Improved permutation in sparse matrix reduction (computer-aided circuit analysis)

The problem of sparse matrix permutation, which is an essential part of the sparse matrix reduction (SMR) method, is discussed. SMR is an efficient method for reducing large sparse equation sets occurring in computer-aided circuit analysis. By permutation of the set of circuit equations, the variable vector is divided into a subvector of dependent variables and a much smaller subvector of independent variables. The permutation method is a carefully balanced trade-off between the minimal number of independent variables and the permutation execution time.< >