Threshold Control Policy for Maintainability of Manufacturing System with Unreliable Workstations

Abstract

This paper is concerned with the maintainability issues and standby workstations provisioning which are the key concerns of the system designers in manufacturing system to overcome the unpredictable interruptions due to workstation failures. An optimal F policy is proposed to control the admission of jobs in the case when the capacity of the system is full. By using queue theoretic approach, the performance of manufacturing system consisting of finite identical workstations in parallel is explored. Various realistic features including minor and major breakdowns of the service facility, degraded failure rate of workstations and controlled admission of failed workstations are also taken into account for modeling the repairable manufacturing system. The spectral method is employed to compute the transient state probabilities for the governing model. Greedy selection and Newton-quasi methods are used to determine the optimal parameter/threshold by minimizing the total cost associated with different activities.