Fuzziness and brand and bound procedures: Applications to separation sequencing

The aim of the paper is to identify optimal and near-optimal distillation sequences according to a particular objective function definition. The potential solutions are valuated from the point of view of practical considerations, rather than purely costing methods. Four rules of thumb addressing column sequencing, expressed in vague and imprecise terms, are implemented and quantified by the fuzzy set theory. The integer programming problem is solved by a Branch and Bound procedure, where the bounding and branching schemes are carried out from the four heuristic rules, with fuzzy quantification. Furthermore the fuzzy set theory is also implemented for modelling data uncertainties, particularly the partial molar flow rates. From the results obtained with the fuzzy branch and bound procedure involving fuzzy data, we conclude that fuzzy set theory can be considered as a promising tool for modelling uncertainty and vagueness, and may have many other applications in Process Systems Engineering. (C) 2000 Elsevier Science B.V. All rights reserved.