An efficient design strategy for solving MDO problems in non-cooperative environments
Anoop R. Nair, Kemper E. Lewis · 8th Symposium on Multidisciplinary Analysis and Optimization · 2000
Multidisciplinary Design Optimization (MDO) problems are optimization problems that describe complex coupled engineering systems. In terms of design, coupling refers to a dependence of one subsystem on some or all of the remaining subsystems. Coupled subsystems generally have one or more design variables in common and due to the inherent coupling between subsystems found in nonhierarchical systems, the design of such systems requires a great deal of coordination, communication, and information exchange between the coupled subsystems. This paper presents a design strategy for solving these types of problems in a non-cooperative environment, where subsystems cannot communicate with each other for a variety of reasons. Limitations of traditional MDO design strategies when applied in a non-cooperative environment as described above are presented. A new “Non-Cooperative” Design Strategy based on Data Fusion is developed to solve MDO problems in non-cooperative environments and the proposed method is demonstrated using two case studies.