A generalized URN problem and its applications

From MaRDI portal





We consider an infinitely large number of managers adopting some new technology which occurs in two types, A and B. We assume that each manager is guided by the following considerations: he analyzes which technology has been adopted by r randomly selected managers and if not less than m of them use A, then he also selects A, otherwise he selects B. Such a decision making rule may seem to be irrational, however, it may be logical in the following situations: if the manager has little information on the comparative gains resulting from the application of the technologies A and B; if he fears risks; if he can gather information only by means of questioning randomly selected managers already using the specific technology and they communicate to him only the type of the technology. It is interesting to find out how the fraction of the managers using only one of the technologies varies, whether one of them covers the weight of the market or whether the proportion of those adopting A (B) tends to a certain limit, and how the number of managers selecting the technology A (B) is growing. It is natural to describe processes of this kind by means of a generalized urn scheme [see \textit{N. L. Johnson} and \textit{S. Kotz}, Urn models and their application (1977; Zbl 0352.60001)]. We imagine an urn of infinite capacity, containing white and black balls. If at each step a certain number of balls are removed from the urn and then, on the basis of a well-defined decision making rule R, one adds to the sample a white or a black ball and together with this additional ball the sample is returned to the urn, then we shall say that a generalized urn scheme with decision making rule R is given. It is necessary to investigate the behavior of the fraction of the balls of each color.











This page was built for publication: A generalized URN problem and its applications

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3318525)