Reformulation of nonconvex quadratic problem


Using the KKT conditions, it is possible to reformulate nonconvex quadratic optimization problems as mixed-integer linear problems. Yalmip has a procedure / solver called KKT to do that. Has CVX a similar method that can be used for nonconvex quadrfatic problems?

(Mark L. Stone) #2

CVX has no built-in method. You’ll have to formulate it yourself as a MILP in CVX … or use another tool, such as YALMIP.