Optimal bounds on expectations of order statistics and spacings from nonparametric families of distributions generated by convex transform order (Q2256088)

Order statistics and their linear combinations play an important role in lifetime analysis. The authors consider random variables with a nonparametric distribution \(F\) which precedes a fixed distribution \(W\) in the convex transform order. If \(W\) is the uniform or the exponential distribution, then the distribution \(F\) has an increasing density and an increasing failure rate. Applying a projection method, the authors derive sharp upper bounds on the expectations of single order statistics and spacings based on i.i.d. random variables from \(F\). These bounds are expressed in terms of the population first two moments. Distributions attaining the bounds are given; general results for distributions with increasing density functions are discussed. The authors illustrate their approach by computing the derived bounds for special sample sizes.