Queueing simulation

This page can be used to conduct a stochastic simulation of a queue, or a system of several queues.

1. Arrival process.

Inter-arrival time distribution   Initial routing to

2. Queueing.

Queue name  Capacity  #servers  Service time distribution  Onward routing to


3. Simulation design.

Number of runs:
Run duration: for warmup, and for data collection.
Random seed:

4.


Notes.

  1. Microsoft Internet Explorer does not work well with this page! Sorry. The Firefox browser is a good alternative.
  2. Larger simulations will take a while. This is a client-side script, so the time it takes will depend on the performance of the Web browser you are using, and of the computer it is running on. Some browsers will annoy you with messages to the effect that “the script is taking too long”; turn these off if at all possible.
  3. The routing boxes can take the name of a single queue, or the names of several queues (separated by commas). With random routing, the probabilities are entered the same way. With join-shortest-queue routing, ties are broken by sending customers to the queue that appears first in the list - this can introduce asymmetry into a system that otherwise looks symmetric.
  4. Simulations of different queueing systems using the same random number seed will have similar elements (e.g. service times) where possible. This makes it possible to compare the systems by the method of common random numbers.
  5. The system starts out empty (no customers) at the beginning of each run.