It is well-known that the optimal beamforming problems for cognitive multicast transmission are indefinite quadratic (nonconvex) optimization programs. The conventional approach is to reformulate them as convex semi-definite ...