A Multiobjective Topology Optimization Method Based on Modified Hypervolume Increment

ABSTRACT The topology optimization involves time‐consuming numerical simulations; therefore, it is desired to achieve a high quality Pareto set with a limit number of optimizations when there are multiple conflicted objectives. In this paper, a high‐performance multiobjective topology optimization method is developed. The basic idea is to maximize the hypervolume, which is the only quality metric with Pareto consistency. The sparse regions of Pareto points are identified by the local potential hypervolume increment, and a modified hypervolume increment function is adopted to search the new Pareto point. The developed method is compared to several existing approaches based on five examples. The results show that the developed method performs the best, and the advantage is more obvious with more objectives.