A recent tractable gradient-based approach that exploits a new two-point topological derivative is extended here to encompass multifidelity modeling. Handling multidisciplinary responses during structural topology optimization is important to a field that has grown rapidly in the last decade, spurred by additive manufacturing. The cellular division method for topological design is more general than conventional optimality criterion methods used most often in industrial structural topology design; however, it is a discrete approach that is daunting without topological derivatives, due to the curse of dimensionality. Results are presented to address that issue for benchmark test cases using the multifidelity approach with two-point topological derivatives to drive cellular development.
Kobayashi et al. (Mon,) studied this question.