Multidisciplinary design optimization based on independent subspaces with common design variables

Designers should use efficient and adequate design methodologies for large scale engineering systems. Multidisciplinary design optimization (MDO) is one of the methodologies which can effectively treat large scale engineering systems. Various MDO methods are briefly explained and compared. Among those MDO methods, multidisciplinary design optimization based on independent subspaces (MDOIS) is a simple and practical method that can be applied to practical MDO engineering problems. However, the current version of MDOIS does not deal with common design variables. A new version of MDOIS is proposed and named as MDOIS/2006. It is a two-level MDO method while the original MDOIS is a single-level method. To validate MDOIS/2006, a mathematical problem as well as a NASA test bed problem is solved. The results are compared to those from other MDO methods. As an engineering example of MDO application, a design problem for the flow patterner is solved. The related disciplines are fluid and structural disciplines. The problem is defined from an MDO viewpoint and solved by MDOIS/2006. The results of the design are discussed.