The Scale-based Product Family Design Based On Improved QFD And Genetic Algorithm

The purpose of mass customization production mode is to produce customized products at the cost of mass production to meet the diversified needs of customers.The core issue of product family design is to maximize product performance and product diversity under limited cost constraints.Therefore,the rational product family design is the basis and guarantee for the implementation of mass customization.Product family design theory consists of modular product family design and scale-based product family design.The former realizes the product's diversification by modifying,adding or deleting the module of products so as to make up various modules combination.But the latter does by changing the value of design parameters of each product so as to make up different values combination.Because scale-based product family is easier to quantify than modular product family and its result of product family design can also meet the customers' needs more widely,so this paper mainly studies scale-based product family design.The core of scale-based product family design is the identification of the platform parameters and the optimization design of the product parameters.In the field of platform parameter identification,now the platform parameters are identified mainly by the parameters' sensitivity or variation coefficient without considering the importance of customer requirements and the influence of the design parameters on customer requirements.So this paper applies QFD into the scale-based product family design.However,the traditional QFD method does not consider the regular change of customer requirements over time when determining the importance of customer requirements,which results in the importance of customer requirements lagging behind the actual situation.And in the determination of relationship matrix elements,there are many shortcomings such as too much dependence on expert experience,strong subjectivity and discretization of values.In view of the above shortcomings,this paper proposes a method to identify the product platform parameters based on improved QFD.First,the importance of customer requirements is determined based on cluster analysis and grey prediction.According to the regular change characteristics of the importance of customer requirements over time,this paper selects the customers and design parameters with high degree of consistency by consistency analysis based on the enterprise transaction history,and calculated the importance of customer requirements over several periods by using fuzzy pairwise comparison method,and analyzes the trend and forecast the importance of customer requirements in the future period by the grey forecasting model.Secondly,based on the sensitivity analysis,the relationship matrix between the customer requirements and the product design parameters is established.In order to overcome the shortcomings of traditional expert scoring method,such as subjectivity and discretization,a sensitivity analysis method is introduced to analyze the influence degree of independent variables on dependent variables,which is used as a basis for determining platform parameters.According to the range of design parameters,the value levels of design parameters is selected.The representative value combinations are selected by orthogonal test table,and the sensitivity is calculated based on the test results.The sensitivity matrix is used as the correlation matrix.In addition,in order to overcome the fuzziness and randomness of QFD evaluation information,cloud computing is applied to deal with the importance of customer requirements,sensitivity and competitive evaluation information of QFD,so that the results will be more accurate and objective.In the field of value optimization of product parameters,compatible decision support problem is the main method to solve scale-based product family design.However,this method can only satisfy the design goals of product family,but can't optimize the performance of product family.Therefore,based on the existing research and according to the inherent characteristics of scale-based product family,this paper proposes a multi-objective and multi constrained genetic algorithm based on crowding distance to solve many problems in the design process of scale-based product family.First,this paper constructs the multiple objective functions according to the enterprise cost,customer satisfaction and performance target,constructs the constraints of the model according to the customer requirements,the relationship between design parameters,and the product's structure,and finally combined with the domain of design parameters establishes the multi-objective optimization model.Then,this paper uses the genetic algorithm based on crowding distance to solve the model,and get the Pareto optimal set of the model through chromosome coding and initialization,decoding,copy selection,crossover and mutation operation.Finally,the optimal solution is chosen by the fuzzy optimization method,and the optimal value of the platform parameters and individual parameters is obtained.The above method is applied to the product family design of general motors,and the results show that in the situation of fixed production cost,the product family using the method in this paper is designed to improve the efficiency of 2%~5% and to reduce the weight of 5%~10% than the original product family,which verifies the effectiveness of this method.