A Population Replacement Strategy Analysis in Multi-objective Optimum Design of Structural Metallic Frames

Abstract

The development of evolutionary computing algorithms has allowed the resolution of new multi-objective optimum design problems in the computational mechanics field. In solid mechanics, they include the optimum design problem of structural metallic frames minimizing simultaneously the constrained weight and the number of different cross section types. It takes into account the discrete optimization with real cross-section types (database that contains normalized and ready-to-buy-in-the-market cross-sections) for real constructible designs considering constraints in terms of stresses, displacements and buckling of the bars. Here an analysis of the population replacement strategy in structural metallic frames multi-objective optimum design is performed using different population sizes in a well-known test case of the literature. Results are presented considering amplitude and approximation of the non-dominated final fronts, comparing the quality of the final designs obtained by the different proposed strategies.