A new special function related to a discrete Gauss-Poisson distribution and some physics of the cell model with Curie-Weiss interactions
O. A. Dobush, M. A. Shpot · 2024
Read A new special function related to a discrete Gauss-Poisson distribution and some physics of the cell model with Curie-Weiss interactions by O. A. Dobush, M. A. Shpot (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-05428.txt.
Introduction
O. A. Dobush and M. A. Shpot introduce a new probability distribution and the special function that normalizes it, motivated by a cell model of simple fluids. The distribution blends a Poisson law with a Gaussian-style penalty on large counts. Its normalizing sum defines the new function, which the authors study in depth. Writing in twenty twenty-four, they derive asymptotic estimates for large values of one parameter. The same treatment covers the function's logarithm and its logarithmic derivative.
Beyond a threshold in the second parameter, the exact function oscillates around its asymptotic form. That wiggle is the mathematical fingerprint of interesting physics. The authors connect the results to the Curie-Weiss cell model, where particles interact in a mean-field way inside cells. The distribution may serve other counting problems with suppressed tails.
Catalog record
| CREATOR | O. A. Dobush, M. A. Shpot |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 51 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-05428
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A new special function related to a discrete Gauss-Poisson distribution and some physics of the cell model with Curie-Weiss interactions O. A. Dobush, M. A. Shpot Abstract Inspired by previous studies in statistical physics [ \big[ see, in particular, Kozitsky at al., A phase transition in a Curie-Weiss system with binary interactions , Condens. Matter Phys. 23 , 23502 (2020) ] \big] we introduce a discrete Gauss-Poisson probability distribution function p G P ( n , z , r ) = [ R ( r , z ) ] − 1 e z n n ! e − 1 2 r n 2 p_{GP}(n;z,r)=\left[R(r;z)\right]^{-1}\frac{{\rm e}^{zn}}{n!}\,\mbox{e}^{-\frac{1}{2}\,rn^{2}} (A1) with support on ℕ 0 \mathbb{N}_{0} and parameters z ∈ ℝ z\in\mathbb{R} and r ∈ ℝ + r\in\mathbb{R}_{+} . The probability mass function p G P ( n , z , r ) p_{GP}(n;z,r) is normalized by the special function R ( r , z ) R(r;z) , given by the infinite sum R ( r , z ) = ∑ n = 0 ∞ e z n n ! e − 1 2 r n 2 , R(r;z)=\sum_{n=0}^{\infty}\frac{{\rm e}^{zn}}{n!}\,{\rm e}^{-\frac{1}{2}\,rn^{2}}, (A2) possessing extremely intersting mathematical properties. We present an asymptotic estimate R ( as ) ( r , z ≫ 1 ) R^{(\rm as)}(r;z\gg 1) for the function R ( r , z ) R(r;z) with large arguments z z , along with similar formulas for its logarithm and logarithmic derivative. These functions exhibit very interesting oscillatory behavior around their asymptotics, for parameters r r above some threshold value r ∗ r^{*} . Some implications of our findings are discussed in the context of the Curie-Weiss cell model of simple fluids. Preprint ArXiv 2412.05428 [math.CA] (2024) with corrections of several misprints Keywords. Phase transition, fluid’s cell model, Curie-Weiss interactions, discrete Gauss-Poisson distribution, special functions, asymptotic behavior 2020 Mathematics Subject Classification. Primary 33C20; Secondary 33C05 1 Introduction Although a number of physical aspects of the cell model with binary Curie-Weiss interactions will be further discussed, the main message of the present communication concerns the special function R(r;z) that normalizes the discrete Gauss-Poisson probability distribution function introduced in ( A1 ). We come up with several useful physical conclusions throughout the paper. However, our primary goal is to bring the essentially unknown mathematical objects ( A1 ) and ( A2 ) arising in a specific branch of theoretical physics to the broad mathematical community. To the best of our knowledge, neither the discrete Gauss-Poisson probability distribution function p_{GP}(n;z,r) nor the special function R(r;z) have been considered in the mathematical literature before. The function R(r;z) seems to appear for the first time in the physical paper [ 1 , p. 440] where the sum of exactly the same functional form as in ( A2 ) resulted from certain integration and started with n=2 . Later [ 2 , p. 20] , an analogous sum appears in even more general form, where the second power of n in the summand is replaced by n^{m} . In different disguises, the function R(r;z) can be found in a series of papers [ 3 , p. 810, (3.15)] , [ 4 , p. 5, (2.15)] , [ 5 , p. 3, (17)] , [ 6 , p. 3, (9)] , where the statistical mechanics calculations were based on grand canonical ensemble (see e.g. [ 7 , Sec. 2.4] ). Apparently, the functional form of the function R(r;z) in ( A2 ) should be characteristic to calculations using this kind of statistics; actually, its general shape is the same as that of the grand-canonical partition function \Xi=\sum_{N=0}^{\infty}z^{N}\,Z_{N}/N! (see [ 7 , (2.4.6)] and ( 5 ) below). Apart from the sum R(r;z) alone, for actual calculations of the mentioned papers several lowest-order moments M_{m}(z,r)=\left[R(r;z)\right]^{-1}\sum_{n=0}^{\infty}\;n^{m}\;\frac{{\rm e}^{zn}}{n!}\,{\rm e}^{-\frac{1}{2}\,rn^{2}},\qquad\qquad m\in\mathbb{N} (1) of the distribution ( A1 ) have been needed. However again, as far as we know, no mathematical properties of the sum R(r;z) or the moments M_{m}(z,r) have been studied in any physically-oriented work, apart from their numerical evaluations using truncated series of exponentials. The truncation of the infinite series in ( A2 ) or ( 1 ) is indeed an important dangerous aspect, which could potentially lead to negative consequences. The reason is that the large- z behavior of a finite sum of exponential functions qualitatively differs (at any order of truncation) from the true asymptotics of the infinite sum. We shall return to this issue in a subsequent publication. Our motivation in the present work consists several aspects. ( i ) First of all, we are quite sure that it is of primary interest to initiate a mathematical investigation of the discrete Gauss-Poisson probability distribution function p_{GP}(n;z,r) , its normalization R(r;z) , and, further, the associated moments M(r;z) . In doing so, we have determined the non-trivial asymptotic behavior of R(r;z) along with some related functions and visualized an intriguing oscillating behavior of \ln R(r;z) in the vicinity of its asymptotics. Of course, it would be a highly interesting challenge to derive any explicit determinations of these functions for some special sets of parameters, although such a goal seems to be rather hopeless for the moment. ( ii ) It is certainly expected that mathematical studies of this kind must have appropriate physical consequences and implications. An example is the issue of convergence of the master integral ( 7 ), involving \ln R(r;z) (see also ( 13 )), in the marginal case of physical stability discussed in Sections 4 and 6 . We hope that the list of such applications will be extended in the future. ( iii ) Our desire is to draw attention of mathematicians to the interesting and practically important mathematical objects discussed in the present communication and trigger their further investigation. ( iv ) On the other side, we observed a lack of explicit analytical results in physical papers where the function R(r;z) and moments M_{m}(z,r) played an essential role, apart from a quite general proof [ 4 ] of the existence of phase transitions in underlying systems. All subsequent work, to be mentioned below, has been dedicated to numerical investigations. Here, we are starting to fill this gap, by presenting several instances of related analytical calculations and their implications. 2 Physical context The special discrete Gauss-Poisson probability distribution function ( A1 ) and associated infinite sums like ( A2 ) and ( 1 ) appear in statistical mechanics investigations of simple classical fluids within the framework of the grand canonical ensemble. The physical basis for the present study is the cell model with binary Curie-Weiss interactions, introduced in [ 8 , 4 ] for the purpose of modelling the simple fluids and describing phase transitions that occur in such physical systems. Extended studies of this model have been subsequently done in [ 9 , 5 ] , and, more recently in [ 6 , 10 ] . In the next section, we are going to describe the underlying physical model while closely following the reference [ 4 ] . 2.1 Two-point interactions and their energy It is assumed that a classical fluid system, which may be in a