Maximal inequalities and their applications to orthogonal and Hadamard matrices (Q2220996): Difference between revisions

Let \((X,\Vert \cdot \Vert)\) be a normed space, \(A = \{a_1, a_2,\dots,a_m\}\), \(1 <m < \infty\), be an arbitrary finite sequence of the vectors from the unit ball \(B_X \subset X\) and let \(G(m) = \{-1, 1\}^m\) be the set of all collections \(\vartheta =\{\vartheta_1,\vartheta_2,\dots,\vartheta_m\}\) of signs, \(\vartheta_i =\pm1, i = 1,2,\dots,m\). The main problem of the paper is to estimate the expression \( S_{X, A} \equiv \min_{\vartheta \in G(m)}\max_{1\leq k \leq m} \vert \vert \sum_{i =1}^k \vartheta_i a_i \vert \vert \). \par The authors prove some maximal inequalities for signs. Their proofs are based on probabilistic arguments. In particular, the authors estimate \[\varrho_{X,A} (\vartheta) \equiv \max_{1\leq k \leq m} \vert \vert \sum_{i =1}^k \vartheta_i a_i \vert \vert ,\] \({\vartheta \in G(m)}\) and \(\inf_{\vartheta \in G(n)} \varrho_{X,A} (\vartheta)\), \(A \in O(n)\) where \(O(n)\) is the orthogonal group.