chiaha-SLXFor classes · Materials

chiaha-SLX course pack

chiaha-SLX is a simulation language you write and run in the browser at slx.chiaha.com/slx. This pack is for a course that teaches discrete event simulation through code: queues, servers, priorities, breakdowns, customers who give up. It has five labs, one per classic problem. Each lab opens a working model with one link, then has students change it, run replications and compare what they see with queueing theory.

What's included

Lab Problem Ideas Class time
Lab 1: The barber shop One barber, customers every 18 ± 6 minutes, haircuts 16 ± 4 Reading a model, replications, run length, utilization 40–50 min
Lab 2: Three bank tellers M/M/3 lobby Erlang C, warm-up, adding and removing a server, an unstable queue 45–50 min
Lab 3: The tool crib One clerk, two classes of mechanics, one served first Priorities, who pays for them, Cobham's formula 45–55 min
Lab 4: A machine that breaks down Jobs on one machine with failures and repairs Availability, frequent short stops against rare long ones 40–50 min
Lab 5: Customers who leave One barber; customers leave if they wait too long Reneging, patience against speed, the M/M/1 check 40–50 min

The labs stand alone. Lab 1 is the gentlest introduction to the language; Labs 2 and 3 assume students have seen M/M/c and M/G/1 formulas, or are seeing them now. Each lab starts with the student part. The instructor part follows it: the answer key, where students get stuck, and a re-run record. Staff and signed-in instructors of an approved course see it; students and visitors see only the student part. (In the source files, the instructor part starts at the line marked INSTRUCTOR ONLY.)

How a lab works

Every lab follows the same pattern:

  1. Open it. A link such as https://slx.chiaha.com/slx?open=deck/barber opens the model in the workspace. The original GPSS deck is in the comment at the top; the SLX below it is what runs.
  2. Run it as written with Runs set to 5 and Show final values ticked. The output prints a report and the model's final values for each run, one after another, each under a --- seed line.
  3. Change one thing in the editor (a time, a capacity, a priority, the run length) and run again. Each task names the exact line to change.
  4. Compare the average of the runs with the earlier case and, where a formula exists, with theory.

Students compute measures from the final values the model prints, for example mean wait = waited ÷ served. Each lab says which ones. A spreadsheet or a calculator is enough.

Runs and plans

Every task is written for 5 replications, so it works on either plan. Each lab needs 5 or 6 runs, so a Free account can do one lab a day with room to spare. Free runs are also counted per network: 10 a day from one network address, however many accounts use it. A class working on the same campus network should be in a registered course, so everyone has the Class plan. The answer keys give the 5-run numbers students should match and, where it helps, 25-run numbers with a 95% confidence interval for discussion.

Students' edits are saved privately in their account (up to 3 models on Free, 50 on Class). If a student wants the original back, the Model button opens it again.

About the numbers in the answer keys

Every number in the keys came from running the model in the chiaha-SLX sandbox, the same runner the workspace uses, with the workspace's defaults: first seed 1, 5 runs unless stated, random-number generator sdidb. Each key says exactly which line was changed. A student who makes the same edit and keeps seed 1 gets the same final values, to the last digit. Measures are shown to four decimals; they are averages of the per-run values.

Theory values are stated next to the simulation results wherever a formula exists (M/M/1, Erlang C, availability, Cobham's formula), with the long run used to check them.

If the chiaha-SLX runtime is updated, results can shift slightly. Re-run the cases in the re-run records before the term starts if you rely on exact values.

Questions or changes

Write to info@chiaha.com.