betaFeedback

WorksMathematics

Partition statistics and universal mixed mock modular forms

Abstract

In this note, we construct combinatorial statistics in terms of partitions, overpartitions, and two-color partitions and relate the corresponding generating functions to the universal mock theta functions. Our framework primarily builds on two qq-series identities of Fine.

Introduction and Statements of Results

In his last letter to Hardy, Ramanujan [3 pp. 354–355] introduced seventeen qq-series, which he called mock theta functions. According to Ramanujan, a mock theta function is a function FF whose asymptotic behavior near roots of unity resembles that of weakly holomorphic modular forms, however, there does not exist a single weakly holomorphic modular form whose difference from FF remains bounded as qq approaches any root of unity radially. In his seminal thesis, Zwegers [11] explained the modularity of Ramanujan’s mock theta functions by “completing” them to so-called harmonic Maass forms.1

Next recall that Dyson [5] introduced the rank2^2 statistic to give a combinatorial explanation of partition congruences of Ramanujan (see [10]). Note that the generating function of the number of partitions with even rank minus the number of partitions with odd rank is Ramanujan’s third order mock theta function

f(q):=∑n≥0qn2(−q)n2.f(q) := \sum_{n \ge0} \frac{q^{n^2}}{(-q)_n^2}.

2020 Mathematics Subject Classification. 05A17, 11P81, 11P84.

Key words and phrases. mock theta functions, overpartitions, partitions, two-color partitions, universal mock theta functions.

Here, and throughout, ∣q∣<1|q|<1 and for a∈Ca\in\mathbb{C} and n∈N0∪{∞}n\in\mathbb{N}_0\cup\{\infty\}, we set (a)n=(a;q)n:=∏j=0n−1(1−aqj)(a)_n=(a;q)_n:=\prod_{j=0}^{n-1}(1-aq^j).

Moreover, MacMahon [9] studied partitions without sequences which are related to mixed mock modular forms.3 More precisely, using work of Fine [6], Andrews [2] showed that the generating function of partitions without sequences is related to Ramanujan’s third order mock theta function χ(q)\chi(q). Note that Ramanujan’s mock theta functions are specializations of the universal mock theta functions (see [8 (3.2) and above (5.1)])

g3(ζ;q):=∑n≥0qn(n+1)(ζ,qζ;q)n+1,g2(ζ;q):=∑n≥0(−q)n(ζ,qζ;q)n+1qn(n+1)2.(1)g_3(\zeta;q):=\sum_{n\geq0}\frac{q^{n(n+1)}}{\left(\zeta,\frac{q}{\zeta};q\right)_{n+1}},\qquad g_2(\zeta;q):=\sum_{n\geq0}\frac{(-q)_n}{\left(\zeta,\frac{q}{\zeta};q\right)_{n+1}}q^{\frac{n(n+1)}{2}}. \tag*{(1)}

Here ζ∈C∖{0}\zeta\in\mathbb{C}\setminus\{0\} and for r∈Nr\in\mathbb{N} and a1,…,ar∈Ca_1,\ldots,a_r\in\mathbb{C}, we let (a1,…,ar)n:=(a1,…,ar;q)n:=(a1)n⋯(ar)n(a_1,\ldots,a_r)_n:=(a_1,\ldots,a_r;q)_n:=(a_1)_n\cdots(a_r)_n.

In [4 Theorem 1.3], Lovejoy, Mahlburg, and the second author gave a combinatorial interpretation of (for d≥3d\geq3 and d>2rd>2r)

(−qr,−qd−r;qd)∞(q2d;q2d)∞g2(−qr;qd).\frac{(-q^r,-q^{d-r};q^d)_\infty}{(q^{2d};q^{2d})_\infty}g_2(-q^r;q^d).

Moreover, [4 (1.2)] yields a combinatorial statistic that is related to (for d≥3d\geq3 and d>2rd>2r)

(−qr,−qd−r;qd)∞g3(−qr;qd).(-q^r,-q^{d-r};q^d)_\infty g_3(-q^r;q^d).

Note that d=2rd=2r was excluded in [4]. In this case, the residue classes rr and −r-r (mod dd) coincide. Thus (−qr,−qd−r;qd)∞=(−qr;q2r)∞(-q^r,-q^{d-r};q^d)_\infty=(-q^r;q^{2r})_\infty and each part ≡r\equiv r (mod 2r2r) comes in two copies. This naturally leads to overpartitions and two-color partitions. This is a key motivation for our combinatorial framework.

In this note, we consider restricted partitions, overpartitions, and colored partitions and relate the corresponding generating functions to the universal mock theta functions. We begin with different qq-hypergeometric series representations of g3g_3 and g2g_2. To the authors’ knowledge these representations are not known.

Theorem 1.1. Let ζ≠qm\zeta\neq q^m for any m∈Zm\in\mathbb{Z}.

(1) We have

g3(ζ;q)=(q)∞(ζ,qζ)∞∑n≥0(ζ,qζ)n(q)nqn.(2)g_{3}(\zeta;q)=\frac{(q)_{\infty}}{\left(\zeta,\frac{q}{\zeta}\right)_{\infty}}\sum_{n\geq0}\frac{\left(\zeta,\frac{q}{\zeta}\right)_{n}}{(q)_{n}}q^{n}. \tag*{(2)}

(2) We have

g2(ζ;q)=(q2;q2)∞(ζ,qζ)∞∑n≥0(ζ,qζ)n(q2;q2)nqn.(3)g_{2}(\zeta;q)=\frac{(q^{2};q^{2})_{\infty}}{\left(\zeta,\frac{q}{\zeta}\right)_{\infty}}\sum_{n\geq0}\frac{\left(\zeta,\frac{q}{\zeta}\right)_{n}}{(q^{2};q^{2})_{n}}q^{n}. \tag*{(3)}

With a suitable specialization of the variables ζ\zeta and qq and after multiplying by appropriate infinite products, the identities in Theorem 1.1 yield the generating functions for classes of restricted partitions. The summation index records either the largest part in a certain residue class or the number of parts in that class, while the remaining factors impose restrictions on the parts in other residue classes.

We now describe these partition statistics explicitly. Here and throughout, we assume d,r∈Nd,r\in\mathbb{N} with d≥2rd\geq2r. First, recall that an overpartition of nn is a partition of nn in which the first occurrence of a part may be overlined. Let p‾(n)\overline{p}(n) denote the number of overpartitions of nn and set p‾(0):=1\overline{p}(0):=1. For instance, p‾(2)=4\overline{p}(2)=4 enumerated by 22, 2‾\overline{2}, 1+11+1, 1‾+1\overline{1}+1. We now consider overpartitions of nn with the following restrictions:

  1. parts are divisible by rr,

  2. parts that are divisible by 2r2r are non-overlined,

  3. the largest part is divisible by 2r2r,

  4. each part ≡r(mod2r)\equiv r\pmod{2r} can appear at most twice and if it appears twice, then one of the two occurrences must be overlined.

Let op⁡r[1](n)\operatorname{op}_{r}^{[1]}(n) denote the number of such overpartitions of nn and set op⁡r[1](0):=1\operatorname{op}_{r}^{[1]}(0):=1. For example, op⁡2[1](12)=5\operatorname{op}_{2}^{[1]}(12)=5 enumerated by 1212, 8+48+4, 8+2‾+28+\overline{2}+2, 4+4+44+4+4, 4+4+2‾+24+4+\overline{2}+2. Let OP⁡r[1](q):=∑n≥0op⁡r[1](n)qn\operatorname{OP}_{r}^{[1]}(q):=\sum_{n\geq0}\operatorname{op}_{r}^{[1]}(n)q^{n} denote the corresponding generating function.

Theorem 1.2. We have

OP⁡r[1](q)=(−qr;q2r)∞2(q2r;q2r)∞g3(−qr;q2r).\operatorname{OP}_{r}^{[1]}(q)=\frac{(-q^{r};q^{2r})_{\infty}^{2}}{(q^{2r};q^{2r})_{\infty}}g_{3}(-q^{r};q^{2r}).

Next, for d>2rd>2r, we consider partitions of nn with the following restrictions:

  1. parts are ≡0,±r(modd)\equiv0,\pm r\pmod d,

  2. each part ≡±r(modd)\equiv\pm r\pmod d is greater than every part divisible by dd.

Let pd,r[2](n)p^{[2]}_{d,r}(n) denote the number of such partitions of nn and set pd,r[2](0):=1p^{[2]}_{d,r}(0):=1.

For example, p4,1[2](5)=3p^{[2]}_{4,1}(5)=3 enumerated by 55, 3+1+13+1+1, 1+1+1+1+11+1+1+1+1. Let Pd,r[2](q):=∑n≥0pd,r[2](n)qnP^{[2]}_{d,r}(q):=\sum_{n\geq0}p^{[2]}_{d,r}(n)q^n denote the corresponding generating function.

Theorem 1.3. We have, for d>2rd>2r,

Pd,r[2](q)=1(qd;qd)∞Q3(qr;qd).P^{[2]}_{d,r}(q)=\frac{1}{(q^d;q^d)_\infty}Q_3\left(q^r;q^d\right).

Next, recall that a two-color partition of nn is a partition of nn in which each part appears in one of two fixed colors, say blue or green. Parts of the same size but with different colors are regarded as distinct and if a part occurs both in blue and in green, then we list the green parts first. Let p2(n)p_2(n) denote the number of two-color partitions of nn and set p2(0):=1p_2(0):=1. For example, p2(2)=5p_2(2)=5 enumerated by 2g2_g, 2b2_b, 1g+1g1_g+1_g, 1g+1b1_g+1_b, 1b+1b1_b+1_b. We consider two-color partitions of nn with the following restrictions:

  1. parts are divisible by rr,

  2. parts divisible by 2r2r can appear only in blue,

  3. each part ≡r(mod2r)\equiv r\pmod{2r} is greater than parts divisible by 2r2r.

Let pr[2](n)p^{[2]}_r(n) denote the number of such two-color partitions of nn and set pr[2](0):=1p^{[2]}_r(0):=1. For example, p1[2](8)=11p^{[2]}_1(8)=11 enumerated by 8b8_b, 6g+2g6_g+2_g, 6b+2g6_b+2_g, 6g+2b6_g+2_b, 6b+2b6_b+2_b, 4b+4b4_b+4_b, 2g+2g+2g+2g2_g+2_g+2_g+2_g, 2g+2g+2g+2b2_g+2_g+2_g+2_b, 2g+2g+2b+2b2_g+2_g+2_b+2_b, 2g+2b+2b+2b2_g+2_b+2_b+2_b, 2b+2b+2b+2b2_b+2_b+2_b+2_b. Let Pr[2](q):=∑n≥0pr[2](n)qnP^{[2]}_r(q):=\sum_{n\geq0}p^{[2]}_r(n)q^n denote the corresponding generating function.

Theorem 1.4. We have

Pr[2](q)=1(q2r;q2r)∞Q3(qr;q2r).P^{[2]}_r(q)=\frac{1}{(q^{2r};q^{2r})_\infty}Q_3\left(q^r;q^{2r}\right).

Now, we consider overpartitions of nn with the following restrictions:

  1. parts are divisible by rr but not by 4r4r,

  2. parts ≡2r(mod4r)\equiv2r\pmod{4r} are non-overlined,

  3. each part ≡r(mod2r)\equiv r\pmod{2r} can appear at most twice, and if it appears twice, it appears once overlined and once non-overlined,

  4. if there are exactly mm parts ≡2r(mod4r)\equiv2r\pmod{4r} (counted with multiplicity), then every part ≡r(mod2r)\equiv r\pmod{2r} is less than 2rm2rm. In particular, if there is no part ≡2r(mod4r)\equiv2r\pmod{4r}, then there is no part ≡r(mod2r)\equiv r\pmod{2r}.

Let op⁡r[2](n)\operatorname{op}^{[2]}_r(n) denote the number of such overpartitions of nn and set op⁡r[2](0):=1\operatorname{op}^{[2]}_r(0):=1. For example, op⁡1[2](6)=3\operatorname{op}^{[2]}_1(6)=3 enumerated by 66, 2+2+22+2+2, 2+2+1‾+12+2+\overline{1}+1. Let OP⁡r[2](q):=∑n≥0op⁡r[2](n)qn\operatorname{OP}^{[2]}_r(q):=\sum_{n\geq0}\operatorname{op}^{[2]}_r(n)q^n denote the corresponding generating function.

Theorem 1.5. We have

OP⁡r[2](q)=(−qr;q2r)∞2(q4r;q4r)∞g2(−qr;q2r).\operatorname{OP}_{r}^{[2]}(q)=\frac{(-q^{r};q^{2r})_{\infty}^{2}}{(q^{4r};q^{4r})_{\infty}}g_{2}\left(-q^{r};q^{2r}\right).

Next, for d>2rd>2r, we consider partitions of nn with the following restrictions:

  1. parts are ≡d(mod2d)\equiv d\pmod{2d} or ≡±r(modd)\equiv\pm r\pmod{d},

  2. if there are exactly mm parts ≡d(mod2d)\equiv d\pmod{2d} (counted with multiplicity), then each part ≡±r(modd)\equiv\pm r\pmod{d} is larger than dmdm.

Let pd,r[4](n)p_{d,r}^{[4]}(n) denote the number of such partitions of nn and set pd,r[4](0):=1p_{d,r}^{[4]}(0):=1. For example, p4,1[4](5)=3p_{4,1}^{[4]}(5)=3 enumerated by 55, 3+1+13+1+1, 1+1+1+1+11+1+1+1+1. Let Pd,r[4](q):=∑n≥0pd,r[4](n)qnP_{d,r}^{[4]}(q):=\sum_{n\ge0}p_{d,r}^{[4]}(n)q^{n} denote the corresponding generating function.

Theorem 1.6. We have, for d>2rd>2r,

Pd,r[4](q)=1(q2d;q2d)∞g2(qr;qd).P_{d,r}^{[4]}(q)=\frac{1}{(q^{2d};q^{2d})_{\infty}}g_{2}\left(q^{r};q^{d}\right).

Finally, for d=2rd=2r, we consider two-color partitions of nn with the following restrictions:

  1. parts are divisible by rr but not by 4r4r,

  2. parts ≡2r(mod4r)\equiv2r\pmod{4r} appear only in blue,

  3. if there are exactly mm parts ≡2r(mod4r)\equiv2r\pmod{4r} (counted with multiplicity), then every part ≡r(mod2r)\equiv r\pmod{2r} is larger than 2rm2rm.

Let pr[3](n)p_{r}^{[3]}(n) denote the number of such two-color partitions of nn and set pr[3](0):=1p_{r}^{[3]}(0):=1. For example, p2[3](8)=10p_{2}^{[3]}(8)=10 enumerated by 6g+2g6_g+2_g, 6b+2g6_b+2_g, 6g+2b6_g+2_b, 6b+2b6_b+2_b, 4b+4b4_b+4_b, 2g+2g+2g+2g2_g+2_g+2_g+2_g, 2g+2g+2g+2b2_g+2_g+2_g+2_b, 2g+2g+2b+2b2_g+2_g+2_b+2_b, 2g+2b+2b+2b2_g+2_b+2_b+2_b, 2b+2b+2b+2b2_b+2_b+2_b+2_b. Let Pr[3](q):=∑n≥0pr[3](n)qnP_{r}^{[3]}(q):=\sum_{n\ge0}p_{r}^{[3]}(n)q^{n} denote the corresponding generating function.

Theorem 1.7. We have

Pr[3](q)=1(q4r;q4r)∞g2(qr;q2r).P_{r}^{[3]}(q)=\frac{1}{(q^{4r};q^{4r})_{\infty}}g_{2}\left(q^{r};q^{2r}\right).

This note is organized as follows. In Section 2, we recall qq-series transformations which are required to prove Theorem 1.1. Section 3 presents the proofs of Theorems 1.1–1.7. Section 4 is devoted to open problems for future research.

Acknowledgements

The authors have received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 101001179).

Preliminaries

In this section we recall qq-series identities4 which are used in the proofs of our results. We start with Fine’s identities. To state these, following [6 equation (1.1)], for a∈Ca \in\mathbb{C}, b≠q−mb \ne q^{-m} for every m∈Nm \in\mathbb{N}, and ∣w∣<1|w| < 1, we define

F(a,b;w;q):=∑n≥0(aq)n(bq)nwn.F(a,b;w;q) := \sum_{n\geq0} \frac{(aq)_n}{(bq)_n}w^n.

By [6 equation (11.3)], we have the following identity.

Lemma 2.1. We have

(bq,wq)∞(q)∞F(0,b;w;q)=11−w∑n≥0(b,w)n(q)nqn.\frac{(bq,wq)_\infty}{(q)_\infty}F(0,b;w;q)=\frac{1}{1-w}\sum_{n\geq0}\frac{(b,w)_n}{(q)_n}q^n.

Next, from [6 (12.3)], we obtain the following lemma.

Lemma 2.2. We have

(1−w)F(0,b;w;q)=∑n≥0(bw)n(bq,wq)nqn2.(1-w)F(0,b;w;q)=\sum_{n\geq0}\frac{(bw)_n}{(bq,wq)_n}q^{n^2}.

Next we require the following transformation from [7 (III.10)].

Lemma 2.3. We have

∑n≥0(aqbc,d,e)n(q,aqb,aqc)n(aqde)n=(aqd,aqe,aqbc)∞(aqb,aqc,aqde)∞∑n≥0(aqde,b,c)n(q,aqd,aqe)n(aqbc)n.\sum_{n\geq0}\frac{\left(\frac{aq}{bc},d,e\right)_n}{\left(q,\frac{aq}{b},\frac{aq}{c}\right)_n}\left(\frac{aq}{de}\right)^n = \frac{\left(\frac{aq}{d},\frac{aq}{e},\frac{aq}{bc}\right)_\infty}{\left(\frac{aq}{b},\frac{aq}{c},\frac{aq}{de}\right)_\infty} \sum_{n\geq0}\frac{\left(\frac{aq}{de},b,c\right)_n}{\left(q,\frac{aq}{d},\frac{aq}{e}\right)_n}\left(\frac{aq}{bc}\right)^n.

Moreover, we require the following identity due to Euler (see [1 (2.2.5)]).

Lemma 2.4. For ∣w∣,∣q∣<1|w|, |q| < 1, we have

∑n≥0wn(q)n=1(w)∞.\sum_{n\geq0}\frac{w^n}{(q)_n}=\frac{1}{(w)_\infty}.

Proof of Theorems 1.1–1.7

We start with the proof of Theorem 1.1.

Proof of Theorem 1.1. Combining Lemmas 2.1 and 2.2, we obtain

∑n≥0(bw)n(bq,wq)nqn2=(q)∞(bq,wq)∞∑n≥0(b,w)n(q)nqn.\sum_{n\ge0}\frac{(bw)_n}{(bq,wq)_n}q^{n^2}=\frac{(q)_\infty}{(bq,wq)_\infty}\sum_{n\ge0}\frac{(b,w)_n}{(q)_n}q^n.

Letting (b,w)=(ζ,qζ)(b,w)=(\zeta,\frac{q}{\zeta}), it follows that

∑n≥0qn(n+1)(ζq,q2ζ)n=(q)∞(ζq,q2ζ)∞∑n≥0(ζ,qζ)n(q)nqn.\sum_{n\ge0}\frac{q^{n(n+1)}}{(\zeta q,\frac{q^2}{\zeta})_n}=\frac{(q)_\infty}{(\zeta q,\frac{q^2}{\zeta})_\infty}\sum_{n\ge0}\frac{(\zeta,\frac{q}{\zeta})_n}{(q)_n}q^n.

This and the definition of g3g_3 from (1.1) conclude the proof of Theorem 1.1 (1). Next, using the definition of g2g_2 in (1.1) and the fact that

lim⁡w→0(−qw)nwn=qn(n+1)2,\lim_{w\to0}\left(-\frac{q}{w}\right)_n w^n=q^{\frac{n(n+1)}{2}},

we have

(1−ζ)(1−qζ)g2(ζ;q)=lim⁡w→0∑n≥0(−q,−qw)n(ζq,q2ζ)nwn.(4)(1-\zeta)\left(1-\frac{q}{\zeta}\right)g_2(\zeta;q)=\lim_{w\to0}\sum_{n\ge0}\frac{(-q,-\frac{q}{w})_n}{(\zeta q,\frac{q^2}{\zeta})_n}w^n. \tag*{(4)}

Now, letting (a,b,c,d,e)=(q,qζ,ζ,−q,−qw)(a,b,c,d,e)=(q,\frac{q}{\zeta},\zeta,-q,-\frac{q}{w}) in Theorem 2.3, we get

∑n≥0(−q,−qw)n(ζq,q2ζ)nwn=(−q,−wq,q)∞(ζq,q2ζ,w)∞∑n≥0(wq,qζ,ζ)n(q,−q,−wq)nqn.\sum_{n\ge0}\frac{(-q,-\frac{q}{w})_n}{(\zeta q,\frac{q^2}{\zeta})_n}w^n=\frac{(-q,-wq,q)_\infty}{(\zeta q,\frac{q^2}{\zeta},w)_\infty}\sum_{n\ge0}\frac{(wq,\frac{q}{\zeta},\zeta)_n}{(q,-q,-wq)_n}q^n.

Taking the limit w→0w\to0 on both sides, and using (3.1), we conclude the proof of Theorem 1.1 (2). □\square

We next prove Theorem 1.2.

Proof of Theorem 1.2. Using the definition of OP⁡r[1](n)\operatorname{OP}^{[1]}_r(n), it is not hard to see that

OP⁡r[1](q)=∑n≥0(−qr;q2r)n2(q2r;q2r)nq2rn.\operatorname{OP}^{[1]}_r(q)=\sum_{n\ge0}\frac{(-q^r;q^{2r})_n^2}{(q^{2r};q^{2r})_n}q^{2rn}.

Here, for n∈Nn\in\mathbb{N}, the largest part is 2rn2rn, while n=0n=0 accounts for the empty partition. Letting q↦q2rq\mapsto q^{2r} and ζ=−qr\zeta=-q^r in Theorem 1.1 (1), we thus obtain the claim. □\square

Now we establish Theorem 1.3.

Proof of Theorem 1.3. By definition of pd,r[2](n)p^{[2]}_{d,r}(n), we have

Pd,r[2](q)=∑n≥0qdn(qd;qd)n(qdn+r,qdn+d−r;qd)∞.P^{[2]}_{d,r}(q)=\sum_{n\geq0}\frac{q^{dn}}{(q^d;q^d)_n(q^{dn+r},q^{dn+d-r};q^d)_\infty}.

Here, for n∈Nn\in\mathbb{N}, the largest part divisible by dd is dndn, while n=0n=0 accounts for partitions with no parts divisible by dd. Letting q↦qdq\mapsto q^d and ζ=qr\zeta=q^r in Theorem 1.1 (1), we thus obtain the claim. □\square

Next, we prove Theorem 1.4.

Proof of Theorem 1.4. By definition of pr[2](n)p^{[2]}_r(n), we obtain

Pr[2](q)=∑n≥0q2rn(q2r;q2r)n(q2rn+r;q2r)∞2.P^{[2]}_r(q)=\sum_{n\geq0}\frac{q^{2rn}}{(q^{2r};q^{2r})_n(q^{2rn+r};q^{2r})_\infty^2}.

Letting q↦q2rq\mapsto q^{2r} and ζ=qr\zeta=q^r in Theorem 1.1 (1), we thus obtain the claim. □\square

Next, we show the following elementary lemma which we require to prove Theorems 1.5–1.7.

Lemma 3.1. For k∈Nk\in\mathbb{N} and m∈N0m\in\mathbb{N}_0, qkm(q2k;q2k)m\frac{q^{km}}{(q^{2k};q^{2k})_m} is the generating function for partitions into exactly mm parts which are ≡k(mod2k)\equiv k\pmod{2k}.

Proof. Introduce a variable ww to record the number of parts. For each allowed part k(2j+1)k(2j+1) (j∈N0j\in\mathbb{N}_0), we have

11−wqk(2j+1)=∑r≥0wrqk(2j+1)r.\frac{1}{1-wq^{k(2j+1)}}=\sum_{r\geq0}w^rq^{k(2j+1)r}.

The term indexed by rr corresponds to taking rr copies of this part. Thus the coefficient of wmw^m in

∏j≥011−wqk(2j+1)=1(wqk;q2k)∞\prod_{j\geq0}\frac{1}{1-wq^{k(2j+1)}}=\frac{1}{(wq^k;q^{2k})_\infty}

is the generating function for partitions into exactly mm parts congruent to k(mod2k)k\pmod{2k}. Letting q↦q2kq\mapsto q^{2k} and w↦wqkw\mapsto wq^k in Theorem 2.4, we obtain

∑n≥0qkn(q2k;q2k)nwn=1(wqk;q2k)∞.\sum_{n\geq0}\frac{q^{kn}}{(q^{2k};q^{2k})_n}w^n=\frac{1}{(wq^k;q^{2k})_\infty}.

Comparing the coefficients of wmw^m on both sides gives the claim. □\square

Now we show Theorem 1.5.

Proof of Theorem 1.5. Let m∈N0m \in\mathbb{N}_0 denote the number of parts ≡2r(mod4r)\equiv2r \pmod{4r} (including repetitions) of an overpartition counted by op⁡r[2](n)\operatorname{op}^{[2]}_r(n). By (2), these parts are non-overlined. By Theorem 3.1 with k=2rk=2r, they are generated by

q2rm(q4r;q4r)m.\frac{q^{2rm}}{(q^{4r};q^{4r})_m}.

By (4), the parts ≡r(mod2r)\equiv r \pmod{2r} lie in {r(2k−1):1≤k≤m}\{r(2k-1):1\leq k\leq m\}. Moreover, by (3), each of these parts appears at most once non-overlined and at most once overlined. Hence these parts are generated by

∏j=0m−1(1+qr(2j+1))2=(−qr;q2r)m2.\prod_{j=0}^{m-1}\left(1+q^{r(2j+1)}\right)^2=(-q^r;q^{2r})_m^2.

Multiplying these two generating functions and summing over mm, we obtain

OP⁡r[2](q)=∑m≥0(−qr;q2r)m2(q4r;q4r)mq2rm.\operatorname{OP}^{[2]}_r(q)=\sum_{m\geq0}\frac{(-q^r;q^{2r})_m^2}{(q^{4r};q^{4r})_m}q^{2rm}.

Letting q↦q2rq\mapsto q^{2r} and ζ=−qr\zeta=-q^r in Theorem 1.1 (2), we thus obtain the claim. □

Next, we prove Theorem 1.6.

Proof of Theorem 1.6. Let m∈N0m\in\mathbb{N}_0 be the number of parts ≡d(mod2d)\equiv d\pmod{2d} (including repetitions) of a partition counted by pd,r[4](n)p^{[4]}_{d,r}(n). By Theorem 3.1 with k=dk=d, these parts are generated by

qdm(q2d;q2d)m.\frac{q^{dm}}{(q^{2d};q^{2d})_m}.

By (2), the parts ≡r(modd)\equiv r\pmod d are at least dm+rdm+r and the parts ≡−r(modd)\equiv-r\pmod d are at least dm+d−rdm+d-r. These parts are therefore generated by

1(qdm+r,qdm+d−r;qd)∞.\frac{1}{(q^{dm+r},q^{dm+d-r};q^d)_\infty}.

Multiplying these two generating functions and summing over mm, we obtain

Pd,r[4](q)=∑m≥0qdm(q2d;q2d)m(qdm+r,qdm+d−r;qd)∞.P^{[4]}_{d,r}(q)=\sum_{m\geq0}\frac{q^{dm}}{(q^{2d};q^{2d})_m(q^{dm+r},q^{dm+d-r};q^d)_\infty}.

Letting q↦qdq\mapsto q^d and ζ=qr\zeta=q^r in Theorem 1.1 (2), we thus obtain the claim. □

Finally, we establish Theorem 1.7.

Proof of Theorem 1.7. Let m∈N0m\in\mathbb{N}_0 denote the number of parts ≡2r(mod4r)\equiv2r\pmod{4r} (including repetitions) of a two-color partition counted by pr[3](n)p^{[3]}_r(n). By (2), these parts appear in blue, so by Theorem 3.1 with k=2rk=2r, they are generated by

q2rm(q4r;q4r)m.\frac{q^{2rm}}{(q^{4r};q^{4r})_m}.

By (2) and (3), the parts ≡r(mod2r)\equiv r\pmod{2r} are at least 2rm+r2rm+r. They are therefore generated by

1(q2rm+r;q2r)∞2.\frac{1}{(q^{2rm+r};q^{2r})_\infty^2}.

Multiplying these two generating functions and summing over mm, we obtain

Pr[3](q)=∑m≥0q2rm(q4r;q4r)m(q2rm+r;q2r)∞2.P^{[3]}_r(q)=\sum_{m\geq0}\frac{q^{2rm}}{(q^{4r};q^{4r})_m(q^{2rm+r};q^{2r})_\infty^2}.

Letting q↦q2rq \mapsto q^{2r} and ζ=qr\zeta= q^r in Theorem 1.1 (2), we thus obtain the claim. □

Open problems

We conclude this paper with the following problems for future research:

(1) Numerically,5 we found the following congruences:

op⁡1[1](12n+8)≡0(mod2),\operatorname{op}_{1}^{[1]}(12n+8) \equiv0 \pmod{2},
op⁡1[1](20n+j)≡0(mod2)with j∈{8,16},\operatorname{op}_{1}^{[1]}(20n+j) \equiv0 \pmod{2}\quad\text{with }j\in\{8,16\},
op⁡1[1](28n+j)≡0(mod2)with j∈{8,16,20}.\operatorname{op}_{1}^{[1]}(28n+j) \equiv0 \pmod{2}\quad\text{with }j\in\{8,16,20\}.

These lead to the following conjecture. Let (⋅⋅)\left(\frac{\cdot}{\cdot}\right) denote the Legendre symbol.

Conjecture 4.1. For any odd prime pp and k∈N0k\in\mathbb{N}_0, we have

2∣op⁡1[1](4pn+4k)for all n∈N0⟺(8k+1p)=−1.2\mid\operatorname{op}_{1}^{[1]}(4pn+4k)\quad\text{for all }n\in\mathbb{N}_0 \Longleftrightarrow\left(\frac{8k+1}{p}\right)=-1.

(2) For op⁡1[2](n)\operatorname{op}_{1}^{[2]}(n), numerically, we observe the following congruences:

op⁡1[2](18n+j)≡0(mod4)with j∈{8,14},\operatorname{op}_{1}^{[2]}(18n+j) \equiv0 \pmod{4}\quad\text{with }j\in\{8,14\},
op⁡1[2](50n+j)≡0(mod4)with j∈{16,26,36,46},\operatorname{op}_{1}^{[2]}(50n+j) \equiv0 \pmod{4}\quad\text{with }j\in\{16,26,36,46\},
op⁡1[2](98n+j)≡0(mod4)with j∈{26,40,54,68,82,96}.\operatorname{op}_{1}^{[2]}(98n+j) \equiv0 \pmod{4}\quad\text{with }j\in\{26,40,54,68,82,96\}.

These suggest the following general pattern.

Conjecture 4.2. Let pp be an odd prime with p≢1(mod8)p\not\equiv1\pmod{8}. Then for all n∈N0n\in\mathbb{N}_0 and 1≤k≤p−11\leq k\leq p-1, we have

op⁡1[2](2p2n+p2−14+2pk)≡0(mod4).\operatorname{op}_{1}^{[2]}\left(2p^2n+\frac{p^2-1}{4}+2pk\right)\equiv0\pmod{4}.

University of Cologne, Department of Mathematics and Computer Science, Weyertal 86-90, 50931 Cologne, Germany

Email address: [email protected]

Email address: [email protected]

References

  1. [1]G. Andrews, The theory of partitions, Cambridge University Press, Cambridge, 1998.
  2. [2]G. Andrews, Partitions with short sequences and mock theta functions, Proc. Natl. Acad. Sci. USA 102 (2005), 4666–4671.
  3. [3]B. Berndt and R. Rankin, Ramanujan: letters and commentary, History of Mathematics 9 (1995), Amer. Math. Soc., Providence, RI.
  4. [4]K. Bringmann, J. Lovejoy, and K. Mahlburg, A partition identity and the universal mock theta function g_2, Math. Res. Lett. 23 (2016), 67–80.arxiv.org/abs/1311.5483
  5. [5]F. Dyson, Some guesses in the theory of partitions, Eureka (Cambridge) 8 (1944), 10–15.
  6. [6]N. Fine, Basic Hypergeometric Series and Applications, American Mathematical Society, Providence, RI, 1988.DOI
  7. [7]G. Gasper and M. Rahman, Basic hypergeometric series, Encycl. of Math. and Applications 35, Cambridge University Press, Cambridge, 1990.
  8. [8]B. Gordon and R. McIntosh, A survey of classical mock theta functions, in partitions, q-series, and modular forms, Dev. Math. 23 (Springer, New York, 2012), 95–144.DOI
  9. [9]P. MacMahon, Combinatory Analysis, Cambridge Univ. Press, Cambridge, U.K., Vol. II (1916), 49–58.
  10. [10]S. Ramanujan, Some properties of p(n), the number of partitions of n, Proc. Cambridge Philos. Soc. 19 (1919), 207–210.
  11. [11]S. Zwegers, Mock theta functions, Ph.D. Thesis, Universität Utrecht (2002).

Paper details

vlnc:W_585mqpkq
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents