Conceptually, my research deals with understanding and modeling the complex systems coupling government, corporations, and consumers in order to minimize the environmental impacts associated with engineered consumer products. Such impacts are directly related to design decisions in straightforward although complex ways. However, they are also highly dependent on how corporations and consumers react to the characteristics of products implied by particular design decisions.

The technical tools I am applying to generate such models include the theory of Multi-Stage Games, Random Utility Models of choice, and Optimization with both Discrete and Continuous Variables. Below I detail specific research activities I am currently undertaking.

One way in which technological and economic knowledge can be integrated in a model of industrial behavior is through game theoretic concepts. Particularly, the modeler assumes that firms seek to maximize some objective (e.g., profits) by choosing product designs from some feasible set of alternatives, and that these objectives are coupled in the sense that any given firm's objective depends on the choices of all firms. I have been examining suitable numerical methods to solve Regulated Multi-Stage Design and Pricing Games, with either (or both) continuous or discrete design variables. The staged nature of these games represents the fact that design decisions must typically be fixed some time in advance (e.g., 3 years for automotive designs), while prices remain flexible, in principal, up until the point-of-sale. In the simplest versions, firms choose designs understanding the competitive pricing responses of other firms to these designs, while assuming competitors keep their designs fixed.

I am applying these methods to induce a representation of three fundamental aspects of real markets - profit-driven design decisions, competition between firms, and consumer preferences for various vehicle attributes - onto technology-cost data such as that used in the National Energy Modeling System, reported by the National Academy (2002), or derived and reported by the Northeast States Center for a Clean Air Future (2004) used to inform automotive fuel economy and emissions policy. Of particular interest is seeing how the "Cost-Effectiveness" (to the consumer) of various technologies (1) changes under competitive pricing assumptions and (2) interacts with profit-driven design decisions and competition under alternative market-based regulatory policies.

I have spent a significant amount of time studying price equilibria as part of the study of Multi-Stage Design and Pricing Games. My primary result in this area involves the characterization of equilibrium prices under Logit, Generalized Extreme Value, and Mixed Logit Random Utility Models of demand as solutions to certain fixed-point equations. This approach has a straightforward generalization to cases where firms face regulatory costs.

For Logit, I have used this equation to provide the first generic equilibrium existence proof for multi-product firm price equilibrium under this class of Random Utility Model. This proof requires tools from differential topology, particularly the Poincare-Hopf theorem, to overcome the lack of quasi-concavity in multi-product firm profit functions. This equation also illuminates some structural properties of price equilibria, some expected and some counterintuitive.

Under Mixed Logit, the fixed-point equation provides a numerical method markedly more efficient and reliable than the Newton-Raphson method. Particularly, the fixed-point iteration requires an order of magnitude less work per iteration than the Newton-Raphson method. Moreover, the Newton-Raphson method (or simple modifications) can be unreliable at finding equilibria, while the fixed-point iteration has shown to be quite robust at finding equilibria. Hopefully, the fixed-point iteration will enable numerical analysis of firms with dynamic decision-making models under complex demand specifications like Mixed Logit.

The automotive industry plays a central role in both Greenhouse Gas (GHG) emissions and materials flows in the United States. Yet, there has been little research that attempts to quantitatively address the consequences of GHG policy on key materials flows. I am part of a multi-university research project dedicated to filling this gap; this project is funded by the US NSF's Division of Civil, Mechanical, and Manufacturing Innovation Biocomplexity MUSES (Materials Use: Science, Engineering, and Society) program. In the Spring of 2007, this project was selected to represent NSF funding at the University of Michigan at the 13th Annual Exhibition and Reception for the Coalition for National Science Funding.

In addition to an application of Multi-Stage Design and Pricing Games, we are developing Life Cycle Assessment methods appropriate for a context where design variables, vehicle characteristics, and fleet character change as a result of policy. In terms of Linear Life Cycle Inventories, this requires modeling how the economic and environmental components of the process technology matrices depend on design characteristics. Since design decisions also change the characteristics "demanded" by consumers, one must also model how functional unit demand changes with design variables. Since demand is typically characterized with stochastic modeling methods (e.g., Random Utility Models), uncertainty also enters the Life Cycle Assessment in a new way.

Under the MUSES project and as part of my dissertation research, I have been involved in the collection of Automotive datasets from a variety of sources: particularly marketing organizations, the EPA, and the NHTSA. We are currently consolidating the various datasets to facilitate the construction of new automotive demand models and a historical analysis of technology adoption under CAFE regulation, among other studies.

(Written in 2005)

Solid Freeform Fabrication (SFF) processes have recently introduced the capability to manufacture components of consumer or industrial products from the ground up. This leads to many challenging engineering design opportunities, particularly the construction of internal structures and heterogeneous materials, neither feasible with traditional manufacturing methods. My role falls with Direct Metal Deposition (DMD), a SFF process for use with atomized metal powders developed at U of M by Professor Jyotiroy Mazumder and the CLAIM group. Processes like DMD are making tremendous economic improvements for producers using injection and die casting molds. We have taken an initial assessment of the relative environmental impacts for manufacture of molds through traditional and DMD methods, hoping that SFF can introduce environmental as well as technical and economic improvements.

(Written in 2007)

During my first two years of graduate school I was part of a project that investigated the environmental benfits of cellular telephone remanufacturing. In collaboration with the Technical University of Berlin and a major cellular telephone manufacturer, a Life-Cycle Assessment of new cellular telephones and remanufactured cellular telephones was undertaken. The (somewhat obvious) conclusion that remanufactured handsets are more environmentally benign than new handsets was verified under the assumption that remanufactured handsets replace new ones. However, remanufactured handsets were, at that time, primarily sold in developing markets as a low-cost market expansion vehicle, in conflict with this assumption. Today there is some availability of remanufactured handsets in mature markets, but it is still largely marginal.

Nonetheless, there is evidence that the availability of information technology, including cellular telephones, carries social benefits to developing markets. These benefits include: increased market empowerment (e.g., knowledge of agricultural prices), increased accessibility of healthcare, entreprenurial activity, and empowerment of women. Environmental economics stipulates that such benefits, in so far as they accelerate development, eventually extend to awareness of environmental impact and environmental efficiency improvement. We did not quantify these aspects of this complex isue, but rather presented them to an otherwise technical audience as an example of the interactions between engineering, technology, and society.