Asymptotic analysis and the Weierstrass theorem in dual spaces (Q6958902)

The paper deals with generalizations of the Weierstrass theorem into settings where the standard coercivity conditions are not present, and the functions are defined in dual spaces endowed with the weak* topology. The authors introduce the notion of an asymptotic function tailored to functions in dual spaces which are infinite-dimensional and weak*-topologized. The properties of the asymptotic function are studied. Thus existing ideas of asymptotic behavior for convex functions are generalized to possibly nonconvex cases in dual spaces.\N\NUsing this idea the authors prove a version of the Weierstrass theorem without requiring coercivity. Instead, they formulate necessary and sufficient conditions involving the asymptotic function itself that guarantee the existence of minimizers for lower semicontinuous functions, and functions that are weakly lower continuous. These results extend earlier work where existence was tied to norm convergence conditions. For quasiconvex functions the authors further refine these conditions. They prove that for such functions, the necessary condition is also sufficient under the proposed framework of the asymptotic function. This gives a complete characterization of existence of minimizers in this broader setting. The significance of the work is in extending classical minimization theory to noncoercive, potentially nonconvex settings in dual spaces and providing a new tool (asymptotic function) that can be used to analyze existence without compactness or coercivity.