A parallel algorithm for the global minimizationof Gibbs free energy

A chemical mixture under conditions of constant temperature and pressure may split intodifferent phases. The number of phases and the composition of each may be determined byglobally minimizing the Gibbs free energy of the system. This can be done by iteratingbetween an easy local minimization problem with a high number of variables and a difficultglobal search and verification problem in a small number of variables. The global problemcan be solved by a branch and bound method, using bounds from interval analysis. Whenimplemented in parallel, the method has lower communication requirements than otherrelated branch and bound approaches for general global minimization. We present a parallelimplementation on a network cluster of workstations that exploits this characteristic. Ondifficult instances, utilizations of over 90% are obtained using up to 14 processors. Thealgorithm copes well with varying workstation loads and has low communication overheads.A method of assessing the performance of a parallel algorithm on a shared heterogeneousnetwork of workstations is developed.