Two-weighted estimates for positive operators and Doob maximal operators on filtered measure spaces

Wei Chen, Chunxiang Zhu, Yahui Zuo, Yong JiaoJournal of the Mathematical Society of Japan 2019 · DOI

Abstract

This research characterizes strong-type and weak-type inequalities with two weights for positive operators within the framework of filtered measure spaces. These estimates serve as probabilistic analogues to the two-weight inequalities for positive operators associated with dyadic cubes in $\mathbb{R}^n$, as established by Lacey, Sawyer, and Uriarte-Tuero. Additionally, the paper establishes several mixed bounds for the Doob maximal operator on filtered measure spaces, including Hytönen–Pérez type and Lerner–Moen type norm estimates. The primary methodology relies on the construction of "principal sets."

Introduction

In harmonic analysis, weighted inequalities are essential for understanding the stability and behavior of operators when the underlying measure is distorted by a weight function. A central challenge in this field is the "two-weight problem": determining the precise conditions under which an operator is bounded between two different weighted spaces. While these conditions are well-defined for dyadic operators in Euclidean space ($\mathbb{R}^n$), extending these results to the more abstract setting of filtered measure spaces—the foundation of martingale theory—has historically been difficult.

Filtered measure spaces lack the rigid geometric structure of Euclidean space, such as the predictable nesting of dyadic cubes. This paper bridges the gap between classical dyadic harmonic analysis and probabilistic martingale theory by providing a complete characterization of two-weight inequalities for positive operators and Doob maximal operators.

Methodology: The Construction of Principal Sets

The authors' approach centers on a technical tool known as the Construction of Principal Sets. This construction serves as a probabilistic analogue to the "sparse families" used in Euclidean harmonic analysis, which have revolutionized the field over the last decade.

Key components of this methodology include:

  • Conditional Sparsity: The authors identify a property (labeled P.3) that allows for refined control over operators by decomposing them into manageable components associated with specific levels of a filtration.
  • Sawyer-type Testing: The boundedness of an operator is proven to be equivalent to "testing" the operator against the characteristic functions of sets within the filtration.
  • Mixed Norm Estimates: For the Doob maximal operator, the authors derive bounds that involve different weight constants (such as $A_p$ and $A_\infty$). These Hytönen–Pérez and Lerner–Moen type estimates provide a more nuanced view of operator behavior than traditional single-constant estimates.

Key Results

Strong and Weak Type Inequalities

The paper provides the first characterization of two-weight strong-type inequalities for positive operators on filtered spaces (Theorem 1.1). The result reveals a fundamental symmetry: an operator is bounded if and only if two specific testing conditions hold—one for the weight and one for its dual weight.

For weak-type inequalities (Theorem 1.6), the authors demonstrate that the requirements are less stringent, requiring only one of the two testing conditions necessary for the strong type.

Sharp Maximal Bounds

Theorem 1.7 establishes sharp norm estimates for the Doob maximal operator. The authors show that if a weight pair $(v, \omega)$ belongs to the $A_p$ class, the operator norm is bounded by a factor involving the $A_p$ and $A_\infty$ constants. These results improve upon existing literature by providing more precise constants and integrating mixed weight theory into the martingale context.

Discussion and Limitations

One significant technical hurdle addressed in the paper is the potential non-measurability of sets defined by weight ratios in filtered spaces. To overcome this, the authors assume weights satisfy an $A_1$ condition, ensuring they do not fluctuate too wildly relative to the filtration. However, it remains an open question whether this $A_1$ assumption is a fundamental requirement or a technical artifact of the current proof method.

Furthermore, while the paper successfully generalizes "local" characterizations, a "global" characterization that does not rely on testing individual sets remains a challenge. Finally, as the current results focus on positive operators, extending these sharp estimates to non-positive operators, such as martingale transforms (the probabilistic version of singular integrals), represents a natural next step for future research.

Summary updated

Abstract

We characterize strong type and weak type inequalities with two weights for positive operators on filtered measure spaces. These estimates are probabilistic analogues of two-weight inequalities for positive operators associated to the dyadic cubes in $\mathbb R^n$ due to Lacey, Sawyer and Uriarte-Tuero [30]. Several mixed bounds for the Doob maximal operator on filtered measure spaces are also obtained. In fact, Hytönen–Pérez type and Lerner–Moen type norm estimates for Doob maximal operator are established. Our approaches are mainly based on the construction of principal sets.
BibTeX
@article{chen2019two,
  author = {Wei Chen and Chunxiang Zhu and Yahui Zuo and Yong Jiao},
  title = {{Two-weighted estimates for positive operators and Doob maximal operators on filtered measure spaces}},
  journal = {Journal of the Mathematical Society of Japan},
  year = {2019},
  doi = {10.2969/jmsj/80058005},
}
Figure