Subgradient projection algorithm. II (Q797969): Difference between revisions

The method presented by the author in Part I [ibid. 35, 111-126 (1982; Zbl 0486.65042)] for the minimization of certain nondifferentiable functions constrained to a convex polytope is extended to the minimization of a strictly convex piecewise smooth function subject to smooth convex constraints. In a sense, it is a subgradient counterpart of the method of \textit{J. B. Rosen} [J. Soc. industr. appl. Math. 9 (1961), 514-532 (1962; Zbl 0231.90048)]. The method uses only easily computable portions of certain \(\epsilon\) -subdifferentials and it avoids jamming. A convergence proof is given.