Constrained inverse formulations in random material design Original Research Article

The focus of this work is on the development of a computational strategy to design materials composed of randomly dispersed particulates suspended in a homogeneous binding matrix. The design objectives are to deliver prescribed macroscopic effective responses while simultaneously obeying constraints that reflect the distortion of the microscale stress fields, as well as the likelihood for fatigue damage. A nonderivative statistical genetic algorithm is developed which can handle difficulties in designing materials with random particulate microstructure due to objective function nonconvexity and lack of objective function regularity. Theoretical aspects are investigated and three-dimensional numerical examples are given.