John Chung-I Chuang, Economies of Scale in Information Dissemination over the Internet

Committee: Alex Hills (EPP/SCS), Granger Morgan (EPP/ECE/Heinz), Marvin Sirbu ­ chair (EPP/ECE/GSIA) and Hui Zhang (SCS)

This dissertation studies economies of scale (EoS) asso-ciated with information disseminated over the Internet. At the product level, such economies may be realized through information bundling, site-licensing and subscriptions. At the information transport level, they may be realized through just-in-time delivery, multi-cast, network caching and replication.

Along the object dimension, a multi-product bundling model with multi-dimensional consumer taste characteristics is developed for goods such as academic journals. Using empirical journal usage data and cost projections for information-delivery over the Internet, the model finds that articles-on-demand should account for a significant fraction of revenue when articles and subscriptions are optimally priced according to a mixed bundling strategy.

Along the receiver dimension, a communication cost model for multicast is developed which demonstrates that multicast group size can serve as an excellent proxy for multicast tree cost. Computer simulations show that, statistically, multicast tree length grows at the 0.8 power of the multicast group size until the point of tree saturation, beyond which additional receivers can be added to the group without further tree growth. This suggests that a two-part tariff may be appropriate for a cost-based approach to multicast pricing.

Along the temporal dimension, EoS savings can be realized through network caching and replication. They can be treated as different service classes within a unified Quality-of-Service framework. A formal resource mapping model is developed which allows services with different traffic profiles and performance specifications to be mapped into an optimal combination of storage and transmission resources.

The work was supported by the NSF grant IRI-9411299, the Council on Library and Information Resources A.R. Zipf Fellowship in Information Management, and Carnegie Mellon.

Heather L. MacLean, Life Cycle Models of Conventional and Alternative-Fueled Automobiles

Committee: Chris Hendrickson (CEE), Lester Lave ­ co-chair (GSIA/EPP/Heinz), Francis McMichael ­ co-chair (CEE/EPP), Mitchell Small (CEE/EPP) and Ron Williams (General Motors)

This thesis uses an economic input-output life-cycle analysis to assess alternative automobile fuel/engine combinations (unleaded gasoline, reformulated gasoline, alcohol/gasoline blends, and compressed natural gas in spark ignition direct and indirect injection engines, neat alcohols in spark ignition direct injection engines, and diesel in compression ignition direct and indirect injection engines).

None of the alternatives emerges as a clear winner. The feasibility of the alternatives depends on the focus of future regulations, government priorities, and technology development. For example, in the near-term if global warming is the primary concern, then direct injection diesel or gasoline automobiles are attractive. If local air quality is also a focus then diesels are less attractive. Compressed natural gas vehicles have the lowest emissions but have difficulty storing sufficient fuel onboard to attain a conventional range.

Although the fuel properties and engine characteristics of the options are diverse, the results of the current assessment with respect to economics, externalities, and vehicle attributes indicate that these are remarkably similar. This is largely a result of automobile manufacturers having to satisfy stringent standards in a highly competitive market. Engineers have learned to overcome the inherent limitations of each fuel and engine type to remain competitive. However, the standards are usually stated in terms of automobile performance, and do not consider the impact of the entire vehicle life cycle. Without a systems level analysis, automobile manufacturers and regulators are led to favor alternatives whose flaws show up only at the systems level.

The work was supported by General Motors, the Green Design Consortium, the Teresa and H. John Heinz III Foundation, Texaco, the Sloan Foundation, and the US EPA.

Davidson Edits Book on Clair Patterson "Clair Cameron Patterson (1922-1995) was one of the world's great scientists. He gave us the first accurate age of the earth, pioneered the concept of contamination control in handling environmental samples, and devoted his life to the study of lead pollution. How one individual could do so much in a lifetime and help the world in so many ways is a story that deserves to be told." Thus, opens the introductory chapter by Cliff Davidson of a new book he has assembled and edited titled Clean Hands: Clair Patterson's crusade against environmental lead contamination.

More than any other investigator, Patterson was responsible for showing that the high lead levels observed throughout the world were not the result of natural background that conventional wisdom claimed they were, but were instead the result of contamination from human use of lead in a wide variety of applications, most notably as an anti-knocking agent in gasoline.

In ten chapters written by Patterson's former students and colleagues, this book recounts the development of his professional career. Working at CalTech, using clean methods that many workers in the field considered extreme and unnecessary, Patterson collected samples in places like the deep ocean and the Greenland ice shelf. Slowly the data began to accumulate indicating that the world was awash in lead released by human activities. Much of the book is devoted to recounting his often lonely and controversial fight against the lead industry and other interested parties which ultimately convinced the scientific and regulatory world that lead posed a serious environmental hazard.

_____________________________________________________________________