Solving Polynomial Equations with Equation Constraints: the Zero-dimensional Case

Abstract: A zero-dimensional polynomial ideal may have a lot of complex zeros. But sometimes, only some of them are needed. In this paper, for a zero-dimensional ideal $I$, we study its complex zeros that locate in another variety $\textbf{V}(J)$ where $J$ is an arbitrary ideal. The main problem is that for a point in $\textbf{V}(I) \cap \textbf{V}(J)=\textbf{V}(I+J)$, its multiplicities w.r.t. $I$ and $I+J$ may be different. Therefore, we cannot get the multiplicity of this point w.r.t. $I$ by studying $I + J$. A straightforward way is that first compute the points of $\textbf{V}(I + J)$, then study their multiplicities w.r.t. $I$. But the former step is difficult to realize exactly. In this paper, we propose a natural geometric explanation of the localization of a polynomial ring corresponding to a semigroup order. Then, based on this view, using the standard basis method and the border basis method, we introduce a way to compute the complex zeros of $I$ in $\textbf{V}(J)$ with their multiplicities w.r.t. $I$. As an application, we compute the sum of Milnor numbers of the singular points on a polynomial hypersurface and work out all the singular points on the hypersurface with their Milnor numbers.