The tradeoffs between cooperative and approximate cooperative formulations in multidisciplinary design
Kemper E. Lewis · 7th AIAA/USAF/NASA/ISSMO Symposium on Multidisciplinary Analysis and Optimization · 1998
A number of approaches to multidiscipli nary design optimization have been used to model collaboration among multiple disciplines. Some of these assume a form of full collaboration where exact representations of shared information are used, while others use approximations of shared information. In this paper, the tradeoffs between using a full form of collaboration or cooperation versus using an approximate form are studied. The tradeoffs studied include solution quality, approximation accuracy, and solution stability. It is shown that for a class of MDO problems, by using approximate forms of collaboration or cooperation high quality solutions can still be found, unknown information can be well approximated, but solution convergence may be slower and at times unstable.