Abstract:
This paper describes a new projection generalized two-step
variable metric method (PTVMM) for solving minimization
problems in the Euclidean space $E^n$ in the case when
function $f(x)$ has prolate level surfaces.
The estimate of rate of convergence of the method in the case of
convex functions is presented. Finally, we indicate, how these
considered methods can be used to solving of testing optimal
control problem. Some results of comparative numerical
experiments are given.
Keywords:minimization on the simple set, projection generalized two-step
two-stage variable metric method, rate of convergence, differencial
equations of movement, optimal control problem, optimization.