Random moment problems under constraints (Q2179594): Difference between revisions

The authors explore the behavior of random moment sequences when certain moments are fixed by a constraint. Let \(\mathcal{P}\left(E\right) \) be the set of all Borel probability measures on a Borel set \(E \subset \mathbb{R}\) possessing moments of all order. Denote by \(m_{j}\left(\mu \right) := \int x^{j} d\mu\left( x\right) \) the \(j\)th moment of a measure \(\mu \in \mathcal{P}\left(E\right) \). A constrain \(\mathcal{C}\) is a finite collection of integer indices \( 1\le i_{1}<\cdots<i_{k}, k \in \mathbb{N} \) and corresponding values \(c_{i_{1}},\ldots,c_{i_{k}} \in \mathbb{R}\). For the constrain \(\mathcal{C}\), define by \[ \mathcal{P}^{\mathcal{C}}\left( E\right) :=\left\lbrace \mu \in \mathcal{P}\left(E \right) \,|\, m_{i_{j}}\left( \mu\right) =c_{i_{j}} \, \forall 1 \le j \le k \right\rbrace \] the space of probability measures on \(E\) fulfilling constrain \(\mathcal{C}\) and by \[ \mathcal{M}^{\mathcal{C}}_{n}\left(E \right) := \left\lbrace\left( m_{1}\left(\mu \right),\ldots,m_{n} \left( \mu\right) \right) \,|\, \mu \in \mathcal{P}^{\mathcal{C}}\left(E \right) \right\rbrace \] the constrained \(n\)th moment space. The set \( \mathcal{M}^{\mathcal{C}}_{n}\left(E \right) \) is a section of \(n\)th moment space \( \mathcal{M}_{n} \left(E \right) := \left\lbrace\left( m_{1}\left(\mu \right),\ldots,m_{n} \left( \mu \right) \right) \,|\, \mu \in \mathcal{P} \left(E \right) \right\rbrace\). The authors equip \( \mathcal{M}^{\mathcal{C}}_{n}\left(E \right) \) with the uniform or more general probability distribution.