scientific article; zbMATH DE number 1779291

From MaRDI portal
Publication:4544328





Let \((\Omega,{\mathcal F},\mu)\) be a measure space with \(\mu\) non-negative and \(\sigma\)-finite, and let \({\mathcal B}_n\) denote the Borel \(\sigma\)-field in \({\mathbb R}^n\). The paper attacks the problem of the characterization of the mappings \(F\) that can be described in the supremal form \[ F(u,B)=\mu-\sup\{f(x,u(x)) : x\in B\} \tag \(*\) \] for every \((u,B)\in L^\infty_\mu(\Omega;{\mathbb R}^n)\times{\mathcal F}\), for some \({\mathcal F}\times{\mathcal B}_n\)-measurable function \(f\colon\Omega\times{\mathbb R}^n\to]-\infty,+\infty]\) such that \(f(x,\cdot)\) is lower semicontinuous for \(\mu\)-a.e. \(x\in\Omega\). It is proved that locality, supremality, and lower semicontinuity conditions of the following type \[ F(u,A)=F(u,B)\text{ for all }u,v\in L^\infty_\mu(\Omega;{\mathbb R}^n) \] and \[ A, B\in{\mathcal F}\text{ with }u=v\;\mu\text{ -a.e. on }B \text{ and }\mu(A\Delta B)=0, \] \[ F\Biggl(u,\bigcup_{i=1}^\infty A_i\Biggr)=\bigvee_{i=1}^\infty F(u,A_i)\text{ for all }u\in L^\infty_\mu(\Omega;{\mathbb R}^n)\text{ and }A_i\in{\mathcal F}, \] \[ \text{for every }B\in{\mathcal F}\text{ the mapping }F(\cdot,B)\text{ is strongly lower semicontinuous in }L^\infty_\mu(\Omega;{\mathbb R}^n), \] imply the existence of a unique \({\mathcal F}\times{\mathcal B}_n\)-measurable function \(f\colon\Omega\times{\mathbb R}^n\to]-\infty,+\infty]\) such that \(f(x,\cdot)\) is lower semicontinuous for \(\mu\)-a.e. \(x\in\Omega\) for which (\(*\)) holds. A counterexample is provided proving that the representation result does not hold anymore if the lower semicontinuity condition is dropped. An analogous representation result is proved by strengthening the lower semicontinuity assumption as for every \(B\in{\mathcal F}\) the mapping \( F(\cdot,B)\) is sequentially weakly* lower semicontinuous in \(L^\infty_\mu(\Omega;{\mathbb R}^n).\) In this case, it is also proved that for \(\mu\)-a.e. \(x\in\Omega\) and for every \(t\in{\mathbb R}\) the level set \(\{s\in{\mathbb R}^n : f(x,s)\leq t\}\) is convex, provided \(\mu\) is nonatomic. The necessity of the weak* lower semicontinuity of a functional, as in \((*)\) provided the above level set convexity of \(f\) is fulfilled, is also proved. Finally, an example proves that the level set convexity of \(f\) does not hold if \(\mu\) has some atoms.




Cited in
(23)








This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4544328)