Identification of Optimal Topologies for Continuum Structures Using Metaheuristics: A Comparative Study

Abstract

The development of low dimensional explicit based topology optimization approaches such as moving morphable components method increased the hopes to develop and expand evolutionary based solutions in the topology optimization of continuum structures. Despite the low dimensionality of the parametrization which helps to increase the efficiency, due to the multimodal behavior of the objective function and the correlation between the design variables more researches should be done to improve the efficiency. This paper is dedicated to comparing nine non-gradient approach based approaches based on the moving morphable parameterization. The algorithms are compared by the convergence speed, the quality of final designs, and the abilities to explore and exploit based on a diversity index. It is demonstrated that only some of these algorithms can lead to globally optimal solutions. This research clarifies the ability of the aforementioned algorithms to solve the topology optimization problem which can help future researchers to develop more suitable and efficient algorithms for this problem.