论文部分内容阅读
本文讨论稀疏矩阵技术中选主元的方法。在对四种有代表性的选主元方案进行综述之后,作者提出两点想法,并证明了两个定理。在此基础上,作者提出了一种新的选主元方案——非零元素最少列——填元数最少的方案。本方案的优点在于改进了效果(填元数和长运算数较少)和提高了予处理速度。从折衷的观点来看,这种新方案比其他四种方案更好。文中给出的三个例子证实了上述结论。
This article discusses the method of selecting principal elements in sparse matrix techniques. After reviewing four representative electoral options, the author proposes two ideas and proves two theorems. On this basis, the author proposes a new scheme of selecting principal components - the least non-zero elements - the least number of elements. The advantage of this scheme is that it improves the effect (fewer elements and long operands) and increases the processing speed. From a compromise point of view, this new solution is better than the other four solutions. The three examples given in the paper confirm the above conclusion.