Using fuzzy sets to evaluate the performance of complex system when parameters are uncertain

Abstract

Reliability models rarely incorporate the uncertainty that is present in most of the reliability parameters, which is an important limitation for the managers of production system. This paper presents a fuzzy-probabilistic methodology that was developed to analyse production systems. The methodology employs probabilistic theory to represent the stochastic behaviour of the failure and repair processes and fuzzy sets to represent uncertainty in the parameters. This approach is based on a fuzzy reliability analysis from which is possible to obtain the fuzzy downtime function of a system in a given period of analysis. This function is a crucial element to obtain two performance measures: a financial measure (the fuzzy cost concerning failures of a production system) and a quality of service measure (the fuzzy frequency of failure on deliveries), as it will be illustrated in the final part of the paper through a practical example