The growth of Taylor series and random Taylor series (Q2720088)
From MaRDI portal
| This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use this page instead for the normal view: The growth of Taylor series and random Taylor series |
scientific article; zbMATH DE number 1610580
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | The growth of Taylor series and random Taylor series |
scientific article; zbMATH DE number 1610580 |
Statements
6 June 2002
0 references
Taylor series
0 references
random Taylor series
0 references
order
0 references
The growth of Taylor series and random Taylor series (English)
0 references
Consider \(f(z)= \sum^\infty_{n=0} b_nz^n\) and \(f_\omega (z)= \sum^\infty_{n=0} Z_n(\omega) b_nz^n\), where \(\{b_n\} \subset\mathbb{C}\) and \(\{Z_n (\omega)\}\) is a sequence of independent and symmetric complex or real random variables of the same distribution and veryfying some conditions in a probability space \((\Omega, {\mathfrak A},P)\). Suppose that \(\varlimsup\root n \of {b_n}=0\) and \(\varlimsup\root n\of {|b_nZ_n (\omega)|}=0\) a.s. \((n\to \infty)\) and let \(M(r,f)= \max\{|f(z)|: |z|=r\}\), \(M(r, f_\omega) =\max\{|f_\omega(z) |:|z|= r\}\). The authors prove that (1) \(\varlimsup (\ln\ln M(r,f)/ \ln\ln r)= \lambda\) \((n\to\infty)\) and (2) \(\varlimsup (\ln\ln M(r, f_\omega)/ \ln\ln r)= \lambda\) a.s. \((n\to \infty)\) if and only if NEWLINE\[NEWLINE\varlimsup [\ln n/(|n|-|n|b_n|/n |)]+ 1= \lambda \in(1,\infty)\;(n\to\infty).NEWLINE\]NEWLINE They replace \(\varlimsup\) in (1) and (2) by lim and, for existence of these limits, give also some necessary and sufficient conditions which are the above condition and an additional condition. The results are analogous to the classical ones.
0 references