Queuing (see Business Analytics Module D) is always a popular topic with students–and evidently with readers of the Wall Street Journal (Oct. 7, 2016) as well. This Journal article is a very basic tutorial on the history (Erlang discussion) and logic of waiting line modeling.
The piece writes: “You’ve probably participated in this familiar dance: Given a choice of checkout lines, you’ve somehow picked the slowest. You could wait it out. You could chassé to another queue. Or you could bail out altogether. After all, no one likes to wait. But are the other lines really faster? When parallel lines feed multiple cashiers, you may not be in the slowest one, but chances are, you also are not in the fastest.”
Prof. Bill Hammack, at the U. of Illinois (YouTube’s “Engineer Guy”), explains it like this in his 4-minute video: “Imagine three lines feeding three cash registers. Some shoppers will have more items than others, or there may be a delay for something like a price check. The rate of service in the different lines will tend to vary. If the delays are random, there are six ways three lines could be ordered from fastest to slowest—1-2-3, 1-3-2, 2-1-3, 2-3-1, 3-1-2 or 3-2-1. Any one of the three (including the one you picked) is quickest in only two of the permutations, or one-third of the time.”
Classroom discussion questions:
1. Why doesn’t every service provider use the multiple-server, single line approach?
2. Explain Erlang’s theory.
Queuing (see Business Analytics Module D) is always a popular topic with students–and evidently with readers of the Wall Street Journal (Oct. 7, 2016) as well. This Journal article is a very basic tutorial on the history (Erlang discussion) and logic of waiting line modeling.