OR Practice—Intelligent Data Compression in a Coal Model

When the U.S. Congress began considering acid rain legislation in the early 1980s, the federal Environmental Protection Administration (EPA) relied on two major models for analyzing the impact of emission reduction policies on the coal and electric utility industries. Debate over the merit of these two models was hampered because they were proprietary, and both had achieved computational tractability through a priori exclusion of the nation's largest and most detailed data bases describing either: (1) the characteristics of individual electric generating units, or (2) coal reserves and washability. Because of general dissatisfaction with the limitations of both models, the EPA contracted with three universities to develop an Advanced Utility Simulation Model (AUSM) that would combine the best features of each of the existing models. This paper describes the approach taken to develop the AUSM, currently used by the EPA, in a manner that achieved computational tractability while making maximum use of the information contained in the available data.