Research On Constraint-Based Product Configuration Method For Mass Customization

The purpose of Mass Customization (MC) is to provide the customized product or service with high efficiency and low cost of mass production. Product configuration technology is an important support tool for the implementation of MC. It can help to improve the efficiency about enterprise capturing the orders from the customers, decrease the time of product design and the lead-time for production, decrease the storage of work-in-process and finished products, increase control of production and enhance the customer satisfaction degree for customized products.According to the analysis for the concept and essential characteristics of MC in details, in combination with the definition, characteristics and advantages of the configurable products, we can indicate that the production and operations based on configurable products can effectively combine the benefits from mass production and one-of-a-kind production in order to effectively implement the strategy of MC in the customer-oriented market environment.Aiming at different complexities with configurable products, three different kinds of constraint satisfaction problem theory models were proposed in the thesis in order to separately deal with product configuration problems with low, middle and high complexity.The first is the product configuration method based on Non-binary Constraint Satisfaction Problem (NCSP). One non-binary backtracking algorithm and five different kinds of non-binary forwards constraint propagation algorithms were proposed for solving NCSP, and the solving performances for those algorithms were compared and analyzed by experimental simulation. The product configuration method based on NCSP was proposed and the experimental simulation for a configurable car indicates that the product configuration method can directly and naturally represent non-binary product configuration knowledge and can also deal with the built product configuration model directly and effectively. Hence, the NCSP-based product configuration method can be used to handle with the product configuration problem with low complexity.The second is the product configuration method based Non-binary Conditional Constraint Satisfaction Problem (NCCSP). We proposed a non-binary conditional constraint satisfaction problem model. One backtracking algorithm and 5 different kinds of forwards constraint propagation algorithms were proposed and the experimental simulations for a configurable car and a configurable hospital monitor indicate that the NCCSP-based product configuration method, on the basis of preserving the advantages of the NCSP-based product configuration method, can directly represent dynamic product configuration knowledge and can also solve the built product configuration models directly and effectively. Hence, the NCCSP-based product configuration method can be used to handle with the product configuration problem with middle complexity.The third is the product configuration method based Generic Conditional Constraint Satisfaction Problem (GCCSP). A generic conditional constraint satisfaction problem model was proposed and the solving algorithm for the GCCSP was given in the thesis. The UML-based product configuration conceptual model was built and the transformation method for the UML-based product configuration conceptual model into GCCSP was proposed. The GCCSP-based product configuration method was proposed and the experimental simulation for a configurable hospital monitor justified the proposed method. The GCCSP-based product configuration method, on the basis of preserving the advantages of NCSP-based and NCCSP-based product configuration methods, can conveniently conceptualize all kinds of product configuration knowledge systematically and universally and can also solve the built product configuration model rapidly and effectively. Hence, the GCCSP-based product configuration method is an ideal method and can be used to handle with the product configuration problem with high complexity.The three different product configuration methods proposed can provide the guidance of methodology for us in order to rapidly develop all kinds of complicate product configuration systems under mass customization. Hence, it is possible for enterprises to quickly respond the individual requirement and improve the core competitive advantage, which further establish the stable foundation for enterprises.