Abstract:
Continuous algorithms are considered in the shape of sets of differential embeddings which are to be used in finding the unconditional minimum of convex functions and solution of convex programming problems. Absolute asymptotic stability of the set of minimum points is investigated for continuous subgradient algorithms without assuming that these sets are limited. Convergence is studied to limited sets of minimum points for continuous algorithms of the subgradient type that are obtained by using penalty functions with a time-varying penalty factor.
Fulltext:
PDF file (1389 kB)