Queue simulator online: watch a line form, change the servers, see the wait
This is the same simulation engine the Simflowly app uses, running in your browser on one queue. Pick a bank, dock or checkout, change the number of servers and watch the line, the busy servers and the average wait update. A full check then runs 30 seeded replications and sets the result beside Erlang C.
How to use the queue simulator
- Choose a scenario: bank branch, dock doors or store checkout.
- Press play and watch the line and the servers: busy servers light up, waiting units queue to the left.
- Change the number of servers. The simulation restarts with the same seed, so the same customers arrive.
- Read the busy share, the line now and the average wait so far.
- Press the full check to run 30 seeded replications and compare the average wait with Erlang C.
Offered load is 3.0 servers of work, so 4 servers are 75% busy in the long run.
What is simulated
- Arrivals are random (Poisson) at the hourly rate of the scenario.
- Service times are lognormal with a spread of cv 0.8 for the bank and checkout and 0.5 for the docks, so they are more regular than the exponential service Erlang C assumes.
- One line feeds all identical servers; nobody leaves the line and nobody jumps it.
- Seed 11. The same customers arrive each time you restart, so you can compare 4 and 5 servers on identical demand.
The bank in numbers (formula, exponential service)
The bank has 40 customers an hour and 4.5 minutes at a teller, so the work is 3.0 tellers. Erlang C gives these averages for the number of tellers (exponential service, so the simulation above with its more regular service shows somewhat lower waits):
| Tellers | Busy | Chance of waiting | Average wait in line |
|---|---|---|---|
| 3 | 100% | 100% | grows without limit |
| 4 | 75% | 51% | 2.3 min |
| 5 | 60% | 24% | 32 sec |
| 6 | 50% | 10% | 9 sec |
| 7 | 43% | 4% | 3 sec |
Formula: Wq = C(c, a) ÷ (c·μ − λ), with a = λ ÷ μ. Press the full check in the simulator to compare with 30 replications of the engine.
What a simulator shows that a formula does not
A formula returns one number. A run shows the shape of the day: the line that builds for ten minutes and clears, then builds again. It also shows what the average hides: with 4 tellers the average wait is 2.3 min by formula, yet the picture is never an average: the line is empty for a while, then several customers arrive close together and it builds. Change the servers from 4 to 5 and watch both the average and the bursts shrink.
Questions
Is this a real simulation or an animation?
It is the real discrete-event engine from the Simflowly app, stepping through arrivals, service starts and service ends in your browser. The line, the busy servers and the numbers are read from the engine at each moment. Nothing is scripted, and the seed is shown so the same run repeats.
Why does the line grow when servers are almost enough?
Arrivals are random, so customers bunch up even when average capacity exceeds demand. With the bank at 4 tellers, work equals 3.0 tellers and they are 75% busy, yet a line still forms. At 100% busy or more the line never clears.
How is this different from the queue calculator?
The calculator gives exact averages for random arrivals and exponential service. This simulator lets you watch one run unfold, uses a more regular service time in each scenario, and runs 30 replications for an average with a 95% interval. Compare both to see how much the service-time spread matters.
Keep going
- Queue calculator (Erlang C): exact averages for your numbers
- Why the wait explodes near 100% busy
- One shared line or one per server
- Dock door calculator
- Erlang C vs simulation: when the formula understaffs
- How the engine was validated against queueing theory
Updated 20 November 2026. Free to use, no account. The calculation runs in your browser; nothing you type is sent to us. The engine behind the simulation checks is validated against exact queueing results.