Mixed-Integer Optimization Techniques for Algorithmic Process Synthesis

Abstract

This paper presents an overview of the mathematical programming approach for process synthesis. First, the methods for process synthesis are reviewed with an emphasis on algorithmic methods. The mathematical programming approach is covered next in a discussion of basic concepts on representations for synthesis, modeling of mixed-integer nonlinear programming (MINLP) problems, MINLP algorithms, and solution strategies. As is shown, these four components are basic elements in the algorithmic methods for process synthesis. Also, it is shown, both through the derivation of methods and their application to several examples, that MINLP optimization has reached a stage where it can solve practical problems of significant size. Finally, several future directions of research are also discussed.