question

Hey @Osama A6, let me point you to some resources that may be helpful in getting your workshop working.

You mention that your operators work in shifts. You can create a Time Table that outlines the start times, end times, and breaks for each shift.

You can add 4 operators per machine and then attach each operator to a different schedule so your operators can rotate working on the machine or taking breaks/going off shift. You can also use this if you have scheduled downtime on a machine that you know will happen at the same time each day/week (for example, M1 goes down 2-2:15PM on Tuesdays for scheduled cleaning). Above is an example Graveyard shift where an Operator will start at midnight, have a break from 03:15 to 03:45 AM, and continue their shift until 06:00 AM. The purple block denotes they're in a "Lunch" down state while the red denotes "scheduled down". You can adjust these states and colors as you like.

---

If you don't know when your machines will break down or you want some variability in downtimes, then you can use an MTBF/MTTR schedule.

This schedule lets you specify around what time the first failure will occur, for how long, and then how long the machine will continue to be up until the next failure. We don't necessarily have a way to input that a machine will be up "90%" of the time, but you can adjust your times and statistical distributions to get close to a desired amount of uptime/downtime ratio. For example, if you set your uptime to be 900 seconds and your downtime to be 100 seconds you'll get an exact ratio of 90% uptime; however, it would be more realistic to set a distribution (like the one in the image) that says uptime is about 1000 seconds and downtime is around 50 - 100 seconds.

---

For your processors, if your processing times are constant, it will be simple to calculate the setup time and input them respectively. However, if your process times are a distribution, then you may have to do something different. You can use an Object Flow (a type of Process Flow), attach each Machine to it, and then have it calculate the necessary process and setup times.

Here's an example Object Flow and example custom code to set the times.

/**Custom Code*/
Object current = param(1);
treenode activity = param(2);
Token token = param(3);
treenode processFlow = ownerobject(activity);

double processTime = triangular(10.0, 20.0, 15.0, getstream(current));
double setupTime = processTime / 10;

// print("processTime:", processTime, "\tsetupTime:", setupTime);

current.setProperty("SetupTime", setupTime);
current.setProperty("ProcessTime", processTime);

When an item arrives at a Processor, a token is created at the "Item Arrives" activity. It moves down to the custom code to calculate and set the processing time (currently a triangular distribution) and the setup time (a tenth of that) to the current processor attached to the Object Flow.

I hope this helps you get started with your model! 3machine-no-shifts_1.fsm