Abstract: A model of describing relationships among functional requirements and structural characteristics, design parameters and uncontrollable factors in non-linear systems is proposed. With this model, system sensitivity is analyzed , and the specific mathematic expression of system sensitivity index Sv2 is given. According to the principle of robust design, an optimization model of robust design is established. Combining this optimization model with the TAGUCHI method, an analysis is carried out to reveal the characteristics of the systems. Based on an analysis of impact of various factors on system robustness, system sensitivity matrix RS is established. Relationships among structural characteristics of system, design parameters and uncontrollable factors can be formulated as a mathematical function. The characteristics of system determine how the design parameters affect the functional requirements of system. Consequently, they affect the distribution of functional requirements. Non-linearity of system can be used to facilitate the selection of design parameters in order to achieve the required functional requirements as well as minimize the variation. When the structural characteristics and selection of design parameters cannot meet the conditions of system robustness, variation of functional requirements could be reduced from the sources by controlling the uncontrollable factors.

Key words: Non-linear system, Product design, Robust design, Sensitivity