Doob's inequalities revisited: A maximal \(H^ 1\)-embedding (Q1107901): Difference between revisions

From MaRDI portal
Set OpenAlex properties.
ReferenceBot (talk | contribs)
Changed an Item
 
Property / cites work
 
Property / cites work: Q3049602 / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q3049603 / rank
 
Normal rank
Property / cites work
 
Property / cites work: Construction of a martingale with given absolute value / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q4150145 / rank
 
Normal rank
Property / cites work
 
Property / cites work: A converse to the dominated convergence theorem / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q5822308 / rank
 
Normal rank
Property / cites work
 
Property / cites work: Every nonnegative submartingale is the absolute value of a martingale / rank
 
Normal rank
Property / cites work
 
Property / cites work: Optimal stopping and best constants for Doob-like inequalities. I: The case \(p=1\) / rank
 
Normal rank

Latest revision as of 17:44, 18 June 2024

scientific article
Language Label Description Also known as
English
Doob's inequalities revisited: A maximal \(H^ 1\)-embedding
scientific article

    Statements

    Doob's inequalities revisited: A maximal \(H^ 1\)-embedding (English)
    0 references
    0 references
    1988
    0 references
    With the aim of finding best constants in Doob's \(L^ p\)-inequalities for nonnegative submartingales the author first establishes a maximal inequality and then an embedding into Brownian motion, with the aid of which he shows that his inequality is sharp. An important fact used is that every nonnegative submartingale can be written as the absolute value of a martingale, first proved by ly continuous spectrum. Theorem 4.1. If for \(p\in {\mathcal P}\) \([c_ 1,\infty)\subset N(w_ p)\subset [c_ 2,\infty)\), \(I(\lambda)=2^{-1}\int ((\xi -\lambda)^{- 1}-\xi (1+\xi)^{-1}d\xi\), \(u=\exp I\), then for \(q\in \sup p p\) there exists a singular measure \(\nu\) on \([c_ 1,c_ 2]\) and solutions \(f_{1,2}\) of \(L(q)p=\xi f\) for a.e. \(\xi\) with respect to \(\nu\) such that if \(g_{\lambda}(x,y)\) is the Green function of L(q), then \[ u(\lambda)g_{\lambda}(x,x)=\int (f^ 2_ 1+f^ 2_ 2)(\xi - \lambda)^{-1}d\nu,\quad \int (f^ 2_ 1+f^ 2_ 2)d\nu =1. \] This theorem extends results by Moser for finite zone potentials to the general ergodic case.
    0 references
    maximal inequality
    0 references
    embedding into Brownian motion
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references