STRATEGIES FOR EQUILIBRIUM-STAGE SEPARATION CALCULATIONS ON PARALLEL COMPUTERS

When multicomponent, multistage separation problems are solved on para llel computers by successive linearization methods, the solution of a large sparse linear equation system becomes a computational bottleneck , since other parts of the calculation are more easily parallelized. W hen the standard problem formulation is used, this system has a block- tridiagonal form. It is shown how this structure can be used in parall elizing the sparse matrix computation. By reformulating the problem so that it has a bordered-block-bidiagonal superstructure, it can be mad e even more amenable to parallelization. These strategies permit the u se of a two-level hierarchy of parallelism that provides substantial i mprovements in computational performance on parallel machines.