论文标题

汉克尔的决定因素和正交多项式,用于扰动高斯的重量:从有限$ n $到大$ n $渐近。

Hankel Determinant and Orthogonal Polynomials for a Perturbed Gaussian Weight: from Finite $n$ to Large $n$ Asymptotics

论文作者

Min, Chao, Chen, Yang

论文摘要

我们在对称扰动的高斯重量$$ w(x; t):= \ mathrm {e}^{ - x^2} \ left(1+t \:x^2 \ right)^λ,\ qquad x x \ in \ qquad in \ quthbb = r} in p(1+t)中,研究了对称扰动的高斯重量$$ W(x; t):= \ MATHRM {E}^{ - x^2} \ qquad x \ in \ mathbb = \ mathbb {r} $。该权重与信息理论中的单用户MIMO系统有关。结果表明,复发系数$β_n(t)$与特定的painlevév vryscentent有关,子领导系数$ \ mathrm {p}(n,t)$满足jimbo-miwa-okamoto $σ$ - pachlevévEquination的pachlevéVequination。此外,我们分别通过$β_N(t)$和$ \ mathrm {p}(n,t)$得出二阶差方程。这使我们能够借助Dyson的库仑液体方法获得$β_N(t)$和$ \ MATHRM {p}(n,t)$的大$ n $全渐近扩展。我们还考虑由扰动的高斯重量产生的hankel行列式$ d_n(t)$。发现$ h_n(t)$,与$ d_n(t)$的对数衍生物相关的数量,可以用$β_N(t)$和$ \ m atrm {p}(n,t)$表示。基于此结果,我们获得了$ h_n(t)$的大$ n $渐近扩展,然后获得hankel decentant $ d_n(t)$的大型扩展。

We study the monic polynomials orthogonal with respect to a symmetric perturbed Gaussian weight $$ w(x;t):=\mathrm{e}^{-x^2}\left(1+t\: x^2\right)^λ,\qquad x\in \mathbb{R}, $$ where $t> 0,\;λ\in \mathbb{R}$. This weight is related to the single-user MIMO systems in information theory. It is shown that the recurrence coefficient $β_n(t)$ is related to a particular Painlevé V transcendent, and the sub-leading coefficient $\mathrm{p}(n,t)$ satisfies the Jimbo-Miwa-Okamoto $σ$-form of the Painlevé V equation. Furthermore, we derive the second-order difference equations satisfied by $β_n(t)$ and $\mathrm{p}(n,t)$, respectively. This enables us to obtain the large $n$ full asymptotic expansions for $β_n(t)$ and $\mathrm{p}(n,t)$ with the aid of Dyson's Coulomb fluid approach. We also consider the Hankel determinant $D_n(t)$, generated by the perturbed Gaussian weight. It is found that $H_n(t)$, a quantity allied to the logarithmic derivative of $D_n(t)$, can be expressed in terms of $β_n(t)$ and $\mathrm{p}(n,t)$. Based on this result, we obtain the large $n$ asymptotic expansion of $H_n(t)$ and then that of the Hankel determinant $D_n(t)$.

扫码加入交流群

加入微信交流群

微信交流群二维码

扫码加入学术交流群,获取更多资源