Inverse Design with Multi-Objective Topology Optimization Based on Binary Rank-1 Perturbations

This contribution shows the extension of the state-of-the-art single-objective binary topology optimization technique to a multi-objective framework. This achievement significantly expands the applicability of the optimization routine since only an inefficient scalarization technique was available before. We utilize the adaptive weighting method, which forces locally optimal solutions to move as close to the Pareto front as possible. The method's efficacy is high, solving multi-objective problems at a computational cost comparable to that of the original algorithm, which solved single-objective problems. All the advantages and features of the original method are preserved. An example is presented in this paper to demonstrate the superb performance of the algorithm.