Queueing network model for a single-operator machine interference problem with external operations

Machine interference is a significant problem in many manufacturing systems. Prior research shows that machine interference may be as high as 10% of machine time. This paper proposes a queueing network model for a single-operator, machine interference problem with external operations, i.e., those tasks that can be completed while the machine is running. The interactions for part/machine and machine/operator are modeled as an open and a closed queueing network, respectively. In the open network, part inter-arrival time follows an exponential distribution. In both networks, service times follow a general distribution that is characterized by their first two moments. An iterative procedure is developed to solve the proposed model. Solution quality is justified by an industry-based case study. Data were collected from the integrated circuit (IC) ink-marking machines of a leading IC packaging company in Taiwan that allowed the construction of an experimental design for computational tests that encompassed a wide range of production scenarios. Empirical results show promise for the proposed methodology in helping to solve industrial problems. Model limitations as well as future research opportunities are discussed.