Reliability

Statistic al Scienc e 2006, V ol. 21, No. 4, 427 DOI: 10.1214 /0883423 06000000664 In th e Public Domain Reliabilit y This sp ecial vo lume of Statistic al Scienc es present s some innov ativ e, i f not p ro v o cativ e, idea s in the area of reliabilit y , or p er h aps more appropr iately n amed, in tegrated system assessmen t. In this ag e of exp o- nen tial gro w th in science, engineering and tec hnol- ogy , the capabilit y to ev aluate the p erform ance, re- liabilit y and sa fet y of complex systems presents new c h allenges. T o da y’s metho dology m ust resp ond to the ev er-increasing demands f or s u c h ev aluations to pro vide k ey information for decision and p olicy mak- ers at all levels of go vernmen t and in dustry—p roblems ranging from in ternational securit y to space explo- ration. W e, the co-edito rs of this v olume and the authors, b eliev e that scien tific progress in r eliabilit y assessmen t requires the dev elopmen t of pro cesses, metho ds and to ols that combine div erse inform a- tion typ es (e.g., exp erimen ts, compu ter sim ulations, exp ert kn o wledge) from div erse sources (e.g., sci- en tists, engineers, busin ess dev elop ers, tec hnology in tegrators, decision mak ers) to assess quan titativ e p erformance metrics that can aid decision m aking under uncertain ty . T hese are highly in terdisciplinary problems. T he pr incipal role of statistical sciences is to bring statistical r igor, thinkin g and metho dology to these problems. Sal lie Kel ler-McNulty is D e an of t he Ge or ge R. Br own Scho ol of Engine ering, Ri c e U niversity, H ouston, T exas 77005 –1892, U S A (e-mail: sal lie@ric e.e du ). Aly son Wilson is a T e chnic al St aff Memb er, Statistic al S cienc es Gr oup, MS F600, L os Alamos National L ab or atory, L os Ala mos, New Mexic o 87545, US A (e-mail: agw@lanl.gov ). Christine Anderson-Co ok is a T e chnic al Staff Memb er, St atistic al Scienc es Gr oup, MS F600, L os Ala mos National L ab or atory, L os Alamo s, New Mexic o 87545 , U SA (e-m ail: c-and-c o ok@lanl.gov ). This is an electronic reprint of the orig inal article published b y the Institute of Mathematical Statistics in Statistic al Scienc e , 20 06, V ol. 2 1, No. 4, 427 . This reprint differ s from the original in pagination and t yp ogr aphic detail. Bedford, Quigley and W alls op en th e issue b y re- viewing the role of exp ert ju dgmen t to supp ort reli - abilit y assessmen ts within the systems engineering design pro cess, with an aim to ward develo ping a framew ork for trac king reliabilit y thr ough the de- sign pro cess. Cham b ers, James, Lambert and V an- der Wiel devel op metho d s for a conte mp orary soft- w are issue—monitoring the health of net wo rked ap- plications. Pe˜ na reviews recen t work in the mo d eling and analysis of recurrent eve nts, a nd Lee and Whit- more r eview fi r st h itting time mo d els, with sp ecial emphasis on regression structur es. Both o f these pa- p ers ha v e wide relev ance in diverse fi elds, includ- ing en gineering, reliabilit y , biomedicine and p ublic health. Wilson, Gra v es, Hamada and Reese review metho dology for addr essing system reliabilit y with limited full system testing, includ ing resource allo- cation considerations. Lindqvist reviews b asic mo d- eling approac hes for failure an d maint enance data from r epairable systems, with emp h asis on imp erfect repair mo dels and th e trend-renewal pro cess. Esco- bar and Meek er review accelerated testing methods , whic h help estimate the f ailure-time distributions or long-term p erformance of comp onents of high- reliabilit y pro d ucts. This vo lume pr esen ts some forw ard lo oking ideas from an international set of r esearc h ers that are fo- cused on the gro wing c hallenges facing statistical sciences in the domain of integ rated system assess- men t. W e h op e y ou enjo y the material. Sallie Keller-McNult y , Co-Editor Alyson Wilson, Co-Editor Christine Anderson-Co ok, Co-Editor 1