Probability in the Engineering and Informational Sciences, 23, 2009, 563–582. Printed in the U.S.A.
doi:10.1017/S0269964809990015
INTRINSIC AGING AND CLASSES OF
NONPARAMETRIC DISTRIBUTIONS
RHONDA RIGHTER
Department of Industrial Engineering and Operations Research
University of California
Berkeley, CA 94720
E-mail: rrighter@ieor.berkeley.edu
MOSHE SHAKED
Department of Mathematics
University of Arizona
Tucson, AZ 85721
E-mail: shaked@math.arizona.edu
J. GEORGE SHANTHIKUMAR
Department of Industrial Engineering and Operations Research
University of California
Berkeley, CA 94720
E-mail: shanthikumar@ieor.berkeley.edu
We develop a general framework for understanding the nonparametric (aging) prop-
erties of nonnegative random variables through the notion of intrinsic aging. We also
introduce some new notions of aging. Many classical and more recent results are
special cases of our general results. Our general framework also leads to new results
for existing notions of aging, as well as many results for our new notions of aging.
1. INTRODUCTION AND SUMMARY
Consider a nonnegative absolutely continuous random variable Y . The random vari-
able Y could be the time to default in a credit risk (e.g., Ammann [1]), the lifetime of
a reliability system (e.g., Barlow and Proschan [2]), or the demand for an item in a
supply chain (e.g., Porteus [15]). Nonparametric (aging) properties of the distribution
function of this random variable often play a crucial role in characterizing the optimal
© 2009 Cambridge University Press 0269-9648/09 $25.00 563