An Evolutionary Clustering Method for Part Family Formation with Multiple Process Plans

This research integrates three techniques: genetic algorithm, constraint satisfaction network and granular computation, into an evolutionary clustering method for part family formation. This method includes two modules: Evolutionary Constraint Satisfaction (ECS) modular and Evolutionary Optimization of Granules (EOG) modular. With this method, a machine/part incidence matrix with multiple process plans can be satisfactorily formed into groups. The principle of the ECS modular is to minimize a predefined objective function under the satisfaction of some constraints and search a set of the best process plan combination for the parts involved. The EOG modular is then applied for clustering the matrix into part families and machine cells, respectively. The EOG integrates granular computation with genetic algorithm. The main contribution of this research is the effectiveness of integrating genetic algorithm, granular computing and the concept of neural network for dealing with large-sized cellular formation problem. This proposed model has been verified and confirmed by its accuracy using several popular cases.