Product variety optimization simultaneously designing module combination and module attributes

This paper proposes a product variety optimization method for both module combination and module attributes of multiple products. As manufacturing competition has become restricted with high profitability and external constraints, simultaneous design of multiple products, called product variety design etc., becomes an important strategy. The systems-based optimal design paradigm is essential for rationalizing such practices, since design for product variety is more complicated than for a single product. In such a direction, we configure an optimization method for simultaneously designing both module combination and module attributes across multiple products. The optimization method hybridizes a genetic algorithm, a mixed-integer programming method with a branch-and-bound technique, and a constrained nonlinear programming method, i.e., a successive quadratic programming method. In its optimization process, the first optimizes the combinatorial pattern of module commonality and similarity among different products, the second optimizes the directions of similarity-based variety, and the third optimizes the continuous module attributes under the others. Finally it is applied to the simultaneous design problem of multiple airplanes to demonstrate its validity and effectiveness.