Genetic encoding for sound package configuration optimization

Abstract

The acoustic properties of poroelastic materials can be accurately predicted by the Biot–Allard model if the required acoustic parameters have been successfully characterized. In conjunction with the transfer matrix approach, one can use the model to analytically design an acoustic liner comprising different materials that can provide better performance or cost reduction advantages over their homogeneous counterparts. However, optimizing the layer configuration remains to a challenging task as it is a combinatorial problem and the search for the optimal configuration can be difficult or expensive if simply based on the build-and-test approach. The Generic Algorithm (GA) was identified as a better optimization method over the gradient-based methods for this task due to its nature in finding the combinatorial solutions. This approach is, to the best of our knowledge, novel in applying the GA to this design problem. The performance of the GA method is compared with the exhaustive search in maximizing the sound absorption coefficient of an acoustic liner with constraints set for its cost, thickness, and total weight. The features of the Generic Algorithm and the preliminary proof-of-concept results will be discussed and presented.