Abstract:

A simple method to construct the quasinormal cone is given in connection with  the nonconvex feasible domain in which the constraint functions are quadratic functions. It has been proven that selected mapping is positive linear independence with regard to constraint gradient, and selected quasinormal cone satisfies the quasinormal cone condition. Therefore, we can solve such non\|convex optimization problems with the help of combined homotopy equation.

Key words: nonconvex optimization, combined homotopy interior point method (CHIP), the quasinormal cone condition, globally method