chiaha-SLXFor classes · Materials

chiaha-SLX Lab 3: The tool crib

Time: 45–55 minutes
You need: a browser, a chiaha-SLX sign-in, and the M/G/1 priority formula (Cobham's formula) for the last part.
Open it: https://slx.chiaha.com/slx?open=deck/tool-crib

Learning goals

The problem

A factory tool crib has one clerk. Two kinds of mechanics come to the window:

Arrive every Clerk takes Priority
Class 1 420 ± 360 seconds 300 ± 90 seconds 1
Class 2 360 ± 240 seconds 100 ± 30 seconds 2

All times are spread evenly over their range. A waiting mechanic with the higher priority number is served first; within a class, first come first served. A mechanic already at the window is never interrupted. The crib is modeled for one 8-hour shift, 28,800 seconds.

Find these lines in the model:

Line in the model What it means
shift = 28800 the run length, in seconds
poisson = FALSE TRUE switches to random (exponential) arrivals, used in tasks 4 and 5
rv_uniform(Arrive1, 60.0, 780.0) priority = 1 class 1 arrivals and their priority
rv_uniform(Arrive2, 120.0, 600.0) priority = 2 class 2 arrivals and their priority
rv_expo(Arrive1, 600.0) priority = 1 and rv_expo(Arrive2, 400.0) priority = 2 the same, for the random-arrivals version
departed1, waited1, departed2, waited2, busy mechanics who reached the window, their total seconds in line, per class; seconds of service

Measures

Tick Show final values and set Runs to 5. For each run:

Average each over the five runs.

Before you run

Q1. How busy should the clerk be? Add up each class's share: mean service time ÷ mean time between arrivals. What does that tell you to expect?

Tasks

Keep Seed at 1 and Runs at 5.

  1. As written. Class 2 (the short jobs) goes first. Run and record the four measures.
  2. Swap the priorities. In the two rv_uniform(...) arrival lines, give class 1 priority = 2 and class 2 priority = 1. Run and record.
  3. No priority. Set both to priority = 1, so the clerk serves in arrival order. Run and record.
  4. Random arrivals, long run. Open the original again (Model button) or undo your edits. Change poisson = FALSE to poisson = TRUE and shift = 28800 to shift = 2880000 (100 shifts). Arrivals are now exponential with means 600 and 400 seconds; service is unchanged. Run and record.
  5. Random arrivals, class 1 first. Keep task 4's edits and swap the priorities in the two rv_expo(...) lines (class 1 priority = 2, class 2 priority = 1). Run and record.

Questions

Q2. Compare tasks 1, 2 and 3. For each class, which order is best and which is worst? Which order gives the lowest mean wait over all mechanics?

Q3. If every mechanic's time costs the factory the same, which order should the crib use? Explain in one or two sentences why serving the short jobs first helps.

Q4. Utilization barely changes between tasks 1, 2 and 3. Why doesn't the order change how busy the clerk is?

Q5. Tasks 4 and 5 are an M/G/1 queue with two non-preemptive priority classes. Use Cobham's formula to compute each class's mean wait for both orders, and compare with your runs. (For a uniform time on [a, b], E[S²] = (b − a)²/12 + mean².)


The answer key is shown to the course's instructor when signed in.