A particle swarm model for estimating reliability and scheduling system maintenance

Abstract

Modifying data and information system components may introduce new errors and deteriorate the reliability of the system. Reliability can be efficiently regained with reliability centred maintenance, which requires reliability estimation for maintenance scheduling. A variant of the particle swarm model is used to estimate reliability of systems implemented according to the model view controller paradigm. Simulations based on data collected from an online system of a large financial institute are used to compare three component-level maintenance policies. Results show that appropriately scheduled component-level maintenance greatly reduces the cost of upholding an acceptable level of reliability by reducing the need in system-wide maintenance.

omplex estimation of status and operability of multilevel distributed systems

“…In general, case real distributed systems can have the form of networks with sophisticated configuration [21]. Availability assessment for the majority of such systems cannot be performed on the basis of simple summation of availability values of their elements using multiplication theorem of probability as it can be done for systems with series-parallel connection of elements [17].…”