Weighted directed graph based part clustering model

To solve problems of planning and reconfiguration for e-manufacturing cells, the parts clustering problems of Product Family (PF) were studied. Through analyzing some key elements in the production of PF parts, i. e. process sequence, preparation time of production and internal transportation cost etc., a part process description model based on Weighted Directed Graph (WDG) was presented. More importantly, according to the criterion of lowest fluctuation of manufacturing cost, a part similarity distance model based on a method of two-grade combination of incidence matrixes was put forward. Referring to these mathematical models, a clustering system for manufacturing of PF parts based on ant colony algorithm was designed and developed. Finally, a software system was developed using Java technology. Analysis of clustering case has demonstrated that this system could rationally and effectively solve the PF parts clustering problems in e-manufacturing system for Mass Customization (MC) mechanism.