Characterization of \(L^1\)-closed decomposable sets in \(L^\infty\) (Q1125307): Difference between revisions

Let \((\Omega,{\mathcal A},\mu)\) be a finite, complete measure space and let \(L^p(\Omega)= L^p(\Omega, \mathbb{R}^n)\), \(1\leq p\leq\infty\). A set \(Q\subset L^p(\Omega)\) is called decomposable, if \(1_Ax+ 1_{\Omega\setminus A}y\in Q\) for all \(x,y\in Q\) and \(A\in{\mathcal A}\). If \(Q: \Omega\to 2^{\mathbb{R}^m}\), then \(\sigma_p(Q)= \{x\in L^p(\Omega): x(t)\in Q(t)\) for a.e. \(t\in\Omega\}\) is always decomposable. It is proved that if \(Q\subset L^\infty(\Omega)\) is nonempty and decomposable, \(1\leq p<\infty\), then the condition \(x\in \text{cl}_pQ\) is equivalent to the following one: there exist a sequence \((A_n)\), \(A_n\in{\mathcal A}\), and a \(\xi\in Q\) such that \(\mu(A_n)\to \mu(\Omega)\) as \(n\to\infty\) and \(1_{A_n}x+ 1_{\Omega\setminus A_n}\xi\in \text{cl}_\infty(\Omega)\) for all natural \(n\). Moreover, the map \(\sigma_\infty\) is a bijection between some sets \(\Gamma(\Omega)\) and \(\Sigma(\Omega)\). Next, there are investigated convexity properties of decomposable sets, support functionals for set-valued maps, and normal and polar cones of decomposable sets in \(L^\infty(\Omega)\).