The present paper describes an approach for the optimal sizing of rotorcraft subfloor structures to minimize structural damage and the incidence of severe injury to the occupants under crash loads. The crash analysis in the present work was modeled through the use of a discrete analysis code (KRASH), which uses lumped masses, linear and nonlinear beam elements, and nonlinear spring elements to model the primary and secondary structures. The optimal load-deflection characteristics of the subfloor components, and the placement of these components in the subfloor region were then obtained through the use of a genetic algorithm-based search procedure. Finally, an inverse design approach was used to recover the dimensions of typical energy absorbing structural components that would yield the load deflection behavior derived in the previous step. A neural network-based response surface was constructed to alleviate the stiff computational requirements of this numerical procedure.