Estimating capacity loadings from workcenter arrival and departure events

Abstract

Abstract Manufacturing firms rely on automatic data collection systems to replace labor-intensive data collection methods. Most systems concentrate on locating and counting parts without regard to transaction times; however, by capturing and processing event times, additional system characteristics can be inferred. Using product arrival and departure event transactions based on bar code technology, this research develops a methodology for estimating the amount of workload in a workcenter with a parallel server structure commonly found in practice. This paper presents an evaluation of the accuracy and precision of the resulting workload estimators under conditions of varying product mix. It was found that the relative proportion of products in the production mix can have a significant effect on estimator performance. For production mixes that change infrequently, the estimator was within 5% of the true workcenter load in 98% of the experiments. The estimator is simple and can be continually updated as new data are collected. This research illustrates the value-added benefits to be gained in utilizing bar code timing information for developing estimates of workcenter load especially within computer-automated shop floor control systems.