We describe a recursive algorithm that decomposes an algebraic set into locally closed equidimensional sets, i.e. sets which each have irreducible components of the same dimension. At the core of this algorithm, we combine ideas from the theory of triangular sets, a.k.a. regular chains, with Gr\"obner bases to encode and work with locally closed algebraic sets. Equipped with this, our algorithm avoids projections of the algebraic sets that are decomposed and certain genericity assumptions frequently made when decomposing polynomial systems, such as assumptions about Noether position. This makes it produce fine decompositions on more structured systems where ensuring genericity assumptions often destroys the structure of the system at hand. Practical experiments demonstrate its efficiency compared to state-of-the-art implementations.
翻译:我们描述一种递归算法, 将一个代数组分解成本地封闭的等维数据集, 即每个组具有同一维度不可减损的组成部分。 在这个算法的核心, 我们结合了三角形组的理论, a. k. a. 普通链条, 与Gr\" obner基底, 来编码和与本地封闭的代数组一起工作。 有了这个算法, 我们的算法就避免了对代数组的预测, 这些代数组是分解的, 在解分解多元系统时经常做出某些通用假设, 比如对诺埃瑟位置的假设。 这使得它在结构更完善的系统中产生精细的分解状态, 以确保通用性假设常常摧毁手头系统的结构。 实际实验证明了它相对于最新执行过程的效率 。