A variant of Atanassov's method for \((t, s)\)-sequences and \((t, \mathbf{e}, s)\)-sequences (Q2509950): Difference between revisions

This article deals with upper star discrepancy bounds for \((t,s)\)-sequences and a generalization of these, namely so-called \((t, \mathbf{e}, s)\)-sequences. The authors use an exact counting argument and a worst-case analysis of signed digits in a method originally proposed by \textit{E. I. Atanassov} [Math. Balk., New Ser. 18, No. 1--2, 15--32 (2004; Zbl 1088.11058)] for showing their bounds. Using this method, the authors improve on several existing discrepancy bounds that have been shown during the past years. The upper bound for the star discrepancy of the first \(N\) points of \((t,s)\)-sequences yields values that are better than previous results in the case of moderate to large dimensions for realistic values of \(N\). The second main result in the paper provides new estimates for the star discrepancy of \((t, \mathbf{e}, s)\)-sequences which improve on the asymptotic behavior of earlier upper bounds.