Access count of this item: 303

Files in This Item:
File Description SizeFormat 
02632771111178373.pdf612.94 kBAdobe PDFView/Open
Title: Random proportional Weibull hazard model for large-scale information systems
Authors: Kobayashi, Kiyoshi  kyouindb  KAKEN_id
Kaito, Kiyoyuki
Author's alias: 小林, 潔司
Issue Date: 2011
Publisher: Emerald Group Publishing Limited
Journal title: Facilities
Volume: 29
Issue: 13/14
Start page: 611
End page: 627
Abstract: [Purpose]– This study aims to focus on asset management of large-scale information systems supporting infrastructures and especially seeks to address a methodology of their statistical deterioration prediction based on their historical inspection data. Information systems are composed of many devices. Deterioration process i.e. wear-out failure generation process of those devices is formulated by a Weibull hazard model. Furthermore, in order to consider the heterogeneity of the hazard rate of each device, the random proportional Weibull hazard model, which expresses the heterogeneity of the hazard rate as random variables, is to be proposed. [Design/methodology/approach] – Large-scale information systems comprise many components, and different types of components might have different hazard rates. Therefore, when analyzing faults of information systems that comprise various types of devices and components, it is important to consider the heterogeneity of the hazard rates that exist between the different types of components. In this study, with this in consideration, the random proportional Weibull hazard model, whose heterogeneity of hazard rates is subject to a gamma distribution, is formulated and a methodology is proposed which estimates the failure rate of various components comprising an information system. [Findings] – Through a case study using a traffic control system for expressways, the validity of the proposed model is empirically verified. Concretely, as for HDD, the service life at which the survival probability is 50 percent is estimated as 158 months. However, even for the same HDD, use environment differs according to usage. Actually, among the three different usages (PC, server, others), failures happen earliest in the case of PCs, which have the highest heterogeneity parameter and a survival probability of 50 percent after 135 months of usage. On the other hand, as for others, its survival probability is 50 percent at 303 months. [Originality/value] – To operationally express the heterogeneity of failure rates, the Weibull hazard model is employed as a base, and a random proportional Weibull hazard model expressing the proportional heterogeneity of hazard rates with a standard gamma distribution is formulated. By estimating the parameter of the standard proportional Weibull hazard function and the parameter of the probability distribution that expresses the heterogeneity of the proportionality constant between the types, the random proportional Weibull hazard model can easily express the heterogeneity of the hazard rates between types and components.
Rights: © Emerald Group Publishing Limited
This is not the published version. Please cite only the published version. この論文は出版社版でありません。引用の際には出版社版をご確認ご利用ください。
URI: http://hdl.handle.net/2433/150117
DOI(Published Version): 10.1108/02632771111178373
Appears in Collections:Journal Articles

Show full item record

Export to RefWorks


Export Format: 


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.