On the stability of solutions to quadratic programming problems (Q5930733)

The paper deals with a class of parametric optimization problems of minimizing a quadratic objective function (which, in general, is not convex) subject to a finite number of linear constraints. Here, the real parameter vector consists of the matrix of the objective function as well as of all linear coefficients describing the problem under consideration. In dependence of this parameter vector, the authors discuss necessary and sufficient conditions for the lower semicontinuity of the point-to-set mappings whose images are the set of local minimizers and the set of global minimizers, respectively.