Queueing network model for a class of material-handling systems

Abstract

A queueing network-based approximation model was developed to evaluate the performance of a fixed-route material-handling system supporting a multipleworkcentre manufacturing facility. The closest transporter allocation scheme was used to assign empty transporters to jobs needing to be moved between various workcentres. This control methodology creates two interesting difficulties that require solving for an accurate approximation model. A model of the distributions of the location of an assigned transporter as a function of the number of available transporters is developed. In addition, it is shown that owing to the strong dependent service time aspect of this situation, the standard queueing network decomposition approach of treating the material-handling systems as a GI/G/c queue can be highly inaccurate. Thus, it is necessary to develop a dependent service-time model for approximating the queueing delay encountered while waiting for an empty transporter assignment.

Toward a relevant agenda for warehousing research: literature review and practitioners’ input

“…4. Automated picking operation: For example, routing and operations sequencing in automated warehouses [67] and positioning the dwell point (location of S/R shuttle during idle time) in a unit-load AS/RS [68]. 5.…”

Transition of interarrival time patterns between automated and manual configurations of assembly systems

“…Approximation models, which are applicable for largescale problems, mainly employ a decomposition approach to divide a manufacturing system into simpler subsystems. While some decomposition approaches limit the layout topology [for example, no cyclic connections may be present in the layout following [10]] or queue-dependent routing [7], other approaches from the class of closed queuing networks with blocking [25] comprise all topological requirements of the systems considered in the present paper. These suitable models, however, assume that considering the average arrival rate for deterministic systems and a probabilistic interarrival time distribution in stochastic systems is sufficient for performance prediction [5].…”

An Analytical Model of Vehicle-Based Automated Material Handling Systems in Semiconductor Fabs

“…The latter results tend to be inaccurate because of the assumption of state-independent service time. Curry et al (2003) propose a more accurate servicedependent queuing network model that generates approximations that are close to the simulation results but the time to solve the analytic model grows exponentially with the number of vehicles. Bakkalbasi (1990) develops analytic models to approximate the empty vehicle travel times for the following dispatching rules: FCFS, Closest Load First, Closest Load with Time Priority, and Furthest Load First.…”