Cooperative coevolutionary topology optimization using moving morphable components

This study aims to complement well-established methods based on gradients, such as standard level-set optimization or density-based methods, and extend the field where numerical optimization of structural topologies can be employed. This is, for example, of interest for design problems related to crashworthiness design. Because of the noisy or discontinuous character of the responses, in these cases it is very hard or nearly impossible to find or approximate meaningful gradient information. Hence, an evolutionary method is attractive because it does not need gradients; however, to ensure numerical efficiency, a low-dimensional explicit representation of the geometric parameters is required, which is available via the moving morphable components (MMC) approach. The aim of this article is to present a new strategy to improve the efficiency and convergence speed of evolutionary topology optimization based on MMC. To do this, an evolutionary algorithm based on coevolution and cooperation is proposed using a divide-and-conquer strategy.

Automated summary

This study introduces a new evolutionary topology optimization strategy to improve efficiency and convergence speed based on the MMC approach, utilizing a divide-and-conquer strategy with coevolution and cooperation.