Discrete Event Simulation

Arrivals and durations are drawn from probability distributions fitted to real data instead of being set to averages. That is the difference from a spreadsheet capacity calculation: queues, blocked aisles and idle stations arise from variability and interaction, and a model fed only averages does not see them. The output is throughput, queue lengths, utilization and order lead times over the simulated period. A detailed digital twin of a warehouse is a DES model kept in step with current data.

Discrete event simulation in practice

One simulation run is one random sample, not an answer. Two runs of the same model with different random seeds give different results. Say ten replications of the current layout average 2,180 order lines per hour with a standard deviation of 60. The 95% confidence interval of that mean is about ±43 lines (2.26 × 60 ÷ √10). A proposed layout that reaches 2,230 lines in one run has proved nothing yet, because a single run of the new layout carries the same spread of about 60 lines. Running both variants on the same random number streams and adding replications narrows the interval before anyone signs off on the investment.

The second trap is the warm‑up period. A model starts with an empty operation – empty buffers, no backlog, idle pickers – so the first simulated hours show more capacity than the real site has. Statistics are collected only after queues and stock levels settle, and the warm‑up is cut from the results. A report that skips replications and warm‑up presents the best case as the expected one.

Logio builds DES models of warehouses, production lines and material flows in dynamic simulation projects and uses them to compare layout, automation and staffing variants before the investment, as part of intralogistics and material flow design.

Related terms

• Digital Twin
• Bottleneck
• Capacity Planning
• Warehouse Automation