A parallel global optimization algorithm for rational separable-factorable functions

Abstract

This paper concludes that we can separately globally minimize every component of a rational separable-factorable function instead of the globally minimizing the whole objective function. This will largely reduce the amount of computation, not only by reducing the size of every subproblem, but also by reducing the number of local minimizers which might be found during the global minimization process. Furthermore, parallel processors can be used in this case to shorten the computing time greatly.