Study of Algorithm and Application of Approximate Groebner Basis

点击翻译按钮获取中文摘要

Based on a proposed "approximate ideal", we constructed a theory of approximate Groebner basis, clarified the instability of Buchberger's algorithm on floating point numbbers using a developed subresultnat-like theory, and proposed an algorithm of approximate Groebner basis by stabilizing Buchberger's algorithm.We also proposed a concept of "approximate singular system" as a multivariate polynomial ideal whose dimension is decreased by a purturbation, and presented an algorithm which recovers the dimension. Applying this operation to algebraic systems of approximately singular type, we proposed a well-conditioning method for such systems. Furthermore, we proposed an error suppressing method and a characteristics extracting method for solving parametric sparse linear systems.