Argues that library book circulation is a binomial rather than a Poisson process, and that individual book popularities are continuous beta distributions. Three examples demonstrate the superiority of beta over negative binomial distribution, and it is suggested that a bivariate-binomial process would be helpful in predicting future book…

Discusses the purpose of mathematical models and reviews the phases of the modeling process--defining performance, representing the problem, predicting performance, estimating parameters, defining optimization criterion, determining solution, and implementing results. Reviews of book-use, resource allocation, and library network models are…