|
Zapiski Nauchnykh Seminarov POMI, 2015, Volume 441, Pages 144–153
(Mi znsl6230)
|
|
|
|
This article is cited in 5 scientific papers (total in 5 papers)
Discriminant and root separation of integral polynomials
F. Götzea, D. Zaporozhetsb a Faculty of Mathematics, Bielefeld University, P.O.Box 10 01 31, 33501 Bielefeld, Germany
b St. Petersburg Department of Steklov Institute of Mathematics, Fontanka 27, 191011 St. Petersburg,
Russia
Abstract:
Consider a random polynomial
$$
G_Q(x)=\xi_{Q,n}x^n+\xi_{Q,n-1}x^{n-1}+\dots+\xi_{Q,0}
$$
with independent coefficients uniformly distributed on $2Q+1$ integer points $\{-Q,\dots,Q\}$. Denote by $D(G_Q)$ the discriminant of $G_Q$. We show that there exists a constant $C_n$, depending on $n$ only such that for all $Q\ge2$ the distribution of $D(G_Q)$ can be approximated as follows
$$
\sup_{-\infty\leq a\leq b\leq\infty}\left|\mathbf P\left(a\leq\frac{D(G_Q)}{Q^{2n-2}}\leq b\right)-\int_a^b\varphi_n(x)\,dx\right|\leq\frac{C_n}{\log Q},
$$
where $\varphi_n$ denotes the probability density function of the discriminant of a random polynomial of degree $n$ with independent coefficients which are uniformly distributed on $[-1,1]$. Let $\Delta(G_Q)$ denote the minimal distance between the complex roots of $G_Q$. As an application we show that for any $\varepsilon>0$ there exists a constant $\delta_n>0$ such that $\Delta(G_Q)$ is stochastically bounded from below/above for all sufficiently large $Q$ in the following sense
$$
\mathbf P\left(\delta_n<\Delta(G_Q)<\frac1{\delta_n}\right)>1-\varepsilon.
$$
Key words and phrases:
distribution of discriminants, integral polynomials, polynomial discriminant, polynomial root separation.
Received: 10.10.2015
Citation:
F. Götze, D. Zaporozhets, “Discriminant and root separation of integral polynomials”, Probability and statistics. Part 22, Zap. Nauchn. Sem. POMI, 441, POMI, St. Petersburg, 2015, 144–153; J. Math. Sci. (N. Y.), 219:5 (2016), 700–706
Linking options:
https://www.mathnet.ru/eng/znsl6230 https://www.mathnet.ru/eng/znsl/v441/p144
|
Statistics & downloads: |
Abstract page: | 205 | Full-text PDF : | 55 | References: | 51 |
|