Relative preference-based product configurator design

Abstract

Product configurators have been considered as a major enabling toolkit for product customization. Configurators interact with customers and transform customers' specific requirements into a set of tangible product specifications. They serve as the bridge between customer needs and companies' offering. Current product configurators often require customers to make choices based on the product design parameters. However, customers may not have enough expertise of the product which they are unfamiliar with. A semantic gap exists between customer requirements and product parameters. It increases the confusion and difficulty that customers encounter in configuration process. This paper attempts to improve the performance of product configurators by bridging the semantic gap between customer needs and detailed product design parameters. Based on historical configuration data, we build a ranking function which leverages on customers' relative preferences towards attribute choices. Then a relative preferences-based product configurator will be developed to bridge the semantic gap by leveraging on the relationship network. When using the relative attribute-based configurator, customers don't need to specify their desired attribute choices one by one. They just need to indicate the relative preferences towards a reference product variant and an updated product variant will be presented based on the ranking function. This shields customers from the potentially demanding and confusing configuration task.