Research On Product Service System Configuration From The Viewpoint Of Customer Perception

Servitization is one of the world's manufacturing development trend.And it is also an effective way to realize our country manufacturing transformation and upgrading.Product service system is a kind of specific model for manufacturing servitization.Customer needs of product service system are necessarily diversified because of the experience nature and the heterogeneity of service.And the diversification of customer needs determines that customized products and services are as the main form of delivery to achieve higher customer value.Then one of the effective ways lead to customized products and services is configuration of products and services.Because of the intangibility and customer experience of service,product service system configuration is more complex and difficult than traditional product configuration.It becomes a key problem that manufacturing servitization faces.Focusing on the key issues of product service system configuration,this dissertation carries out the following research points based on analysis of past research and advanced industry practice.The main contents include:1.Customer needs identification and transformation.Customer needs are classified into two categories that are functional needs and perception needs.Factor analysis and semantic differential method are used to gain the information about customer perception.And then,multiclass support vector machine model is built on the basis of customer needs-product service system data pairs.The model realizes a transformation from customer needs to product service systems.To run the model,RBF function is viewed as kernel function.Meanwhile,OVO strategy is used to model multiclass support vector machine,cross validation and grid search method are combined to seek the best parameter pair(C~*,?~*).Finally,a case study validates the proposed approach.Furthermore,it is found that synthetically considering customer functional needs and perception needs contributes to higher classification accuracy.2.Mining rules that describe the relationships between product components service modules and customer perception.To satisfy customer perception needs,it is necessary to discern the product and service components which trigger a certain customer perception.To obtain the relationships,a rough set information table is firstly built.Then customers are classified into different groups by cluster analysis with squared Euclidean distance and Ward minimum variance method.After that,dominance-based rough set runs on each customer group to obtain the relationships between product service components and customer perception which are described as“If…Then…”rules.At last,case study demonstrates and validates the proposed approach.3.Realizing interactive configuration of product service system.Firstly,interactive configuration process is described as a recalling process of content-addressable memory which is conducted by Hopfield neural networks.Then three modules that are associative memory module,spurious memory mutation module,and rules generation module are emphasized.After that,four interactive configuration scenarioes that are constant customer needs,changeable customer needs,configuration scheme breach consistency and rule-oriented configuration,are designed to validate the proposed method.Finally,the performance of the proposed method is evaluated by mutation effectiveness,response time,and configuration process visualization.