Checkout & Drive-Through

Customers arrive at random and wait in line: compare separate lines, one line, an express lane, drive-through windows and venue gates.

  • Waiting
  • Being served
  • Leaving
  • Server open
  • Server closed
100%

Customers arrive at random, wait in line and are served; the counter compares the waits of each layout. Tap a cashier or gate to close or open it; Rerun plays the same arrivals again with your changes.

Setup
Scene
Layout
4 cashiers
160 per hour

How is Checkout & Drive-Through working for you?

About Checkout & Drive-Through

Scenes
Store, drive-through, venue gates
Exact waits
Erlang C (M/M/c) and M/M/1
Little's law
L = λW (Little 1961)
Clock
Event by event, in seconds

Checkout & Drive-Through is a queueing simulation: customers arrive at random, wait in line, are served and leave, and the tool keeps the books on every wait. Each event, an arrival or the end of a service, is taken from a time-ordered queue of events, so the clock jumps from one to the next and every wait is computed exactly. The customers on the canvas walk to their place in line, so the queue looks the way it is.

Three families of layout share the model. In a store, each cashier can have a separate line, which customers pick by length and leave for a line two shorter, or one line can feed every cashier, so nobody is served out of turn; an express lane serves only customers with few items. A drive-through runs order point, pay window and pickup window in series, with room for only a few cars between them, so a slow window holds up the windows before it. Venue gates send people through security lanes and then turnstiles while arrivals rise toward kick-off.

Where exact results hold, the chart draws them beside the measured mean wait: one line to c cashiers with exponential times is an M/M/c queue, whose wait follows Erlang's C formula, and separate lines joined at random are c M/M/1 queues. Little's law, that the mean number in line is the arrival rate times the mean wait, holds in every scene. Tap a cashier or gate to close it, compare layouts in the Runs table, and export a PNG, video, GIF or CSV.

How to use Checkout & Drive-Through

  1. Pick a scene. Choose Store, Drive-through or Venue gates, then a Layout: separate lines, one line or an express lane in a store; one lane or two order lanes at a drive-through.
  2. Set the traffic. Set Servers from 1 to 8 and Arrivals from 10 to 400 per hour. Above one customer per mean service time per server, the lines grow without end.
  3. Watch the lines. Blue customers are waiting, yellow ones are being served and gray ones are leaving. The counter shows the time, the customers served, the mean and longest wait, and the share served before someone who arrived earlier.
  4. Compare with the exact wait. With Service spread on Exponential, one line in a store draws Erlang C's exact mean wait on the chart; separate lines with Line choice At random and Switching off draw the M/M/1 wait of each line.
  5. Close a server. Tap a cashier or a security gate to close or open it; its line joins the others. Rerun plays the same arrivals again with your changes, Restart draws new arrivals, and Reset returns every setting to its default.
  6. Compare runs. After one hour of model time each run adds a row to the Runs table: the mean and longest wait, customers served per hour, the share served out of order and the mean time from arrival to leaving.
  7. Export. Save a PNG, a video or GIF, or record the first 30 seconds again. The CSV files hold the mean wait every 30 s of model time and the Runs table.

Frequently asked questions

Is one line faster than separate lines?

For the same cashiers and customers, one line feeding every cashier gives a shorter mean wait and serves everyone in order, because no cashier sits idle while another line waits. With 4 cashiers, 160 customers an hour and a mean service of 75 s, Erlang's C formula gives a mean wait of 74 s for one line against 375 s for four separate lines joined at random.

What is Little's law?

The mean number of customers in a line equals their arrival rate times their mean wait, L = λW, for any queue that settles down, whatever the arrival or service times. John Little proved it in 1961; the tool's readout lets it be checked in every scene.

Why do drive-through lanes back up?

Because the windows are in series with little room between them: when the pickup window is slow, the cars waiting for it fill the space, the pay window cannot send its car on, and the order point stops too. Once the lane before the order point is full, arriving cars drive on.

Are the drive-through times real?

No. The order, pay and pickup shares of the service time and the room between windows are model settings; the published drive-through studies could not be opened when this tool was built.

References

  1. Little, J. D. C. (1961). A proof for the queuing formula: L = λW. Operations Research 9(3), 383–387.
  2. Kendall, D. G. (1953). Stochastic processes occurring in the theory of queues and their analysis by the method of the imbedded Markov chain. The Annals of Mathematical Statistics 24(3), 338–354.
  3. Larson, R. C. (1987). Perspectives on queues: social justice and the psychology of queueing. Operations Research 35(6), 895–905.
  4. Rothkopf, M. H., & Rech, P. (1987). Perspectives on queues: combining queues is not always beneficial. Operations Research 35(6), 906–909.