)\,p(b)\). Recently, DeSalvo and Pak [8] proved that the sequence \(\{p(n)\}\) is log-concave for \(n>25\).
Theorem 1.1
(Bessenrodt–Ono 2016) Let a, b be natural numbers. Let \(a+b >9\), then
Bessenrodt and Ono provided proof, based on a theorem of Rademacher [18] and Lehmer [15]. They speculated at the end of their paper that a combinatorial proof could be possible. Shortly after their paper was published, Alanazi, Gagola, and Munagi [1] found such a proof. Chern, Fu, and Tang [6] generalized Bessenrodt and Ono’s theorem to k-colored partitions \(p_{-k}(n)\) of n.
Theorem 1.2
(Chern, Fu, Tang 2018) Let a, b, k be natural numbers. Let \(k>1\), then
except for \((a,b,k) \in \left\{ (1,1,2), (1,2,2), (2,1,2), (1,3,2), (3,1,2), (1,1,3)\right\} \).
Let \(P_n(x)\) be the unique polynomial of degree n satisfying \(P_n(k)=p_{-k}(n)\) for all \(k \in \mathbb {N}\) (see [9]). In this paper we prove the following results:
Theorem 1.3
Let \(n \in \mathbb {N}\) and \(x \in \mathbb {R}\) with \(x \ge 1\). Then
Moreover, let \(n+1\) be an odd prime number. Then, there exists \(x_n \in (0,1)\), such that
Our main result is the following extension of the Bessenrodt–Ono type inequality.
Theorem 1.4
Let \(a,b \in \mathbb {N}\), \(a+b>2\), and \(x >2\). Then
The case \(x=2\) is true for \(a+b>4\).
This gives a precise answer to a conjecture stated in [11]. The theorem is proven by induction. The proof uses a special formula for the derivative of \(P_n(x)\), the inequality (1.3) for \(k=2\), a bound of Lehmer (5.2), and Theorem 1.3. The inequality (1.3) for \(k=2\) is proven by Chern, Fu, and Tang [6].
The values in Table 1 for \(1\le a,b\le 10\) imply that inequality (1.6) can also hold for values of x smaller than 2. If we assume \(a\ge b\) then except for \(b=1\) and \(a=1,2,3\) all values are less than 2. For values of x larger than the numbers (rounded to two decimal places) in position \(\left( a,b\right) \) inequality (1.6) holds.
Bessenrodt–Ono-type inequalities also appeared in works by Beckwith and Bessenrodt [4] on k-regular partitions and by Hou and Jagadeesan [12] on the numbers of partitions with ranks in a given residue class modulo \(t=3\). Recently, Males [16] obtained results for general t. Dawsey and Masri [7] obtained new results for Andrews spt-function. It is very likely that some of these recent results can be extended to an inequality of certain polynomials.
This paper is organized as follows. In Sect. 2, we introduce the polynomials \(P_n(x)\) and provide basic properties. In Sect. 3, we give numerical evidence for Theorem 1.3 and Theorem 1.4. In Sect. 4, we prove Theorem 1.3 and in Sect. 5, we prove Theorem 1.4.
2 Partitions and Polynomials
A partition \(\lambda \) of a positive integer n is any non-increasing sequence \(\lambda _1,\lambda _2, \ldots , \lambda _d\) of positive integers, whose sum is n. The \(\lambda _i\) denote the parts of the partition. The number of partitions of n is denoted by p(n) (see [2, 17]).
Example
The partitions of \(n=4\) and \(n=5\) are
Hence, \(p(4)=5\) and \(p(5)=7\). Note that p(200) is already equal to 3972999029388.
A partition is called a k-colored partition of n if each part can appear in k colors. Let \(p_{-k}(n)\) denote the number of k-colored partitions of n (see [6]). Note that \(p_{-1}(n)= p(n)\) and \(p_{-k}(n) < p_{-(k+1)}(n)\). The generating function of \(p_{-k}(n)\) is given by
Here, \((a;q)_{\infty }= \prod _{n=0}^{\infty } (1 - a \, q^n)\) is the q-Pochhammer symbol.
Definition
We define recursively a family of polynomials \(P_n(x)\). Let \(\sigma (n) := \sum _{d \mid n} d\) denote the sum of divisors of n. Then let \(P_0(x):=1\) and
It is known that \(P_n(k) = p_{-k}(n)\) for \(k \in \mathbb {N}\) (see [9] and compare the generating functions of \(P_{n}\left( k\right) \) and \(p_{-k}\left( n\right) \)). Let \(p_{-k}(0):=1\). If we put \(q:= \mathrm {e}^{2 \pi \mathrm {i} \tau }\) with \(\tau \) in the upper complex half-plane, then
We have \(P_0(x)=1\), \(P_1(x) = x\), \(P_2(x) = x/2 \, (x+3)\), \(P_3(x) = x/6 \, (x^2 + 9x +8)\). In addition, we have \(P_n(x) = x/n! \, \cdot \, {\widetilde{P}}_n(x)\), where \({\widetilde{P}}_n(x)\) is a monic polynomial of degree \(n-1\) with positive integer coefficients. There is a useful formula for the derivatives \(P_n'(x)\) for all \(n \in \mathbb {N}\) [10]:
3 Numerical Evidence for Theorems 1.3 and 1.4
3.1 The Difference Functions
In this section, we study the difference function
for \(n\in \mathbb {N}\). The results will be utilized to prove Theorem 1.3. We first observe that
This is true since \(P_n(x)\) is a polynomial of degree n of leading coefficient 1/n!. We also deduce that \(\Delta _n(0)=0\) for \(n\ge 1\). We are especially interested in the largest non-negative real root \(x_n\) of \(\Delta _n(x)\) (compare Fig. 1), since \(P_{n+1}(x) > P_n(x)\) for \(x > x_n\).
By (3.2) and continuity for all values of x larger than this root the difference function is positive. The real roots of \(\Delta _1(x)\) are \(\{-1,0\}\) and the real roots of \( \Delta _2(x)\) are
We can see already that \(\Delta _n(x)\) is not always positive, here \( \Delta _2 ( x) <0\) for
3.2 The Multiplicative Differences Functions
Let \(a,b \in \mathbb {N}\). The Bessenrodt–Ono inequality \(p(a) \, p(b) > p(a+b)\) is always satisfied for all \(a,b \ge 2\) and \(a+b >9\). Since the inequality is symmetric in a and b we assume \(a \ge b\). It was also shown [5] that there is equality for \((a,b) \in \left\{ (6,2), (7,2), (4,3)\right\} \). The inequality fails completely for \(b=1\) and
while it is true for the remaining cases \((a,b)\in \left\{ (4,4),(5,4),\left( 6,3\right) \right\} \).
The BO for 2-colored partitions is true for all \(a,b \in \mathbb {N}\) except for \((a,b)=(1,1)\), where \(p_{-2} (1) \, p_{-2}(1) < p_{-2}(2)\). Let \(a \ge b\). Then, we have equality for \((a,b) \in \{ (2,1),(3,1)\}\). The BO for 3-colored partitions holds for all \(a,b \in \mathbb {N}\) except for \((a,b)=(1,1)\), where we have equality. If \(k \ge 4\), then BO is fulfilled for all \(a,b \in \mathbb {N}\) (see [6]).
Let \(P_{a,b}(x):= P_a(x) \, P_b(x) - P_{a+b}(x)\). Then \(P_{a,b}(0)= 0\), \(P_{a,b}'(0) = - \sigma (a+b)/(a+b)\) and
In contrast to \(\Delta _n(x)\), the polynomials \(P_{a,b}(x)\) appear to have only one root \(x_{a,b}\) with a positive real part. Table 1 records these roots for \(1 \le a,b \le 10\). The root distribution in Table 1 explains all exceptions in the papers [5] and [6]. In Fig. 2 we have displayed the single positive root \(x_{a,1}\) of \(P_{a,1}(x)\) for \(1 \le a \le 100\).
Conjecture
It seems that, in general, \(P_{a,1}(x)\) has exactly one positive real root (and no non-real roots) and that the limit exists and is equal to 1.
4 Proof of Theorem 1.3
Proof of Theorem 1.3
We prove the first part of Theorem 1.3 by induction.
Claim: \(P_{n+1}(x) > P_n(x)\) for all \(n \in \mathbb {N}\) and \(x \ge 1\). Let \(n=1\). We have \(\Delta _1(x)> 0\) for all \(x>0\). Suppose that \(\Delta _m(x)>0\) is true for all real numbers \(x \ge 1\) and integers \(1 \le m \le n-1\). Using (2.3) and the induction hypothesis, we obtain for \(P_{n+1}'(x)\) the strict lower bound:
Hence, \(P_{n+1}'(x) > P_{n}'(x)\). Property (1.1) for the partition function provides
As a corollary, we obtain that \(\Delta _n'(x) >0\).
We finally prove (1.5). We have already observed \(\Delta _{2}\left( x\right) <0\) for \(0<x<\sqrt{10}-3\) in Sect. 3.1. Since \(\Delta _n(0)=0\) and
it is sufficient to show that \(\Delta _n'(0) <0\) for \(n+1\) an odd prime. We have
using that \(\sigma (n) > n+1\).
5 Proof of Theorem 1.4
5.1 Lehmer’s Bound
Let \(\mu (n):= \frac{\pi }{6} \sqrt{24n-1}\). Rademacher [18] proved the formula:
Here, \(A_k^{*}(n) = \frac{1}{\sqrt{k}} A_k(n)\) are real numbers, where \(A_k(n)\) is a complicated sum of 24kth roots of unity. Lehmer [13,14,15] obtained the estimate
for all \(n, N \in \mathbb {N}\). DeSalvo and Pak [8] recently utilized the case \(N=2\) and proved that p(n) is log-concave for all \(n>25\). They proved two of Chen’s conjectures. For our purpose the case \(N=1\), which was studied by Bessenrodt and Ono [5], is more convenient. They obtained:
This is the effective estimate for \(P_{m}\left( 1\right) \) that we use throughout our proof.
5.2 Proof of Theorem 1.4 When \(b = 1\)
In the following, we utilize the well-known upper bound:
Proposition 5.1
Let \(n \in \mathbb {N}\) and \(x \in \mathbb {R}\). The inequality
holds for all \(x>2\) and \(n\ge 2\). In the case \(n=1\) it holds for \(x>3\).
Proof
The case \(n=1\) is easy to see, since \(P_{1,1}(x) = x/2 \, (x-3)\). We prove the proposition by induction on n. Let \(n=2,3 \). Then \(P_{2,1}(x) = x/3 \, (x^2-4)>0\) for \(x >2\) and \(P_{3,1}\left( x\right) =\frac{x}{8}\,\left( x-2\right) \left( x+1\right) \left( x+7\right) >0\) for \(x>2\). Suppose that \(n\ge 4 \) and that \(P_{m,1}(x) >0\) for \(2 \le m \le n-1\) and \(x>2\). It is sufficient to prove that \(\frac{\mathrm {d}}{\mathrm {d}x} P_{n,1}(x) > 0\), since we already know that \(P_{n,1}(2) \ge 0\) ([6], see Theorem 1.2). The derivative of \(P_{n,1}(x)\) is equal to
We recall that \(P_n(x)\) has non-negative coefficients. It is, therefore, sufficient to show that
for \(n\ge 4 \). Since \(P_n(2)> P_n(1)\), we can now use (5.2). Suppose \(n\ge 49\). Then
For
we see that \( 17/8 \left( 1+\ln \left( 2n\right) \right) \le 17/8 \left( \ln \left( 98\right) +\frac{n}{49} \right) \le f\left( n\right) \). Thus for \(n\ge 58\) it holds that \(P_{n}\left( x\right)>P_{n}\left( 1\right) > 17/8 \left( \ln \left( 98\right) +\frac{n}{49} \right) \ge 17/8 \left( 1+\ln \left( 2n\right) \right) \). It remains to check \(P_{n}\left( 2\right) >\frac{17}{8}\left( 1+\ln \left( 2n\right) \right) \) for \(4 \le n\le 57\). In this case we have \(\left( 1+\ln \left( 2n\right) \right) <6\). Since \(P_{n}\left( 2\right) \) increases monotonously in n by Theorem 1.3 we also have \(\frac{17}{8}\left( 1+\ln \left( 2n\right) \right) < 20=P_{4 }\left( 2\right) \le P_{n}\left( 2\right) \) for \(4 \le n\le 57\). \(\square \)
We deduce from the proof of Proposition 5.1 the following property.
Corollary 5.2
Let \(n \ge 2\) and \(x>2\) or \( n \ge 4\) and \(x \ge 1\). Then
5.3 Proof of Theorem 1.4 When \(b > 1\)
Proof of Theorem 1.4
We show (1.6) by induction on \(n=a+b\). For \(n=a+b=3\) we have \(P_{1}\left( x\right) P_{2}\left( x\right) -P_{3}\left( x\right) =\frac{x}{3}\left( x^{2}-4\right) >0\) for all \(x>2\).
Suppose \(n\ge 4\) and \(P_{A}\left( x\right) P_{B}\left( x\right) >P_{A+B}\left( x\right) \) for all \(3\le A+B\le n-1=a+b-1\) and \(x>2\). Without loss of generality we assume \(a\ge b\ge 2\). (The case \(b=1\) was proved in Proposition 5.1.)
We have \(P_{a}\left( 2\right) P_{b}\left( 2\right) \ge P_{a+b}\left( 2\right) \) for \(a+b\ge 3\) by Theorem 1.2 of [6]. If we can now show that
the proof is completed, as this implies \(P_{a}\left( x\right) P_{b}\left( x\right) >P_{a+b}\left( x\right) \) for all \(x>2\).
Note that \(P_{A}\left( x\right) P_{0}\left( x\right) =P_{A}\left( x\right) \). Thus,
We now consider
for each k separately. From Theorem 1.3, we know that \(P_{a+b-k}\left( x\right) \) increases faster than \(P_{b-k}\left( x\right) \) for \(x\ge 1\). Hence, to show that (5.8) is positive, it is enough to show this for \(1\le x\le 2\).
Using (5.2) for \(k<b\)
and the last is positive if and only if
Now
On the other hand
If a fulfills
then a also fulfills (5.9). We show (5.10) for \(a\ge 94\) in Lemma 5.3 below. This implies that (5.8) is positive for \(x=1\). Theorem 1.3 then implies that (5.8) is positive for \(x\ge 1\).
We used PARI/GP to check that Eq. (5.8) with \(x=2\) is positive for \(1\le k<b\le a\le 93\). Again Theorem 1.3 implies (5.8) is positive for \(x\ge 2\).
What remains to be considered is (5.8) for \(k=b\). This follows from Corollary 5.2 setting \(n=a\).
Thus, for all \(a\ge b\ge k\ge 1\) the value of (5.8) is positive. Hence, also
for \(x>2\). This completes the induction step, as explained in the beginning of the proof. \(\square \)
In the following, we provide a proof of (5.10) for \(a\ge 94\). With more work and tighter bounds, it is possible to show (5.10) for \(a\ge 34\).
Lemma 5.3
For \(a\ge 94\), Eq. (5.10) holds, that is
Proof
Since \(a>0\)
Since \(a\ge 81\), we have
Thus, (5.10) is fulfilled if
and this is fulfilled for \(a\ge 94>\left( \left( \frac{1}{81}+\frac{\ln \left( 162\right) }{81}\right) \left( \frac{1}{81}+1\right) \frac{9}{4}\right) /\left( \frac{1}{6!}\left( \frac{\pi }{3}\right) ^{6}\right) \). \(\square \)
References
A. A. Alanazi, S. M. Gagola III, A. O. Munagi: Combinatorial proof of a partition inequality of Bessenrodt–Ono. Ann. Comb. 21 (2017), 331–337.
G. E. Andrews: The Theory of Partitions. Cambridge University Press, Cambridge (1998).
G. E. Andrews, K. Eriksson: Integer Partitions. Cambridge University Press, Cambridge (2004).
O. Beckwith, C. Bessenrodt: Multiplicative properties of the number of\(k\)-regular partitions. Ann. Comb. 20 No. 2 (2016), 231–250.
C. Bessenrodt, K. Ono: Maximal multiplicative properties of partitions. Ann. Comb. 20 No. 1 (2016), 59–64.
S. Chern, S. Fu, D. Tang: Some inequalities for\(k\)-colored partition functions. Ramanujan J. 46 (2018), 713–725.
M. Locus Dawsey, R. Masri: Effective bounds for the Andrews spt-function. Forum Math. 31(3), 743-767 (2019)
S. DeSalvo, I. Pak: Log-concavity of the partition function. Ramanujan J. 38 (2015), 61–73.
B. Heim, F. Luca, M. Neuhauser: On cyclotomic factors of polynomials related to modular forms. Ramanujan J. 48 (2019), 445–458.
B. Heim, M. Neuhauser: Polynomials related to powers of the Dedekind eta function. Integers 18 Paper No. A97 (2018).
B. Heim, M. Neuhauser: Variants of a partition inequality of Bessenrodt–Ono. Research in Number Theory (2019), 5: 32.
E. Hou, M. Jagadeesan: Dyson’s partition ranks and their multiplicative extension. Ramanujan J. 45 No. 3 (2018), 817–839.
D. H. Lehmer: On the Hardy–Ramanujan series for the partition function. J. Lond. Math. Soc. 12 (1937), 171–176.
D. H. Lehmer: On the series for the partition function. Trans. Amer. Math. Soc. 43 (1938), 271–295.
D. H. Lehmer: On the remainders and convergence of the series for the partition functions. Trans. Amer. Math. Soc. 46 (1939), 362–373.
J. Males: Asymptotic equidistribution and convexity for partition ranks. Ramanujan J. (2020). https://doi.org/10.1007/s11139-019-00202-8
K. Ono: The Web of Modularity: Arithmetic of the Coefficients of Modular Forms and q-series. Conference Board of Mathematical Sciences 102 (2003).
H. Rademacher: A convergent series for the partition function\(p(n)\). Proc. Natl. Acad. Sci. USA 23 (1937), 78–84.
Acknowledgements
Open access funding provided by Projekt DEAL. We thank the RWTH Aachen University and the Graduate School: Experimental and constructive algebra for their support and the two referees for carefully reading the manuscript and their excellent comments.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Heim, B., Neuhauser, M. & Tröger, R. Polynomization of the Bessenrodt–Ono Inequality. Ann. Comb. 24, 697–709 (2020). https://doi.org/10.1007/s00026-020-00509-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00026-020-00509-0