A Genetic Algorithm integrated into product design for modular identification (ID: 6-117)

This paper presents a method to identify module from product structure on the base of information dependency. Supporting product designed by family or platform is the virtue of modularization design, which makes the modules of system have a good defined interface. Modular partition is an important step in modularization design, which directly effects on product's function, performance, and cost. Usually, the relationship among the components is the most important factor in modular partition. The modular partition has obvious objective, for different products, engineers must consider from different aspects for finding module based on concretely requirements. But in complex system, there are many interaction relations of components. it is difficult to find out the best modular partition. GA (Genetic Algorithm) has considered as one of better means to solve modular problems than other methods. In this paper, a method of Design Structure Matrix (DSM) has been adopted to identify the architecture of complex products whose decomposition is unknown in advance. Consequently, the two basic operators of GA are designed (crossover and mutation) for cluster problem and the optimization of the DSM is implemented by GA. The feature of the method is convenient to find a proper module in condition of interacting information. A testing case is carried out to show the effects of optimizing the modular partition in complex system with a genetic algorithm and satisfactory result is obtained.