Abstract:
This article formulates and proves several useful correlations between support functions of convex sets and projection operations over them, such as asymptotic equivalence of projection operations and computation of support functions for general convex closed bounded sets, as well as equivalence between least-norm and regularized convex suplinear optimization problems. These results generalize previously known equivalences for linear optimization problems and provide new and greatly simplified proofs for them. Illustr. 1, bibliogr. 10.
Keywords:convex optimization, regularization, projection, support function.
Fulltext:
PDF file (295 kB)