抄録
The entire design activity of an engineering system is performed through the integration of distributed partial design problems. A better design solution requires their superior coordination. In general, a design problem is multi-objective, and its optima is defined as a set of Pareto optimal solutions. This means that its final solution cannot be fixed without any outside commitment. Such relationship gives design freedom to each part, and simultaneously acts as outside commitments. This paper proposes an optimization-based method for effectively supporting designers in distributed design process based on Pareto optimality. In the method, the above freedom is modeled as linear convex regions and domains represented with a set of representative points, refinement operation is mathematically formulated over such points. The optimization procedure is implemented with physical programming for efficiently reflecting designer's preference. The proposed method is applied to preliminary airplane design problem to demonstrate its validity.
