This research is related to the multidisciplinary design of complex systems. We focus on the optimization activity of the embodiment design phase according to "System Engineering" formalism. In such an activity, developers may encounter some difficulties mainly due to a coupling between the disciplines and a multiplicity of the design objectives. To provide designers with a methodology and tools to help them when choosing the product architecture and components, we are interested in Multidisciplinary Design Optimization (MDO) methods as well as in multiobjective optimization. Our contribution aimed at reducing the duration of the optimization activity and has two main issues. On the one hand, two research methods based on meta-heuristics can quickly determine an efficient solutions set. On the other hand, the application of discrimination methods provides designers with a help tool for analyzing the set of computed efficient solutions. Our proposals are illustrated by two examples in robotics.
