Abstract
This paper tells the story how a MAPLE-assisted quest for an interesting undergraduate problem in trigonometric series led some "amateurs" to the discovery that the one-parameter family of deterministic trigonometric series Sp: t {mapping} Σn∈ℕsin(n-pt), p > 1, exhibits both order and apparent chaos, and how this has prompted some professionals to offer their expert insights. As to order, an elementary (undergraduate) proof is given that ∀t∈ℝ, with explicitly computed constant αp. As to chaos, the seemingly erratic fluctuations about this overall trend are discussed. Experts' commentaries are reproduced as to why the fluctuations of Sp(t) - αpsign(t) {pipe}t{pipe}1/p are presumably not Gaussian. Inspired by a central limit type theorem of Marc Kac, a well-motivated conjecture is formulated to the effect that the fluctuations of the ⌈t1/(p+1)⌉-th partial sum of Sp(t), when properly scaled, do converge in distribution to a standard Gaussian when t→∞, though-provided that p is chosen so that the frequencies {n-p}n∈ℕ are rationally linear independent; no conjecture has been forthcoming for rationally dependent {n-p}n∈ℕ. Moreover, following other experts' tip-offs, the interesting relationship of the asymptotics of Sp(t) to properties of the Riemann ζ function is exhibited using the Mellin transform.
Original language | English (US) |
---|---|
Pages (from-to) | 572-600 |
Number of pages | 29 |
Journal | Journal of Statistical Physics |
Volume | 150 |
Issue number | 3 |
DOIs | |
State | Published - Feb 2013 |
All Science Journal Classification (ASJC) codes
- Statistical and Nonlinear Physics
- Mathematical Physics
Keywords
- Deterministic chaos
- Fourier transform
- Kac central limit theorem
- Markov-Lévy method of characteristic functions
- Mellin transform
- Riemann ζ function
- Sine series
- Steinhaus notion of statistical independence of functions
- Tempered distributions