Optimal Design of Stiffened Composite Panel for Performance and Manufacturing Considerations

Abstract

Abstract This paper describes a design study in which a stiffened composite wing panel is configured for a combination of performance and manufacturing related requirements. The principal focus of the paper resides in demonstrating the adaptation of newly emergent soft-computing methods for a variety of sub-tasks that constitute the design process. These sub-tasks include function approximations, modeling of processes that lack a good analytical description, and design optimization in a space that consists of a mix of integer, discrete, and continuous design variables. Soft computing techniques discussed in this context include function approximations using back-propagation neural networks, modeling of the composite panel fabrication process using evolutionary fuzzy models, and the application of genetic algorithms and immune network modeling to the optimization problem.