On balancing variances of station processing times in unpaced lines

For an unpaced line with equal mean station processing time for all stations, the paper studies how the line’s utilization factor is affected by different patterns of allocating processing-time variances among the stations. A literature review shows that earlier results on this question are ambiguous and inconclusive. Using long simulation runs (up to 240000 production units) and considering ‘larger’ lines with up to 19 stations and 6 units/buffer, the paper identified three desirable variance-allocation characteristics: ‘bowl-shape’, ‘symmetry’ and ‘spike-shape’. Conceptual generalization of these characteristics enable us to explain and reconcile the seemingly contradictory performance of a wide variety of variance-allocation patterns considered in this as well as the earlier studies.