Research On Constraint Based Product Configuration Method And Product Configuration Optimization

Three types of product configuration method, traditional Constraint Satisfaction Problem (CSP) based, Dynamic Constraint Satisfaction Problem (DCSP) based, product configuration ontology and Constraint Satisfaction Problem based, were developed and researched in this paper. The traditional CSP based configuration method can represent product configuration knowledge directly and naturally and has well adaptability when product is evolving, it can also solve the configuration problem directly and effectively, thus can be used for the implementation of production configuration on small-scale configurable product. For the DCSP based configuration method, on the basis of preserving the knowledge representation and solving ability of traditional CSP, it can further directly represent both dynamic configuration knowledge and product composition knowledge, accordingly have more knowledge representation ability and can be used for the implementation of product configuration on middle-scale configurable product. As an optimal configuration method, product configuration ontology and CSP based configuration method can be used for the implementation of product configuration on large-scale configurable product. The simulation experiments and its analysis on traditional CSP and DCSP based configuration methods are also performed by using Heli H2000 series forklift as a research object, and the validity of all the proposed solving algorithms is verified according to the simulation result. Further more, the solving of all three configuration methods is proved to be NP-complete.Product configuration optimization is also researched in this paper and three configuration optimization methods, Interactive Constraint Satisfaction Problem (ICSP) and Model Based Diagnosis (MBD) based, HCP-net (Hybrid Conditional Preference-net) based, Branch and Bound (BB) based, were developed. The ICSP and MBD based configuration optimization method can provide precisely optimized result, however, a large number of search work and user involved activities must be performed to achieve this, thus can be used for the situation that the user both has rigorous requirements on optimized result and will to spend enough time on configuration process. Compared with ICSP and MBD based configuration optimization method, the HCP-net based method can provide relatively less precise optimization result, in the mean time, not so large number of search work and user involved activities will be performed in the optimization procedure, thus can be used for the situation that the user has not so rigorous requirements on optimized result and will to spend some specified time on configuration process. For the BB based configuration optimization method, the veracity of the obtained optimized result is no higher than HCP-net based method, however, it has the least search work and user involved activities among all three configuration optimization methods, thus can be used for the situation that the user has common requirements on optimized result and is not willing to spend time on configuration process. The simulation and analysis on HCP-based and ICSP and MBD based configuration optimization methods are also performed by using Heli H2000 series forklift.