Solving the general quadratic programming problem in a finite number of steps (Q382174)

An iterative method consisting of a finite number of steps is developed for the general inequality constrained convex quadratic programming problem. For that, the problem is replaced equivalently by an unconstrained minimization problem with a differentiable convex function. The method itself combines the gradient method with an Armijo type step-size rule for this function with the computation of orthogonal projections to solutions of systems of linear equations which are related to the Karush-Kuhn-Tucker conditions of the original problem and formed by working index sets which, after a finite number of iterations, coincide with the set of active indices in a solution of the original problem.