Master's Research

Title: Multi-Attribute Optimization Based on Conjoint Analysis

Wright State University, Dayton, OH, 2007

Download

Abstract:

Over the last thirty years, there were tremendous advances in multidisciplinary design optimization in reducing computational cost, developing algorithms for efficient sensitivity analysis in reaching an optimum. Most of these efforts assumed a single objective (attribute) function and a multitude of constraints. Very little work has been done in including the designer's preferences as part of the optimization scheme and in addressing the ability to handle multiple attributes simultaneously. This need to develop a systematic method for including designer's preferences is the main focus of this research work. The concept of modeling preferences among multi-attribute alternatives is prevalent in consumer product marketing, and the current work adopts a widely used marketing method known as conjoint analysis. This method is often implemented to assess the individual part-worth of the attributes, which provides insightful knowledge of the products and is then further used to create new products in the market. Conjoint analysis can be integrated with optimization techniques for engineering applications. In this work, a novel method of combining advances from management science and engineering disciplines is implemented. Details about the conjoint analysis algorithm is discussed with an example of designing a pizza. Furthermore, the method is applied to structural engineering applications such as a cantilever beam, a fixed plate, and a composite lightweight torpedo demonstrating the advantages of the method.