A sharp partitioning-inequality for non-atomic probability measures based on the mass of the infimum of the measures (Q1077801): Difference between revisions

If \(\mu_ 1,...,\mu_ n\) are non-atomic probability measures on the same measurable space (S,\({\mathcal F})\), then there is an \({\mathcal F}\)- measurable partition \(\{A_ i\}^ n_{i=1}\) of S such that \(\mu_ i(A_ i)\geq (n-1+m)^{-1},\) where \(m=\| \bigwedge^{n}_{i=1}\mu_ i\|\) is the total mass of the largest measure dominated by each of the \(\mu_ i's\); moreover, this bound is attained for all \(n\geq 1\) and all m in [0,1]. This result is an analogue of the bound \((n+1-M)^{-1}\) of \textit{J. Elton}, \textit{T. Hill}, and \textit{R. Kertz} [Optimal-partitioning inequalities for non-atomic probability measures. Trans. Am. Math. Soc. 296, 703-725 (1986)] based on the mass M of the supremum of the measures; each gives a quantitative generalization of a well-known cake-cutting inequality of \textit{K. Urbanik} [Fund. Math. 41, 150-162 (1954; Zbl 0056.055)] and of \textit{L. Dubins} and \textit{E. Spanier} [Am. Math. Monthly 68, 1-17 (1961; Zbl 0108.316)].