This article describes the scientific contributions of Milton Sobel. It motivates his research by considering his family background, his war experiences, and his mentors and fellow students at Columbia University. His research in sequential analysis, selection, ranking, group testing, and probabilistic combinatorics are highlighted.

This article is concerned with a loading-dependent model of cascading failure proposed recently by Dobson, Carreras, and Newman [6]. The central problem is to determine the distribution of the total number of initial components that will have finally failed. A new approach based on a closed connection with epidemic modeling is developed. This allows us to consider a more general failure model in which the additional loads caused by successive failures are arbitrarily fixed (instead of being constant as in [6]). The key mathematical tool is provided by the partial joint distributions of order statistics for a sample of independent uniform (0,1) random variables.

We consider the sojourn time V in the M/D/1 processor sharing (PS) queue and show that P(V > x) is of the form Ce−γx as x becomes large. The proof involves a geometric random sum representation of V and a connection with Yule processes, which also enables us to simplify Ott's [21] derivation of the Laplace transform of V. Numerical experiments show that the approximation P(V > x) ≈ Ce−γx is excellent even for moderate values of x.

Recently, Pellerey, Shaked, and Zinn [6] introduced a discrete-time analogue of the nonhomogeneous Poisson process. The purpose of this article is to provide some results for stochastic comparisons of the epoch times and the interepoch times of those processes. Also, we show the relationships between these processes and discrete record values and we provide several results for discrete weak record values.

The concept of generalized order statistics was introduced as a unified approach to a variety of models of ordered random variables. The purpose of this article is to investigate conditions on the distributions and the parameters on which the generalized order statistics are based to establish the likelihood ratio ordering of general p-spacings and the hazard rate and the dispersive orderings of (normalizing) simple spacings from two samples. We thus strengthen and complement some results in Franco, Ruiz, and Ruiz [7] and Belzunce, Mercader, and Ruiz [5]. This article is a continuation of Hu and Zhuang [10].

We investigate some new properties of mean inactivity time (MIT) order and increasing MIT (IMIT) class of life distributions. The preservation property of MIT order under increasing and concave transformations, reversed preservation properties of MIT order, and IMIT class of life distributions under the taking of maximum are developed. Based on the residual life at a random time and the excess lifetime in a renewal process, stochastic comparisons of both IMIT and decreasing mean residual life distributions are conducted as well.

We consider funding an interest-bearing warranty reserve with contributions after each sale. The problem for the manufacturer is to determine the initial level of the reserve fund and the amount to be put in after each sale, so as to ensure that the reserve fund covers all of the warranty liabilities with a prespecified probability over a fixed period of time. We assume a nonhomogeneous Poisson sales process, random warranty periods, and a constant failure rate for items under warranty. We derive the mean and variance of the reserve level as a function of time and provide a robust heuristic to aid the manufacturer in its decision.

In this article, we obtain error bounds for exponential approximations to the classes of weighted residual and equilibrium lifetime distributions with monotone weight functions. These bounds are obtained for the class of distributions with increasing (decreasing) hazard rate and mean residual life functions.

In this article, we consider an insurance risk model where the claim and premium processes follow some time series models. We first consider the model proposed in Gerber [2,3]; then a model with dependent structure between premium and claim processes modeled by using Granger's causal model is considered. By using some martingale arguments, Lundberg-type upper bounds for the ruin probabilities under both models are obtained. Some special cases are discussed.