Conjoint-analysis-based multiattribute optimization: Application in acoustical design

The concept of customer preference, or product value, prevalent in economics and management science, is just beginning to be used in engineering design. This concept and the associated measurement approaches offer us a theoretically appealing way to aggregating customer preferences for multiple product attributes into a single objective function, representing total product value, which may then be maximized. Among these methods, conjoint analysis has emerged as the most popular approach in marketing to estimate the value that customers attach to different features of a product that can be at different levels. In this paper, we use conjoint analysis in a novel way to assess a product designer's preferences for addressing a practical problem in acoustical design. We incorporate a designer's preferences for reducing noise in a curved pressure vessel excited with broadband noise. The shell is part of a large industrial machine. The "product" here refers to a broadband vibration absorber(s) attached to the structure. Through direct interaction with the design engineer, we elicit his/her preferences for various alternative design configurations and specify an aggregate value function. We then apply optimization techniques, interfaced to simulation codes, to maximize the value function. We show that this method provides more economical designs compared to certain conventional formulations. We conclude by summarizing some limitations to this research, which point to several future research opportunities.