A SYNTHESIS SPARSE MATRIX ALGORITHM FOR COMPUTER AIDED CIRCUIT DESIGN
Senquan Yang · Chinese Journal of Computers · 1984
A Synthesis Sparse Matrix Algorithm (SSMA) is described in this paper. In this algorithm, the pivots are so selected that the numerical stability of the elimination process and the sparsity of the coefficient matrix are maintained, and the factorizations of L U are adjusted. SSMA is suitable for solving high order sparse linearized algebra equations in computer aided circuit design.A study of SSMA shows that for most practical circuits the precision is close to that of the complete pivot elimination method, the amount of fillings is less than 1.8n, and that of long operations is from 4n to 15n, where n is the order of the matrix. Moneover, the memory capacity and the CPU times are drectly pre-portional to the total amount of nonzeros in the coefficient matrix.