Determining New System Design Requirements to Optimize Fleet Level Metrics Under Uncertainty

Abstract

Traditional approaches to design and optimize a new system often do not consider how the operator will use this new system alongside the other existing systems. This handoff between the designs of the new system and how this new system operates with the group of systems leads to the sub-optimal performance of the new system when measured with respect to system-level objective. Aircraft design choices made to meet a set of requirements dictate the performance of the aircraft, and the aircraft performance influences how the operator might use the aircraft. Further, the presence of uncertainties in predictions of the new aircraft performance and costs and uncertainties in the amount of payload to transport further exacerbate the problem of determining these requirements. Recent efforts have posed approaches to address this problem, but generally with a deterministic perspective. This research adopts a previously developed subspace decomposition approach and integrates features from robust/reliability based optimization to address the uncertainties and solves two application problems a military and a commercial airline application. The result demonstrates the ability of the framework to identify the design requirements for the new aircraft, and a posterior analysis indicates that the framework acceptably handles the uncertainties.