On Equivalence of Anchored and ANOVA Spaces; Lower Bounds

Peter Kritzer,Friedrich Pillichshammer,G. W. Wasilkowski
DOI: https://doi.org/10.48550/arXiv.1511.05674
2015-11-18
Abstract:We provide lower bounds for the norms of embeddings between $\boldsymbol{\gamma}$-weighted Anchored and ANOVA spaces of $s$-variate functions with mixed partial derivatives of order one bounded in $L_p$ norm ($p\in[1,\infty]$). In particular we show that the norms behave polynomially in $s$ for Finite Order Weights and Finite Diameter Weights if $p>1$, and increase faster than any polynomial in $s$ for Product Order-Dependent Weights and any $p$.
Numerical Analysis
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the equivalence between weighted anchored spaces and ANOVA spaces. More specifically, it is about the lower bound estimation of the embedding norms of these spaces. In more detail: 1. **Research Background**: - Anchored spaces and ANOVA spaces are two spaces used to represent multivariate functions, and they are widely used in areas such as numerical integration and function approximation. - ANOVA decomposition involves variance calculation and is difficult to be numerically processed in practice; while the truncated dimension of the anchored space is small and easy to utilize. 2. **Core of the Problem**: - The author focuses on the equivalence of these two types of spaces, that is, for a given weight \(\gamma\), whether the anchored space and the ANOVA space are equivalent, and how the growth rate of the embedding norm is. - Specifically, they study the behavior of the embedding norm under different weight types and provide lower bound estimates of the embedding norm. 3. **Main Objectives**: - Provide lower bound estimates of the embedding norm, especially for finite - order weights, finite - diameter weights, and product order - dependent weights. - Determine the growth rate of the embedding norm under these weight types to judge whether the anchored space and the ANOVA space are polynomially equivalent or uniformly equivalent. 4. **Specific Problem Description**: - For finite - order weights, when \(p > 1\), the embedding norm shows a polynomial growth of \(s\), that is, \(\Theta(s^{q/p^*})\), where \(q\) is the order of the weight and \(p^*\) is the conjugate index of \(p\). - For finite - diameter weights, the growth rate of the embedding norm is \(\Theta((s - q)^{1/p^*})\). - For product order - dependent weights, the growth rate of the embedding norm is faster than any polynomial. Through the study of these problems, the author aims to further understand the properties of these two types of spaces and their application potential in high - dimensional numerical analysis. ### Formula Summary - \(p^*\) is the conjugate index of \(p\), satisfying \(\frac{1}{p}+\frac{1}{p^*}=1\). - The lower bound formula of the embedding norm is: \[ \sup_{\{c_u, h_u\}_{u \in U}} \left( \sum_{v \in U} \gamma_v^{-p} \int_{D_v} \left| \sum_{w \subseteq [s]\setminus v} c_{v \cup w} \int_{D_w} h_{v \cup w}([x_v; t_w]) \prod_{j \in w} (1 - t_j) dt_w \right|^p dx_v \right)^{1/p} \] \[ \left( \sum_{u \in U} c_u^p \right)^{-1/p} \] - For finite - order weights, the asymptotic behavior of the embedding norm is \(\Theta(s^{q/p^*})\). - For finite - diameter weights, the asymptotic behavior of the embedding norm is \(\Theta((s - q)^{1/p^*})\). - For product order - dependent weights, the growth rate of the embedding norm is faster than any polynomial. Hope this information can help you better understand the core problems and research content of this paper. If you have any further questions, please feel free to ask!