Unimodality of the Andrews-Garvan-Dyson cranks of partitions (Q2054238)

Dyson's rank and Andrews-Garvan-Dyson crank are two fundamental statistics in the theory of partitions. The rank of a partition was introduced by \textit{F. J. Dyson} [``Some guesses in the theory of partitions'', Eureka 10--15 (1944)] as the largest part of the partition minus the number of parts. The crank of a partition was defined by \textit{G. E. Andrews} and \textit{F. G. Garvan} [Bull. Am. Math. Soc., New Ser. 18, No. 2, 167--171 (1988; Zbl 0646.10008)] as either the largest part if the partition contains no ones, or the number of parts larger than the number of ones minus the number of ones. In this paper under review, the authors investigate the distribution of the Andrews-Garvan-Dyson cranks of partitions. Let \(N(m,n)\) and \(M(m,n)\) denote the number of partitions of \(n\) with the Dyson rank \(m\) and the Andrews-Garvan-Dyson crank \(m\), respectively. They prove that the sequence \(\{M(m,n)\}_{|m|\leq n-1}\) is unimodal for any \(n\geq44\). The unimodality of \(\{M(m,n)\}_{|m|\leq n-1}\) is related to the monotonicity properties of two partition functions \(p_r(n)\) and \(pp_r(n)\), where \(p_r(n)\) and \(pp_r(n)\) was, respectively, defined by \[ \sum_{n=0}^\infty p_r(n)q^n=\dfrac{1}{(q^2;q)_{r-1}}\qquad\text{and}\qquad \sum_{n=0}^\infty pp_r(n)q^n=\dfrac{1}{(q^2;q)_{r-1}(q^2;q)_r}, \] where \[ (a;q)_n=\prod_{k=0}^{n-1}(1-aq^k). \] The authors prove that \(p_r(n)\geq p_r(n-1)\) for any \(r\geq5\) and \(n\geq14\) and \(pp_r(n)\geq pp_r(n-1)\) for any \(r\geq3\) and \(n\geq8\). Based on the monotonicity properties on \(p_r(n)\) and \(pp_r(n)\), they prove that \[ M(m,n)\geq M(m,n-1)\quad\text{for any }n\geq14,\;0\leq m\leq n-2, \] and \[ M(m-1,n)\geq M(m,n)\quad\text{for any }n\geq44,\;1\leq m\leq n-1. \tag{1} \] The authors further find that the inequality (1) implies that the sequence \(\{M(m,n)\}_{|m|\leq n-1}\) is unimodal for any \(n\geq44\) upon the symmetry of \(M(m,n)\). Moreover, the authors also provide a proof of an upper bound for \(\mathrm{ospt}(n)\) conjectured by \textit{S. H. Chan} and \textit{R. Mao} [Adv. Math. 258, 414--437 (2014; Zbl 1294.11179)] in view of (1), where the function \(\mathrm{ospt}(n)\) denotes the difference between the first positive rank and crank moments, is given by \[ \mathrm{ospt}(n)=\sum_{m=0}^\infty m\big(M(m,n)-N(m,n)\big). \]