Bio-Inspired Clustering of Complex Products Structure based on DSM

Abstract—Clustering plays an important role in the decomposition of complex products structure. Different clustering algorithms may achieve different effects of the decomposition. This paper aims to proposes a bio-inspired genetic algorithm that is implemented based on its reliable fitness function and design structure matrix (DSM) for clustering analysis of complex products. This new bio-inspired genetic algorithm captures the features of DSM, which is base on the biological evolution theory. Examples of these products include motorcycle engines that are presented for clustering. The five cluster alternatives are obtained from the regular clustering algorithm and the bio-inspired genetic algorithm, while the best cluster alternative comes from the bio-inspired genetic algorithm. The results show that this algorithm is well adaptable, especially when the product elements have complicated and asymmetric connections. Keywords—Bio-inspired computation; genetic algorithm (GA); reliable fitness function; DSM; complex products; clustering I