How many dock doors do you need? Calculator with a peak-hour simulation

Multiply trucks per hour at your peak by the hours each truck holds a door: that is how many doors are busy on average. You need more than that, because trucks arrive in clumps. Enter your peak, dwell time and what happens when every door is busy, and compare Erlang B, Erlang C and a simulation.

Doors busy on average2.24 trucks/h × 33 min, the cross-dock template
Template's 2 doors111% loaddemand above capacity: the yard queue never clears
With 3 doors12.9 minaverage wait, simulated, 74% busy

Your peak hour

How much does that time vary?
When every door is busy

Starting values are an example, not data about your site. Use the busiest hour you actually plan for, not the daily average. The target is your business choice, not a standard.

Result

Doors to start with6 Erlang C
Doors busy on average3 6/h × 0.5 h
Busy share at 6 doors50%
Doors each method asks for
  • Average load only4 doors
  • Erlang B6 doors
  • Erlang C6 doors
Doors by method, with the formula result and the simulated result for each
MethodDoorsFormula saysSimulation, your dwell spread
Average load only
fewest doors that are not overloaded
450.9% wait, avg 15.3 minpress the check
Erlang B
at most 10% of trucks turned away
69.9% wait, avg 59 secpress the check
Erlang C
at most 10% of trucks wait
69.9% wait, avg 59 secpress the check

Highlighted row is the one that matches how your yard works. Formulas assume exponential dwell times (the “high” setting). Simulation uses your dwell spread, lognormal, and random arrivals at the hourly rate you entered. The yard is the doors' waiting line; door types, appointments and drivers who leave are not modelled.

Simulate this in Simflowly

Quick answer

doors busy on average = trucks per hour × hours per truck = 4 × 0.56 = 2.22 doors

With that many doors you are 100% busy and the yard never clears, so you always need more. How many more depends on how much waiting (or turning away) you accept and on how much dwell time varies. The calculator above does the arithmetic for three methods.

Three methods, one answer

  • Average load. Enough doors that load is below 100%. It is a floor, not a plan.
  • Erlang B. Counts the share of trucks turned away when every door is busy and no truck waits. It depends only on the average dwell, not on its spread.
  • Erlang C. Counts the share of trucks that wait when the yard is unlimited. It assumes exponential dwell times, which is a very wide spread; real dwell is often tighter, so it tends to overstate waits.
  • Simulation. Runs trucks through the doors with your dwell spread and arrival rate, with seeded random numbers so the result repeats. It is the only one of the three that can also show the 95th percentile wait.

Worked example: the cross-dock template

Simflowly's cross-docking template has 4 trucks an hour and a receiving dwell that is triangular, 20 to 50 minutes with a most likely 30, so 33 minutes on average. That is 2.22 doors busy on average. The template gives the receiving area 2 doors, which cannot keep up. These numbers are illustrative inputs, not data from any real dock.

Receiving doors in the cross-dock template at 2, 3 and 4 doors: busy share, wait and queue
DoorsBusyAverage wait95th percentile waitLongest line seen (avg of runs)
2111% loadoverloadedn/a105+
374%12.9 min52.7 min8.7
456%2.5 min16.8 min5.5

With 2 doors the wait is not a steady-state number: it only grows with how long you run (the queue reached about 105 trucks within 10 days), so none is shown. Seed 7, 20 replications of 10 days, first day discarded, random arrivals, triangular dwell. Going from 3 to 4 doors takes the average wait from 12.9 min to 2.5 min, at 74% and 56% busy.

Peaks: why a daily average hides the wave

Take the same trucks per day and the same 3 doors, but bunch the arrivals into an early-afternoon wave instead of spreading them evenly. The wave below is an illustrative shape, not data: it peaks at 2.39 times the average rate and is scaled so the daily total is identical (mean multiplier 1.00). It uses the hourly arrival profile that Simflowly sources support.

Same trucks per day, flat against wave arrivals, at 3 doors
Arrival pattern, 3 doorsBusyAverage wait95th percentile wait
Even across 24 hours74%12.9 min52.7 min
Afternoon wave, same total73%2.1 h5.2 h

Hourly multipliers used: 0.24, 0.24, 0.24, 0.24, 0.24, 0.36, 0.72, 1.07, 1.43, 1.67, 1.91, 2.15, 2.39, 2.39, 2.15, 1.67, 1.31, 0.95, 0.72, 0.60, 0.48, 0.36, 0.24, 0.24 (hour 0 to 23). Daily averages of busy share look similar; the waits do not. That is the reason to size on the peak hour and to simulate the day. A staffing or door schedule that changes by the hour is not modelled: doors stay constant.

When this model is wrong

  • Appointments. Scheduled arrivals are far more regular than random ones, so waits are lower than shown. Lower the arrival variability in a full model instead of using this page.
  • Door types. Reefer, hazmat or oversize doors are not interchangeable; one pool of identical doors is assumed.
  • Drivers who leave. Trucks that give up after a long wait are not modelled. A yard limit is the closest stand-in: trucks above it are lost.
  • Dwell depends on the crew. A fixed dwell is assumed. If unloading speed depends on labour, model the labour as its own station.

Questions

How many dock doors do I need for X trucks a day?

Start with the peak hour, not the day. Peak-hour trucks multiplied by the hours each truck holds a door gives the doors busy on average. Then add doors until the chance that a truck waits (or is turned away) meets the target you choose. A daily average understates the peak, so the answer for a day total is only as good as the shape of that day.

Why is the average not enough?

Trucks bunch up, and the line grows faster than it shrinks as doors approach 100% busy. In the worked example below, the same trucks per day give an average wait of 12.9 min on a flat day and 2.1 h on a day with an early-afternoon wave, at the same average busy share.

Erlang B or Erlang C for dock doors?

Use Erlang B when a truck that finds every door busy leaves and is lost. Use Erlang C when it waits in the yard without limit. Most yards queue, so Erlang C is the usual start, and a yard with a hard limit sits between the two: the calculator handles that with a finite-yard formula and the simulation.

Should inbound and outbound share doors?

It depends on your timing and layout, and there is no universal ratio, so this page does not quote one. Model inbound and outbound as separate stations in Simflowly and compare the two layouts.

What does this model leave out?

Door types, appointment schedules, drivers who leave after waiting too long, detention cost and labour. Dwell time is one average with a spread, and arrivals are random at the hourly rate you enter. If those matter, treat the result as a first cut.

Is the result free to open in Simflowly?

A single-station model like this fits the Free plan, which simulates up to 5 stations. You need a free account to save it.

Keep going

Updated 17 November 2026. Free to use, no account. The calculation runs in your browser; nothing you type is sent to us.