Configurators have been generally accepted as important tools to interact with customers and elicit their requirements in the form of tangible product specification. These interactions, commonly called product configuring process, aim to find the best match between customers’ requirements and company’s offerings. Therefore an efficient configurator should take both product structure and customers’ preferences into consideration. In this paper, we present a novel iterative method of attributes selection for product configuring procedure. The algorithm is based on Shapley value, a concept used in game theory to estimate the usefulness of certain entities. It iteratively selects the most relevant attribute from the remaining attributes pool and proposes it for customers to configure. Thus it obtains customers’ specification in an adaptive manner in the sense that different customers may have different query sequences. Information content is used as the measure of usefulness. As a result, the most uncertainty can be eliminated and product development team has a better understanding of what customers want in a fix time horizon. Maximum a posteriori criterion is also exploited to give product recommendation based on the partially configured product configuration. Thus the customized 1-to-1 configuring procedure is presented and the recommendation can converge to a customer’s target with fewer interactions between the customers and designers. We also use a case of PC configurator to exemplify and test the viability of the presented method.

This content is only available via PDF.
You do not currently have access to this content.