Study of aging properties of a binary random variable (Q1369644)

The identification of special families of distribution functions as mathematical models of aging and wear has been discussed in detail by the author and \textit{L. P. Khomenko} [ibid. 28, No. 3, 403-421 (1992), resp. ibid. 1992, No. 3, 92-113 (1992)]. It is usually not difficult to prove that these families (classes) of distributions are included in one another. A more difficult issue is strictness of inclusion: many different distributions have to be examined as counter-examples. Here we consider the distribution of a binary random variable as a prototype, and derive bounds on its parameters for different aging properties of the distribution. The analysis reveals some new regularities of aging distributions, as well as interrelationships and differences between ``aging'' and ``rejuvenating'' distributions.