Selection of a Mixed-Integer Nonlinear Programming (MINLP) Model of Distillation Column Synthesis by Case-Based Reasoning

The paper presents a new application of the case-based reasoning method for finding a mixed-integer nonlinear programming (MINLP) model with superstructure and a solution of the corresponding distillation synthesis problem by suggesting an initial point for performing design and optimization of the system. A case library has been built from earlier published distillation problems with reproducible MINLP models. When solving a new problem, the most similar case to the target is found in the case library during the retrieval process in two steps: (i) first, a set of matching cases is retrieved, using inductive retrieval; (ii) the cases in the retrieved set then are ranked according to their similarity to the target case, using the nearest-neighborhood method.