A generalized successive overrelaxation method for least squares problems (Q1267026)

A new iterative method (called generalized successive overrelaxation) is given for solving large sparse least squares problems and computing the minimum norm solution to underdetermined consistent linear systems. The method is convergent if the matrix has full column rank. The method involves a matrix \(P\) as a preconditioner and a parameter \(\rho\) as an accelerator parameter. By suitable choosing of matrix \(P\) parallel computations are possible.