Research On Automotive Mass Customization Mechanism Based On AHP-BP Neural Network

At present,there are still huge challenges both in cost and speed for automotive companies to implement MC mode.The research on automotive MC is mainly based on qualitative analysis and lack of quantitative analysis.And further exploration is needed to find the balance between mass production mode and customized production mode.This article studies the customer demand hierarchy model,demand cluster analysis and recombination analysis,demand conversion and simulation prediction in the context of automotive MC.The main research results are as follows.（1）This article establishes a hierarchical model and indicator system of customer demand based on AHP.According to the system,a survey questionnaire is designed to obtain original data.Taking a specific sample as an example,AHP is used to analyze customer requirements quantitatively,and complete customized demand index weights are obtained,which converts the subjective demand information into specific weights data.（2）With the help of SPSS,cluster analysis is performed based on the samples obtained by AHP.Taking demand variables as the object,R-type algorithm is used for cluster analysis,and divide total demand variables into five categories.Demands of different categories are reorganized,thus forming four different product levels,namely primary product group,intermediate product group,advanced product group and premium product group.Taking intermediate group as an example,a precise customer demand index system is established.Taking customers as the object,K-means algorithm is used for cluster analysis,and total sample are divided into four categories,which compresses customer demand information effectively.（3）Through QFD this article establishes a house of quality for intermediate group that expresses the relationship between customer demand and functional modules.According to this house of quality,Delphi method is used to obtain the weight data of functional modules.These data,together with the customer demand weight data obtained by AHP,provide training samples for the BP neural network.With the help of Matlab,the corresponding BP network model is constructed,then uses the samples to train the network and obtains good accuracy and effect.Finally,the network is simulated.The simulation results can predict and guide how to reasonably allocate resources such as funds,human resources and material resources according to the weight data of each submodule.This article explores the overall implementation mechanism of MC in automotive industry,and explores the resolution of the contradiction between mass production and customized production.It provides certain reference significance and practical value for guiding automobile manufacturers to meet customers’ customization needs to the greatest extent with the least comprehensive cost and shortest product cycle.