An improved robust topology optimization approach using multiobjective evolutionary algorithms

• Our approach allows designers to make a-posteriori tradeoffs between objectives.
• Our robustness measures are the expected compliance and the expected variance.
• Sizing is directly incorporated into the problem formulation.
• We analyzed different sources of variability (force and material properties).
• We obtained structures that were more robust than the deterministic approaches.

Robust topology optimization has gained importance during the last years. This paper presents a robust approach to topology optimization using multiobjective evolutionary algorithms. A key contribution of our approach is that our optimization model handles structural robustness through the first two objectives, namely, the expected compliance and its variance; whereas a third objective incorporates the volume of the structure and tackles the sizing optimization problem. Finally, a major contribution of the proposed approach is that it returns a Pareto frontier showing the designer an array of possible solutions and unveiling the existing tradeoff between the different problem objectives, namely the expected compliance, variance of compliance, and volume of the structure.