SUBJECT: Ph.D. Proposal Presentation
BY: Benjamin Lee
TIME: Monday, July 22, 2013, 9:00 a.m.
PLACE: Love Building, 109
TITLE: Towards Efficient Value-Driven Design and Decision Making in Building Energy Modeling
COMMITTEE: Dr. Chris Paredis, Chair (ME)
Dr. Godfried Augenbroe (ARCH)
Dr. Bert Bras (ME)
Dr. Julie Linsey (ME)
Dr. Brian German (AE)


Within the design community, a growing number of researchers have shown interest in the notion of a Value-Driven Design (VDD) approach, in which designers, engineers, and managers focus on maximizing the 'value' of a particular design, rather than a set of specified requirements. The principles of VDD are simple and robust, and the rigor of decision theory is leveraged to provide a strong mathematical basis for rationality for decision making in design. It has been argued that a value-driven approach can eliminate troubling issues that arise during the design process, such as cost and schedule overrun, by taking a holistic perspective on how individual design decisions impact high-level attributes of concern. However, a common criticism of Value-Driven Design, and similarly focused approaches, is that while they possess a robust mathematical foundation, they are difficult to use in practice because methods for their implementation have not been sufficiently developed. Further, some argue that less rigorous methods appear to provide reasonable results in practice, and so VDD is not necessary. To reconcile these disparate viewpoints on the merits of VDD, it must be shown that VDD makes a valuable contribution to the design process, and the cost to actually implement VDD in practice must be reduced. This leads to the motivating question which will drive this research:
How can the application of value-driven approaches be made more pragmatic for large scale engineering problems?
As a response, I propose that investigating effective characteristics of value models will lead to insights about its applicability as well as identify opportunities for reducing the cost of its implementation. Specifically, I propose to investigate the characteristics of an effective value model within the domain context of Building Energy Modeling (BEM). BEM will serve as a good domain for the investigation of VDD, as it involves several elements of complex systems similar to other domains: multiple stakeholders, integration of disparate systems, significant design and operation uncertainties, competing objectives, and large investment requirements among others. Further, BEM currently struggles with the task of advancing the state of the art in building energy efficiency, while still meeting user comfort and other aesthetic requirements.