A Proportionality Principle for Partitioning Problems
Main result: Let \(\mu_ 1,...,\mu_ k\) denote nonatomic probability measures on a measurable space (S,\({\mathcal F})\), \(\Pi_ k=\Pi_ k({\mathcal F})\) the collection of all \({\mathcal F}\)-measurable partitions \(\{A_ 1\}^ k_{i=1}\) of S, and \(p_ k\) the set \(\{\underset \tilde{} p=(p_ 1,...,p_ k)\in {\mathbb{R}}^ k:\) \(p_ i\geq 0\), \(i=1,...,k\), \(\sum^{k}_{j=1}p_ j=1\}\). Furthermore \(M_{n,k}\) is introduced as the set of all real-valued \(n\times k\) matrices, where in particular \({\underset \tilde{} \mu}(\underset \tilde{} A)\) is the \(n\times k\) matrix with entries \(\mu_ i(A_ j)\), \(i=1,...,n\), \(j=1,...,k\), \(\underset \tilde{} A\) denotes the partition \(\{A_ i\}^ k_{i=1}\), and \({\underset \tilde{} \mu}=(\mu_ 1,...,\mu_ k)\). It is shown, that for any \(f: M_{n,k}\to {\mathbb{R}}\) the inequality \[ \sup \{f({\underset \tilde{} \mu}(\underset \tilde{} A)): \underset \tilde{} A\in \Pi_ k\}\geq C(f)=\sup \{f\left( \begin{matrix} \underset \tilde{} p_ 1\\ \underset \tilde{} p_ n\end{matrix} \right): \underset \tilde{} p_ 1=...=\underset \tilde{} p_ n\in {\mathcal P}_ k\} \] is valid. Furthermore, if C(f) if attained for \(\underset \tilde{} p\in {\mathcal P}_ k\), then there exists a measurable partition \(\underset \tilde{} A=\{A_ i\}^ k_{i=1}\) satisfying \(f({\underset \tilde{} \mu}(\underset \tilde{} A))=C(f).\) The proof is based on Lyapunov's theorem concerning the convexity of the range of \({\underset \tilde{} \mu}\), from which for any given \(\underset \tilde{} p\in {\mathcal P}_ k\) the existence of a partition \(\underset \tilde{} A=\{A_ i\}^ k_{i=1}\) with \({\underset \tilde{} \mu}(\underset \tilde{} A)=(\underset \tilde{} p...\underset \tilde{} p)'\) follows. Applications are given to classical and recent fair-division problems, and to statistical decision theory, mathematical physics, Banach space theory, and inequalities for continuous random variables.
- Optimal-Partitioning Inequalities for Nonatomic Probability Measures
- Optimal Partitioning of a Measurable Space
- A sharp partitioning-inequality for non-atomic probability measures based on the mass of the infimum of the measures
- Fair Division without Additivity
- A sharp nonconvexity bound for partition ranges of vector measures with atoms
- Determining a Fair Border
- Equipartitioning common domains of non-atomic measures
- How to Cut A Cake Fairly
- scientific article; zbMATH DE number 3816010 (Why is no real title available?)
- scientific article; zbMATH DE number 3097423 (Why is no real title available?)
- Liapounoff's Theorem for Nonatomic, Finitely-Additive, Bounded, Finite-Dimensional, Vector-Valued Measures
- Partitioning general probability measures
- Quelques théorèmes sur les mesures
- Relations among certain ranges of vector measures
- Sur la division pragmatique
- Sur les fonctions-vecteurs complètement additives
- On the Yao-Yao partition theorem
- Fair division of a measurable space
- A sharp partitioning-inequality for non-atomic probability measures based on the mass of the infimum of the measures
- A generalization of Lyapounov's theorem
- Partitioning inequalities in probability and statistics
- Bounds for \(\alpha\)-optimal partitioning of a measurable space based on several efficient partitions
- Interval Partitioning Arising from Size-Biased Picking
- A Lyapunov-type theorem for nonadditive vector measures
- Optimal-Partitioning Inequalities for Nonatomic Probability Measures
- A multidimensional Lyapunov type theorem
- Partitioning general probability measures
- Equipartitioning common domains of non-atomic measures
This page was built for publication: A Proportionality Principle for Partitioning Problems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3793425)