A novel approach for neuro-fuzzy system-based multi-objective optimization to capture inherent fuzziness in engineering processes

Abstract

We deal with multi-objective optimization problems in various fields and in some of them, the objectives are found to be conflicting in nature. We obtain multiple optimal or near-optimal solutions of the problem using a multi-objective evolutionary algorithm (MOEA). In this study, an approach is proposed for enhancing the use of MOEA to establish important input–output relationships of some manufacturing processes. In the proposed approach, after getting an initial set of Pareto-front data points through MOEA, the trade-off solutions are used to train a neuro-fuzzy system (NFS) utilizing an EOA. This trained NFS is then used to get a modified Pareto-front and the modified trade-off solutions are clustered using different clustering algorithms. These clustered solutions are then analyzed to establish the relationships among decision variables and objectives. These principles will surely enrich the knowledge of designers and inspire them to apply this approach for a broad range of practical problems. The data related to two different engineering problems are used to show the applicability of the proposed approach.