Nonconvex quadratic programming, semidefinite relaxations and randomization algorithms in information and decision systems (Q2722588)

First, some basic properties of a nonconvex quadratic optimization problem are outlined, including its complexity properties. Next, two convex relaxation schemes (based on linear and semidefinite programming, respectively) are presented. A randomization technique (based also on semidefinite programming relaxation) for solving the original problem is discussed. At the end, some applications of nonconvex quadratic programming to graph partitioning problems, to robust optimization problems and to control problems are provided.NEWLINENEWLINEFor the entire collection see [Zbl 0961.00036].