A Hybrid Search algorithm for multicriteria optimization and evaluation in mechatronic products design

Abstract

This study is related to the application of artificial intelligence approaches to the design of mechatronic products. The purpose is to help designers to choose design parameters according to technical specifications and customers' needs. Because of the high number of parameters and possible combinations, a heuristic based approach is investigated for optimization activity when the task is a multiple objective combinatorial one and when the criteria to be optimized cannot be weighted or ordered. A hybrid search (HS) algorithm based on some metaheuristic mechanisms is proposed for optimization without resuming the multiple objectives to a single one. The specified algorithm derives the best (not dominated) solutions set: the Pareto front, from the set of acceptable solutions defined by the design parameters, and is quite faster compared to exhaustive methods. Self-organizing Maps are then used to analyze and evaluate the obtained fronts. Our approach is fully illustrated in the case of the design of a 2-degrees of freedom robot.