English

Random Polynomials in Several Complex Variables

Complex Variables 2024-01-29 v2 Probability

Abstract

We generalize some previous results on random polynomials in several complex variables. A standard setting is to consider random polynomials Hn(z):=j=1mnajpj(z)H_n(z):=\sum_{j=1}^{m_n} a_jp_j(z) that are linear combinations of basis polynomials {pj}\{p_j\} with i.i.d. complex random variable coefficients {aj}\{a_j\} where {pj}\{p_j\} form an orthonormal basis for a Bernstein-Markov measure on a compact set KCdK\subset {\bf C}^d. Here mnm_n is the dimension of Pn\mathcal P_n, the holomorphic polynomials of degree at most nn in Cd{\bf C}^d. We consider more general bases {pj}\{p_j\}, which include, e.g., higher-dimensional generalizations of Fekete polynomials. Moreover we allow Hn(z):=j=1mnanjpnj(z)H_n(z):=\sum_{j=1}^{m_n} a_{nj}p_{nj}(z); i.e., we have an array of basis polynomials {pnj}\{p_{nj}\} and random coefficients {anj}\{a_{nj}\}. This always occurs in a weighted situation. We prove results on convergence in probability and on almost sure convergence of 1nlogHn\frac{1}{n}\log |H_n| in Lloc1(Cd)L^1_{loc}({\bf C}^d) to the (weighted) extremal plurisubharmonic function for KK. We aim for weakest possible sufficient conditions on the random coefficients to guarantee convergence.

Keywords

Cite

@article{arxiv.2112.00880,
  title  = {Random Polynomials in Several Complex Variables},
  author = {Turgay Bayraktar and Tom Bloom and Norm Levenberg},
  journal= {arXiv preprint arXiv:2112.00880},
  year   = {2024}
}

Comments

This replaces and improves a previous version which has a gap in the proof of the higher codimension case at the end