Linear complexity algorithms for semiseparable matrices (Q1075738): Difference between revisions

For the solution of linear systems with coefficient matrix \(R=D+S\), where S is a semiseparable matrix and D a diagonal matrix, recursive algorithms are introduced that use the LDU factorization of R into factors \((L+S_ L)D(I+S_ U)\), where the semiseparability property is retained. The complexity for these algorithms is analyzed and an efficient updating method for the increase of system dimensions is proposed.