SOLVING POLYNOMIAL PROGRAMS VIA CONVEX QUADRATIC REFORMULATION WITH ITERATIVE FACTORIZATION METHOD

The paper provides a mechanism for reformulating polynomial program using quadratic equations, achieved by an iterative factorization process. Our approach involves reformulating a polynomial programme using quadratic equations. This is done by taking into account the precise range of auxiliary variables and lowering their dimension. This method uses iterative factorization on monomials with a degree greater than two and the widest range to make a quadratic programme with the smallest range of auxiliary variables. Furthermore, the quadratic program is convex with an α-under estimator to obtain a convex quadratic program and solve it. The iterative factorization method lets you change the range of auxiliary variables, make the dimensions of auxiliary variables smaller, and keep adding new constraint functions. Therefore, the convex relaxation of the polynomial programme will be more powerful.