Optimal design of structures using metaheuristics and metamodeling

Computation-intensive structural design optimization problems are common in engineering. The computation burden is often caused by expensive simulation models and becomes a problem in practice. To address such a challenge, metamodeling techniques are often chosen to improve the optimization efficiency. Metamodels, or surrogate models, are very popular methods and are considered a valuable tool to support a wide scope of activities in modern engineering design. The paper discusses strategies to obtain suitable metamodels to assess their quality concerning prediction, considering machine learning techniques used in the literature. A multi-objective structural optimization problem is used to illustrate the applicability of the approach, considering as objective functions the structure’s weight, and critical load factor concerning the structure’s global stability. The numerical results demonstrate the efficiency and computational advantages of the proposed methodology.