This report summarizes the research to date performed by the author and his students on the use of modern multiparameter estimation techniques in the building of an attrition rate generator in support of the USMC Officer Planning and Utility System (OPUS). Three main areas are identified: The cell aggregation problem; the specifics of parameter estimation; the need to match forecasting techniques to the specific application. Most of the effort has been in the first two of these areas and much has been learned. The aggregation problem, i. e. , the grouping of personnel cells into an appropriate number having common size and attrition behavior, has emerged as the most important problem that requires immediate attention. Its resolution is expected to lead to a clear policy for multiparameter estimation. Estimation and forecasting are both impacted by the nature of the data base. It is likely that specific applications will use differing data bases and differing statistical techniques as well. (sdw)