A stochastic model of an information center
Jack Minker · NASA Technical Reports Server (NASA) · 1969
In a recent paper [ l ] the author investigated a stochastic model relevant t o information handling centers best typified by computer u t i 1 i ties and document storage and retrieval centers.teristics o f information centers were evaluated for retirement policies t h a t govern when items are retired from a primary store t o a less accessible store.The results obtained assumed t h a t the primary store was of unbounded capacity.sider the case where the primary store has a finite capacity.The growth charac-In this paper we remove this restriction and con-A set of integral equations is derived for the expected number of items i n the primary store.arrival d i s t r i b u t i o n for documents, the request distribution, and the parameters associated w i t h the retirement policy.No particular limiting assumptions have been made w i t h respect t o the form of the distributions.The set of integral equations are solved for document arrivals thatThe integral equations depend only upon the follow a Poisson distribution.The expected value of the size of the store approaches the result given i n [l] as M, the size of the primary store, becomes unbounded. . INTRODUCTIONIn this paper we shall be concerned w i t h a mathematical model t h a t describes a portion of the operation of an information center or a computer u t i l i t y .Although many papers have been written concerning librar i e s and information centers, relatively few papers describe mathematical models.An extensive bibliography on papers discussing the use o f librar i e s has been prepared by DeWeese [Z].graphy generally present a summary of data collected without reference t o mathematical models f o r describing or predicting the use of books.The papers cited i n that biblio-Jain [3] reviews some twelve mathematical models that predict the use of books.I n addition, Jain develops a model of his own.The author [ I ] has developed a model of a library t h a t makes use of the chara c t e r i s t i c s that describe book use to determine the expected size of a primary data store.paper.models t o describe how books are used, while we would employ the results