ROBUST MULTI-OBJECTIVE OPTIMIZATION APPLIED TO ENGINEERING SYSTEMS DESIGN

Abstract

IN ENGINEERING SYSTEMS DESIGN, THEORETICAL DETERMINISTIC SOLUTIONS CAN BE HARDLY APPLIED DIRECTLY TO REAL-WORLD SCENARIOS. BASICALLY, THIS IS DUE TO MANUFACTURING LIMITATIONS AND ENVIRONMENTAL CONDITIONS UNDER WHICH THE REAL SYSTEM WILL OPERATE. THEREFORE, A SMALL VARIA-TION IN THE DESIGN VARIABLES VECTOR CAN RESULT IN A MEANINGFUL CHANGE ON THE THEORETICAL OPTIMAL DESIGN AS REPRESENTED BY THE MINIMIZATION OF THE CORRESPONDING VECTOR OF OBJECTIVE FUNCTIONS. IN THIS CONTEXT, IT IS IMPORTANT TO DEVELOP METHODOLOGIES THAT ARE ABLE TO PRODUCE SOLUTIONS (EVEN SUBOPTIMAL) THAT ARE LESS SENSITIVE TO PERTURBATIONS IN THE DESIGN VARIABLE VECTOR AND, CONSEQUENTLY, LEADING TO A ROBUST OPTIMAL DESIGN. IN THIS CONTRIBUTION, FIRST THE PROPOSED APPROACH IS TESTED ON VARIOUS MATHEMATICAL FUNCTIONS. THEN, THE METHODOLOGY IS APPLIED TO THE DESIGN OF TWO REPRESENTATIVE ENGINEERING SYSTEMS THROUGH MULTI-OBJECTIVE OPTIMIZATION USING THE FIREFLY COLONY ALGORITHM IN ASSOCIATION WITH THE EFFECTIVE MEAN CONCEPT IS PRESENTED. THE RESULTS OBTAINED DEMONSTRATE THAT THE DESIGN STRATEGY CONVEYED REPRESENTS AN INTERESTING ALTERNATIVE APPROACH TO OBTAIN ROBUST DESIGN FOR A NUMBER OF ENGINEERING APPLICATIONS.