| Online data in product design and development,e.g.,sales data and customer review data,generates in large amounts,which is owing to the rapid expansion of information and network technologies.The expansion helps customers more easily to gain and learn product attributes and alternative options on one hand,which increases the dimensionality of requirements,and leads to the requirement variety and makes it harder to satisfy customers in high degree.On the other hand,the expansion brings more intense market competition for products,as a result of the convenience and ease to comprehend and dig out characteristics of products from competitors.The mapping process from customer requirements to product configurations is a necessary and vital part to complete product design and development tasks,and has significant impact on promoting market competition and lowering the market risk of products and companies.The mapping process includes two procedures,namely requirements identification,and product configuration.However,the traditional methods regarding to the two procedures currently applied in mapping process fail mapping process in reducing time,cost and promoting satisfaction in real-world design cases recently,which gives insufficient competition of products.Against those backgrounds,this thesis proposes an intelligent mapping process from customer requirements to product configurations based on online data,which is intended to take the advantages of the present information and network technologies and online data,to decrease the time and cost in product design tasks,as well as increase customer satisfaction on the end products.The proposed intelligent mapping process views mapping process as classification tasks between requirements from customers and feasible configuration plans of product,i.e.,the product configuration plan that best satisfy the requirements of a customer is classified or selected to be assigned to the class of customers that shares the same requirements.Then,according to the classification results and objective constraints,designers and companies could decide the final product configurations to be manufactured and marketed in the next stage.The general contents and achievements of this thesis are as follows:Firstly,this thesis analyzed the backgrounds and problems,put forward the research purpose and significance,and proposed a technical solution of intelligent mapping process from customer requirements to product configurations based on online data.The solution figured out how to intelligently process the mapping tasks,and how to deal with the sparse high dimensional data under intelligent mapping conditions,also it provided a new perspective and way for mapping tasks.According to the technical solution,this thesis drew the research map,confined the contents,determined the research methods and tools,and studied related methods.Secondly,this thesis focused on foundational techniques to realize the technical solution,including 3 mainly aspects.(a)Research on the intelligent classifiers.It took those widely utilized intelligent classifiers of different types in real-world engineering problems to conduct a comparative research to find out which types or classifiers are more applicable in mapping process.(b)Research on the dimensionality reduction methods under ordinary data settings.Those generally applied dimensionality reduction methods in engineering situations were taken into research in this study,to fix those methods that have better performance applying with intelligent classifiers in mapping process.And,(c)Research on the dimensionality reduction methods under sparse high dimensional data settings.It put forward a whole new dimensionality reduction methods based on local linear embedding,for improving the performance facing with the situation where data is of sparse and high dimensionality and traditional methods would be of poor performance or even fail to process.Thirdly,combined with the two kinds of online data,and the researches on foundational techniques,this thesis proposed three intelligent mapping approaches,namely intelligent mapping approach based on sales data,intelligent mapping approach based on customer positive review data,and intelligent mapping approach based on sales data and customer review data.Those approaches were studied and introduced thoroughly,and were implemented in real world design cased to testify their advantages over traditional mapping solutions in time,cost and customer satisfaction.This study raised a technical solution of intelligent mapping process,put forward the most capable intelligent classifiers and dimensionality reduction methods,and proposed three intelligent mapping approached based on online data.The researches and achievements provide designers and companies with a new perspective and way to map customer requirements into product configurations in product design and development tasks,which adds the value to the usability of online data,and would help companies save time,cost and promote customer satisfaction,and increase market competition as well as reduce market risk,in return. |
