Medina J S, Prosmiti R, Villarreal P, Delgado-Barrio G, Alemán J V
Instituto de Física Fundamental, IFF-CSIC, Serrano 123, Madrid ES-28006, Spain.
Phys Rev E Stat Nonlin Soft Matter Phys. 2011 Dec;84(6 Pt 2):066703. doi: 10.1103/PhysRevE.84.066703. Epub 2011 Dec 16.
An approximation to the Fourier transform (FT) of the Kohlrausch function (stretched exponential) with shape parameter 0 < β ≤ 1 is presented by using Havriliak-Negami-like functions. Mathematical expressions to fit their parameters α, γ, and τ, as functions of β (0 < β ≤ 1 and 1 < β < 2) are given, which allows a quick identification in the frequency domain of the corresponding shape factor β. Reconstruction via fast Fourier transform of frequency approximants to time domain are shown as good substitutes in short times though biased in long ones (increasing discrepancies as β → 1). The method is proposed as a template to commute time and frequency domains when analyzing complex data. Such a strategy facilitates intensive algorithmic search of parameters while adjusting the data of one or several Kohlrausch-Williams-Watts relaxations.
通过使用类哈弗里利亚克-内加米函数,给出了形状参数为(0 < β ≤ 1)的科尔劳施函数(拉伸指数函数)的傅里叶变换(FT)的近似值。给出了将其参数(α)、(γ)和(τ)拟合为(β)((0 < β ≤ 1)和(1 < β < 2))的函数的数学表达式,这使得能够在频域中快速识别相应的形状因子(β)。通过快速傅里叶变换将频域近似值重构到时域,在短时间内显示为良好的替代方法,尽管在长时间内存在偏差(随着(β → 1),差异增大)。该方法被提议作为分析复杂数据时在时域和频域之间转换的模板。这种策略有助于在调整一个或几个科尔劳施-威廉姆斯-瓦茨弛豫的数据时,对参数进行密集的算法搜索。