Global optimality conditions in maximizing a convex quadratic function under convex quadratic constraints (Q1850844)

The article characterizes globally optimal solutions for the problem of maximizing a convex quadratic objective function under several convex quadratic constraints. The characterization is given in terms of a primal optimality condition which can be seen as a certain mixture of first and second order conditions. The main idea for the proof is to specialize a more general global optimality condition for convex maximization in terms of \(\epsilon\)-subgradients -- due to the same author -- to the quadratic programming case.