Lane selection in an AGV-based asynchronous parallel assembly line

Automated Guided Vehicles (AGVs) have the flexibility to meet the material handling requirements of asynchronous assembly lines. Furthermore, AGVs can select from among alternative paths in accordance with a prescribed lane selection rule, thereby facilitating the use of parallel server workstations. In this paper, motivated by our work in the automotive industry, several new lane selection rules are proposed. One of these, First Available Server/First Available Buffer/Balanced Work Content (FAS/FAB/BWC), is compared to two existing rules: Alternating Server, and First Available Server/First Available Buffer/Expected Completion Time (FAS/FAB/EC). Three performance measures — job throughput, workload balance among servers, and alteration of input job sequence — are employed. A SIMAN simulation model of a small, asynchronous parallel assembly line is used to study the impact on system performance of both the lane selection rule and the number of parallel servers. Interaction between these two factors is studied through ANOVA. A number of interesting findings are reported for results of the lane selection rules with respect to the three performance measures. Their interpretation is used to motivate further research.