GOALI: A Choice Modeling Approach to Incorporating Heterogeneous Consumer Preferences into Enterprise-Driven Product Design

点击翻译按钮获取中文摘要

The research objective of this GOALI award is to develop an information-based hierarchical choice modeling approach for managing and analyzing consumer preference data from multiple sources. This approach will enable the transformation of qualitative marketing preferences to quantitative performance measures in order to manage the complexity of demand modeling for large-scale artifacts. Advanced probabilistic choice modeling approaches will be used to capture customer preferences and market heterogeneity. Model fusion (data enrichment) techniques will be developed to effectively combine revealed and stated preference data from various sources to improve prediction accuracy and model validity. Finally, the integration of the choice model into an enterprise-driven design framework will be demonstrated in the design of vehicle interior package dimensions, considering multidisciplinary tradeoffs.If successful, the proposed research will offer a general demand modeling and enterprise-driven design approach that can be applied to large-scale systems, engineered consumer goods, and mobility/prosthetic devices. Our research will leverage the results from existing discrete choice analysis techniques in market research and transportation engineering, and extend their use in engineering design. The strong collaboration between the university and industry will ensure that the resulting technology is successfully transferred and implemented. The proposed research has far reaching consequences for design education by instilling rigor into capturing customers' preference in the context of product usage through analytical modeling. Results of the research will be incorporated into undergraduate and graduate design courses in mechanical engineering. Industry co-PIs will have significant involvement in the project through collaborations with the research team, service on student thesis committees, invited research seminar presentations, and guest lecturer opportunities.