Optimality of two-parameter iterative processes in the constrained gradient method (Q1264097)

Various iterative schemes of constrained minimization are currently known. It has been shown that the constrained gradient method admits a fairly arbitrary choice of both the direction of descent and the step length in each iteration. The possibility of controlling the iterative process by a proper choice of these parameters poses the problem of optimal choice of control parameters in each step of the solution of the minimization problem. In this note we use the approach of \textit{S. N. Perfilov} and \textit{R. F. Khabibullin} [Issled. Prikl. Mat. 8, 41-50 (1979)] to explore the optimality of the two-parameter iterative process for the constrained gradient method of \textit{V. G. Karmanov} [``Mathematical programming'' (Russian) (1975; Zbl 0349.90075)].