The Surrogate Blend Optimizer employs objective functions describing the differences between computed and specified values of properties. Minimization occurs in a two-step process. Initial minimization uses a Genetic algorithm. This is due to the high number of iterations required by the Genetic algorithm to approximately locate a region of the global minimum. Secondly, the Powell minimization converges more rapidly and accurately to the minimum indicated by the Genetic algorithm.

The first step is the global minimization step, which is based on a Genetic algorithm. This method samples all regions of the mole fraction space and gradually settles towards lower function values. The second step is a local minimization using the Direction Set method that is based on Powell’s conjugate direction method.