A flexible clustering approach for virtual cell formation considering real-life production factors using Kohonen self-organising map

In dynamic production environment, quick adaptation for new product design is an important issue to achieve competitive edge. To address this issue, virtual cellular manufacturing has been evolved to improve shop flexibility and setup efficiency of cellular manufacturing system. The virtual manufacturing cell (VMC) involves grouping of parts to form part families and grouping of machines to form machine cells without physical boundaries of cells. Kohonen's self-organising map (KSOM), an unsupervised neural network technique, is employed in this work for cell formation because of its flexibility in dynamic clustering of parts and machines. The proposed KSOM network approach considers real life production factors like processing time, operation sequence, routing flexibility, machine capacity, machine flexibility and demand to design machine cells. A numerical example from literature is used to illustrate the proposed methodology. The methodology is tested on benchmark problems of different size and performance measures such as group technology efficiency (GTE) and exceptional elements (EE) are evaluated. The results indicate that the proposed approach is quite capable of solving different sizes of problems and outperforms the existing methods in some cases.