Simulation study for the optimal repair capacity of an IT maintenance center based on LRD failure distribution

Consider various host systems that consist of both hardware and software. There is an IT center which governs the maintenance activities of all of those host systems. The center's major concern is deciding the proper level of repair capacity for various kinds of failures due to hardware, softwa...

Full description

Saved in:

Bibliographic Details
Published in:	Computers & operations research 2004-04, Vol.31 (5), p.745-759
Main Authors:	Sohn, So Young, Oh, Ki Yeol
Format:	Article
Language:	eng
Subjects:	Computer peripherals Long range dependent property Mathematical models Monte Carlo simulation Multi-response optimization Records maintenance Software Statistical analysis Studies System failure
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Consider various host systems that consist of both hardware and software. There is an IT center which governs the maintenance activities of all of those host systems. The center's major concern is deciding the proper level of repair capacity for various kinds of failures due to hardware, software, and operator related problems where corresponding failure and repair times may exhibit fairly large variations. In this paper, we use Monte Carlo simulation in order to decide the optimal repair capacity when both failures and repair times of individual host systems exhibit long-range dependent (LRD) property. In terms of utilization rate, average and maximum waiting times, we suggest the optimal capacity of maintenance. Various sensitivity analyses follow. Scope and purpose Recent trend in Information Technology in terms of maintenance is the care of both hardware and software. The main feature of the failures of software and hardware in the advanced IT system would be variability in both failure and repair times which cannot be well explained by the classical queueing models. We use a heavy tailed Pareto distribution with covariates in order to represent such phenomena and derive the optimal maintenance capacity for various breakdowns. The result of our analysis is expected to provide the practical implication of the maintenance planning.
ISSN:	0305-0548 1873-765X