The Most High-Impact Production Flow Details

Many teams come into simulation with unrealistic, Excel-driven expectations—“if I double the machines I double the output,” only to find that small changes send throughput predictions wildly off. Recognizing the non-linear effects of process-time variability, rework loops, reentrant flows or buffer blocking, in this article we’ll show - backed by peer-reviewed and industry studies - how even a handful of high-impact flow-model details can swing throughput by double-digit percentages.

The percentage ranges shown below are illustrative examples from published studies; not predictions of how the impact might swing in any particular scenario.

1. Process-Time Variability

When the coefficient of variation (CV) of cycle times increases, queue-induced waiting inflates and chokes throughput.

Example: –35 % throughput as CV rises from 0.2 to 0.3 [1].

2. Rework Loops

Inspection or test stations that reroute parts back through repair create feedback flows, inflating cycle-time variance.

Example: –15 % to –20 % net throughput under moderate (≈ 5 %) defect rates [2].

3. Reentrant Routing

Multiple visits to the same resource compound queues and blocking.

Example: –25 % to –50 % difference in throughput from reentry dynamics alone when modeling a relatively simple fab with 172 processes on 24 stations [3].

4. Feeder-Lines

Small feeder systems—like vibratory bowls, hoppers or pick-and-place loaders—deliver parts into your main line. Assuming they never run out hides starvation and blocking.

Example: Throughput overestimated by 5 % to 20 % when feeder dynamics are not correctly captured [4].

5. Changeover Dynamics

Changeovers and setup variation influence batch sizes and thus utilization.

Example: 40 % to 50 % difference in throughput from capturing setup-detail [5], and up to 121 % from SMED .[6]

6. Batching & Lot Sizes

Larger lot releases increase WIP and waiting; smaller batches improve flow at the expense of more setups.

Example: Throughput can vary by –20 % to +15 % depending on lot-size tuning [7].

7. Rework Loops

Process yield may send material to rework loops which introduce extra queues and variation, reducing net output beyond the scrap rate losses from waste alone.

Example: Net throughput difference of 5 % to 15 % [8].

8. Human Availability & Shift Patterns

Assuming constant operator availability overlooks breaks, shift changes and skill-level variation.

Example: Throughput variance of –5 % to –15 % compared to idealized availability [9].

Conclusion

Each missing detail can skew throughput by ±10 %, ±50 %, or more. Linelab lets you easily define all these critical elements — stochastic times, rework paths, feeder lines, reentrant loops, setups, batching, maintenance, yield and labor availability. Adjust parameters and run every “what-if” in minutes, quantifying ± percent-level impacts without rebuilding from scratch and making investment-grade recommendations with confidence.

References

Govindarajan, M., & Kumar, S. (2024). Variability propagation in manufacturing systems: the impact of the processing time distribution on the inter-departure time. Journal of Manufacturing Systems. Retrieved from https://www.tandfonline.com/doi/full/10.1080/21681015.2024.2346080
Flapper, S. D. P., Bekker, R., & Hoeven, J. (2004). Performance analysis of production systems with rework loops. International Journal of Production Research, 42(22), 4709–4732. https://doi.org/10.1080/07408170490458553
Arabaca, F., & Cosgun, B. (2018). Simulation Analysis of Segmented CONWIP: Application to Reentrant Flow Lines. ExtendSim Conference Paper. Retrieved from https://extendsim.com/images/downloads/academic/adopters/arabaca-conwip2018.pdf
Li, J., Alden, J., Rabaey, J. (2005). Approximating feeder line reliability statistics with partial data collection in assembly systems,.Computers & Industrial Engineering, 48(2), 181-203. https://doi.org/10.1016/j.cie.2005.01.006
Şahin, R., & Koloğlu, A. (2022). A case study on reducing setup time using SMED on a turning line. Gazi University Journal of Science. https://doi.org/10.35378/gujs.735969
Ang, Z., Cheah, C. K., and Prakash, J. (2025). “Enhancing Throughput in Labor Intensive Assembly Lines.” E3S Web of Conferences 603:04021. https://doi.org/10.1051/e3sconf/202560304021.
Chung, K., & Nelson, B. L. (1994). Batch size effects in the analysis of simulation output. Operations Research, 30(3), 556–567. https://doi.org/10.1287/opre.30.3.556
Li, J. (2004). Performance analysis of production systems with rework loops. IIE Transactions, 36(8), 755–765. https://doi.org/10.1080/07408170490458553
Song, H., Tucker, A., Murrell, K, Kauffman, M. (2013). The Impact of Pooling on Throughput Time in Discretionary Work Settings. Academy of Management Annual Meeting Proceedings. https://doi.org/10.5465/ambpp.2013.10886abstract