Multidomain Demand Modeling in Design for Market Systems.

Abstract

Consumers make choices based not only on functional product attributes (e.g., fuel economy) but also on non-functional attributes (e.g., vehicle form). Consequently, ignoring non-functional product attributes in demand modeling can lead to product designs less attractive to consumers. This dissertation focuses on two major non-functional product attributes: (i) aesthetic product form as a perceptual product attribute and (ii) services as external product attributes.

A limitation in conventional discrete choice analysis is that it handles functional and non-functional attributes within a single demand model. An aesthetic product form is generated by a potentially huge number of geometric variables; thus, it cannot be quantified simply and it is difficult to integrate with functional attributes. Similarly, when considering services, it is challenging to incorporate the relationship (or channel) between product and service attributes (or multiple providers) into a single demand model.

This dissertation proposes a multidomain demand modeling approach to integrate functional and non-functional attributes, whose values are decided by different design domains, into a single demand model. We employ consumer choice models from Marketing, systems design optimization from Engineering, machine learning algorithms and human-computer interaction from Computer Science, and location network models from Operations Research within a design optimization framework. This work addresses three demand models: (i) a demand model for engineering and industrial design, (ii) a demand model for engineering and service design, and (iii) a demand model for engineering and operations design. The benefits of this unified approach is demonstrated through three respective design applications including gasoline vehicle design, electric vehicle and charging station location design, and tablet and e-book service design.

The contribution of this research is in helping resolve trade-offs between conflicted design domain decisions, by integrating disparate attributes into a multidomain demand model. This work consequently extends the scope of Design for Market Systems from product design to business model design by considering external product attributes.