Application of Alternative Multidisciplinary Optimization Formulations to a Model Problem for Static Aeroelasticity

Abstract

A new model problem for static aeroelasticity is introduced and used to illustrate several alternative approaches for formulating multidisciplinary design optimization problems. The alternatives are distinguished by the kind of analysis problem feasibility that is maintained at each optimization iteration. In the familiar "multidisciplinary feasible" approach, the full multidisciplinary analysis problem is solved at each iteration of the optimizer. At the other end of the spectrum is the "all-at-once" approach where none of the individual analysis discipline equations is guaranteed to be satisfied until optimization convergence. In between lie other possibilities that amount to enforcing feasibility of the individual analysis disciplines at each optimization iteration, but allowing the coupling between the disciplines to be incorrect until optimization convergence. Results are given for these three approaches applied to the new model. In general, delaying feasibility until optimality reduces the total amount of computing work.