Improving collaborative optimization for MDO problems with multi-objective subsystems.

This paper investigates collaborative optimization (CO) for multidisciplinary design optimization problems with multi-objective subsystems. A multi-stage optimization strategy is presented in the system level to enhance the performance of CO that may be weakened by the initial point in some cases. Resolving the multi-objective problem in subsystems with the preference-based algorithm requires setting the preference values for all the objective functions. Therefore, we transform the incompatibility function into a disciplinary constraint aiming to avoid providing the preference value, and propose two transformation methods, namely, closest distance and relaxed distance. At the end of the paper, the results of two engineering examples demonstrate the performance of the multi-stage optimization strategy, and the effectiveness of the transformation for the incompatibility function. [ABSTRACT FROM AUTHOR]