question

@Sudheer R

If I understand you correctly, you want a Task Executor to automatically know how many items are "missing" from the queue and get that many and load them.

You could create a model where Process Flow creates a token when an item exits a queue and assigns a Task Executor to go pick up a flow item and put it in the Queue that you want to keep full, but you cant dynamically change the amount a Task Executor is supposed to pick up once a task has been assigned. In other words: once the Task Executor has been assigned to pick up 1 flow item, you cant tell it to pick up 2.

The other issue you would run into by doing this with Process Flow is the trigger to tell the Task Executor to pick up a flow item would be an item leaving the Queue you want to keep full. Those items do not leave the Queue in the same instance, and therefore would assign work to the task executor in two separate instances.

I have 2 ideas about how to make this work. They depend on how your model is going to work. They will both take a bit of time to create.

If you have items leaving the queue and you can guarantee that the TE that is loading the queue will be able to re fill the queue before more is taken, you can have a variable that represents the amount that is pulled from the queue, and assign the task executor to pull that many from the queue. You would do this by pushing all the items in the queue before the queue you want to keep at 3 to a list, and having the Task Executor pull that many items from that list and load. Again, this will only work if you know the amount leaving the queue, and the TE can fill it before it leaves.

The other way I can think of is more flexible, but has more potential to go wrong. You would push all of the items in the Queue before you the queue you want to keep full to a list. You would generate a token every time an item leaves the queue you want to keep full. You would also have a global variable that increments every time an item leaves the queue, and decrements every time an item is reintroduced to the queue. You could also conceivably make this variable a pointer to the Queues content minus 3. Once you generate a token, you would have a decide with this criteria: if the Task Executor is acquired, or the content of the queue is 3, sink the token. Otherwise, you would acquire the Task Executor, have them travel to the queue, check the number in the global variable, pull that many items from the list and save them in an array in a label, load them, and unload them in the queue you want to keep full, and then release the TE and sink.

This solution has alot of potential to go wrong, so if you can I would implement the first. Feel free to reply if you have questions about these solutions since they are a bit complicated.

This is the default logic of the standard Queue. If the queue's capacity is 3, and it has 1 in it currently, it will receive 2 simultaneously.

queuecapacity3.fsm