Real zeros of a class of hyperbolic polynomials with random coefficients (Q1751352)

Summary: We have proved here that the expected number of real zeros of a random hyperbolic polynomial of the form \(y=P_n(t) = \sqrt{(^n_1)}a_1\) cosh \(t + \sqrt {(^n_2)}a_2\) cosh \(2t+ \dots+ \sqrt {(^n_n)}a_n\) cosh \(nt\), where \(a_1,\dots,a_n\) is a sequence of standard Gaussian random variables, is \(\sqrt n/2 +o_p(1)\). It is shown that the asymptotic value of expected number of times the polynomial crosses the level \(y = K\) is also \(\sqrt n/2\) as long as \(K\) does not exceed \(\sqrt {2^ne^{\mu(n)}}\), where \(\mu(n) = o(n)\). The number of oscillations of \(P_n(t)\) about \(y=K\) will be less than \(\sqrt n/2\) asymptotically only if \(K= \sqrt {2^ne^{\mu(n)}}\), where \(\mu(n) = O(n)\) or \(n^{-1}\mu(n) \to\infty\). In the former case the number of oscillations continues to be a fraction of \(\sqrt n\) and decreases with the increase in value of \(\mu(n)\). In the latter case, the number of oscillations reduces to \(o_p(\sqrt n)\) and almost no trace of the curve is expected to be present above the level \(y=K\) if \(\mu(n)/(n \log n) \to \infty\).