Journal of Complexity ( IF 1.7 ) Pub Date : 2019-10-08 , DOI: 10.1016/j.jco.2019.101439 J. Dingess , G.W. Wasilkowski
The paper deals with linear problems defined on -weighted Hilbert spaces of functions with infinitely many variables. The spaces are endowed with zero-mean Gaussian measures which allows to define and study -truncation and -superposition dimensions in the average case and probabilistic settings. Roughly speaking, these -dimensions quantify the smallest number of variables that allow to approximate the -variate functions by special ones that depend on at most -variables with the average error bounded by . In the probabilistic setting, given , we want the error with probability . We show that the -dimensions are surprisingly small which, for anchored spaces, leads to very efficient algorithms, including the Multivariate Decomposition Methods.
中文翻译:
-的平均和概率设置中的叠加和截断维 变量线性问题
本文讨论了定义在 具有无限多个变量的函数的加权希尔伯特空间。这些空间具有零均值高斯度量,可以用来定义和研究-截断和 -在平均情况和概率设置中的叠加维度。粗略地说,这些尺寸量化最小数量 变量的近似值 -最多依赖的特殊函数的变量函数 -平均误差范围为-的变量 。在概率环境中,给定,我们想要的错误 很有可能 。我们表明尺寸小得令人惊讶,对于锚定空间,这导致了非常有效的算法,包括多元分解方法。