Finding the optimum sequence of distillation columns using the A* search procedure

The synthesis of separation sequences is a well known problem in the process synthesis domain. A number of solution techniques has been proposed. These may be broadly classified into two general categories, namely discovery methods based upon the application of heuristics and algorithmic methods. The former procedures can produce conflicting results whilst the latter require considerable mathematical expertise. It is possible to identify the optimal sequence by exhaustive search but complete enumeration and evaluation of all alternatives is a daunting task for problems involving more than a few components. In this study, an alternative approach is proposed based upon an efficient search technique developed in the AI (Artificial Intelligence) field and known as A* search. It dramatically reduces the search space and can guarantee a globally optimal solution. The results are presented for the identification of an optimal configuration for multicomponent distillation sequences assuming sharp separations.