Mesh dependency of stress-based crossover for structural topology optimization

This paper presents a genetic algorithm (GA) with a stress-based crossover (SX) operator to obtain a solution without checkerboard patterns for multi-constrained topology optimization problems. SX is based on the element stress. On one hand, smaller mesh size is required to improve the accuracy of structure analysis results. On the other hand, the computation cost of genetic algorithms for structural topology optimization problems (STOPs) increases with a more detailed mesh. Therefore, it is necessary to discuss the mesh dependency of SX for STOPs. Here, the mesh dependency of SX has been investigated through experiments with four different sized meshes. Furthermore, a comparison of evolutionary structural optimization (ESO) and SX is discussed.