SDP-quality bounds via convex quadratic relaxations for global optimization of mixed-integer quadratic programs

We consider the global optimization of nonconvex mixed-integer quadratic programs with linear equality constraints. In particular, we present a new class convex relaxations which are derived via cuts. To construct these cuts, solve separation problem involving matrix inequality special structure that allows use specialized solution algorithms. Our cuts nonconvex, but define feasible set when intersected show our an outer-approximation semi-infinite program under certain conditions is equivalent to well-known semidefinite relaxation. The implemented in solver BARON, and tested by conducting numerical experiments on large collection problems. Results demonstrate that, for test problems, lead significant improvement performance BARON.