A routine for solving sparse systems in structural analysis

Abstract A computer subprogram for the direct in-core solution of large symmetric systems of linear algebraic equations is presented. The procedure employs a sparse matrix technique and allows for symmetrical pivoting and reordering to improve numerical stability and reduce core allocation. The work is especially oriented to problems deriving from minimum or stationary principles in structural analysis. A FORTRAN print-out of the complete package is given in the Appendix.