Quadratic optimization of fixed points for a family of nonexpansive mappings in Hilbert space (Q2388418): Difference between revisions

The paper under review treats algorithms for finding a minimum of a quadratic operator over the intersection of finitely many operators in Hilbert space. These operators are assumed to be nonexpansive self-maps of a closed convex subset. The operator defining the second-degree part \(A\) in the cost function is assumed to be selfadjoint and strongly positive with \(I - A\) being less than 1 in norm. An iteration procedure is defined where the new iterates are in an affinely linear way defined by the previous iterates. Under boundedness, unboundedness and bounded variation type conditions on the parameters of the iteration procedure, a convergence theorem and a corresponding proof of the unique minimizer theorem are carefully provided.