Necessary Optimality Conditions for Control Problems and the Stone--Cech Compactification
There are diverse situations in infinite dimensional control theory where, to obtain necessary optimality conditions (for instance of the type of Pontryagin's maximum principle), one is forced to consider dual spaces of nonseparable spaces. An example are the parabolic equations in \(L^\infty(\Omega)\) spaces. Here, the ``final condition for the costate is a sort of distribution based not on ordinary \(\sigma\)-additive measures but on elements of the dual of \(L^\infty(\Omega)\) (finitely additive measures). Another instance is considered by the authors in this paper; if one solves a parabolic equation in a domain \(\Omega\) in a time interval \([0, T]\) with an initial condition in \(L^\infty(\Omega)\) (rather than in \(C(\overline \Omega)),\) the solution belongs not to \(C([0, T] \times \overline \Omega)\) but to the space \(BC((0, T] \times \overline \Omega)\) of bounded continuous functions in the noncompact set \((0, T] \times \overline \Omega.\) If one deals with pointwise state constraints, there is a choice; either consider directly the dual of \(BC((0, T] \times \overline \Omega)\) (consisting of finitely additive measures) or use instead the dual of \(C(((0, T] \times \overline \Omega)^\#),\) where \(^\#\) indicates the Stone-Čech compactification; this dual consists of \(\sigma\)-additive measures. Either choice entails a certain ``weirdness (as happens with things associated with the full axiom of choice), but this seems to be a built-in feature of the problem. The authors illustrate the use of the Stone-Čech compactification for a parabolic distributed parameter system with pointwise state constraints, and announce a treatment of more general problems in forthcoming papers.
- Optimality conditions for state-constrained Dirichlet boundary control problems
- scientific article; zbMATH DE number 1500628 (Why is no real title available?)
- scientific article; zbMATH DE number 1759161 (Why is no real title available?)
- The maximum principle for control systems described by linear parabolic equations
- Minimax Dirichlet boundary control problem with state constraints
This page was built for publication: Necessary Optimality Conditions for Control Problems and the Stone--Cech Compactification
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4253021)