Optimal design of a lightweight sandwich panel for mechanical and thermal properties

The optimization of a lightweight sandwich panel is explored using a genetic algorithm and the Pareto dominance in order to satisfy specific mechanical and thermal properties for transport applications. Genetic algorithms are particularly suited for problems with both continuous and discrete variables as for the design of sandwich panels that can be described as the combination of a geometrical problem (defining the thickness of each layer) and a material selection problem. We will assess how the Pareto-set optimization enables us to find a compromise between different performances in order to develop multifunctional properties. The large number of materials that can be considered, the number of design variables (twice the number of layers) and the nature of the algorithm can hamper the convergence. Therefore, a space reduction method using material selection techniques is employed to fasten the convergence.