Task Sequence Delay for Multiple Capacity

Hi,

I would like to learn how task sequences should be set for the case that a TE needs to handle multiple flowitems simultaneously.

As a flow item enters the system, it generates and dispatches a TS that basically make some delay of TE(robot). The delay time is known and fixed (3 sec).

Since the capacity of robot is more than one, if two flowitems come into system at the same time, I expect the robot handles (delay) the two flowitem simultaneously. However, when I use TASKTYPE_DELAY for each flowitem, it dispatches sequentially and takes total of 6 secs, even though the robot capacity is set to 2.

Could anyone help me to write the TS to handle for this case?

I use V4.3 and attached my testing model for your review.

Thank you for your help in advance.
Sung

Sung,

It is usually best to try and use the Flexsim objects how they are intended to be used. So, I have modified your model to use the robot the way it is normally used. I raised all of the speeds to very high numbers so that you wouldn't have any extra time caused by travelling. I also removed all of the on entry and message code in the model.

Next, I added an on load trigger to send a delayed message in zero time if the content of the robot is one (the first item only). That message creates a preempting task sequence that has the 3 second delay that you wanted.

For testing I upped the max content to 3 and have 3 items arriving at time zero.

I hope this gets you what you want,
Brandon

Hi, Brandon,

Thank you for the explanation and modified model. Yes, it works fine for the model assumption that is described in my first post. However, I am sorry, I think I didn't put the assumptions or model description well enough for what I am looking for.

The flowitems in the system could arrive at different time points. In other words, one flow item arrives at time 0, but the second flow item arrives at 0.5 sec later. Or, as I described in the previous post, two flow items could arrive at the same time at time 0. The time delay for the robot process is still 3 secs.

When two (or more) flowitems arrives exactly at the same time, your code would work just fine. If there is a little interval between two flowitems, the later item still needs to wait until the first item is completed. What I want to do is though, the later part can be processed while the first item is being processed due to capacity of two or more of the robot (just like a processor with capacity two) The total processing duration for the example (0.5 interval) should be 3.5 sec.
Time 0: first item arrives, start processing,
time 0.5: second item arrives, check capacity, start processing
time 3: the first item is completed,
time 3.5 the second item is completed

I understand that it would be achieved if I use a Processor whose capacity is set to two. Unfortunately, the processor needs to be TE so that it can do other tasks (i.e. delivery or transport) for other items. The model that I attached is a made-up testing model to represent this capacity issue only. The original model has other stream of flow item that triggers other tasks on the robot.

Can you please help me again to achieve this?

Thank you again for your help.
Sung

Sung,

I am not clear on what you are trying to do. If you want the transport to always handle two items then use the queue to batch the items together. That way you can even add a maximum wait time. You can also slow down the transport so that it doesn't move so fast and this will make it pick up more than one flow item when it gets there. However, once it starts the delay task it won't pick up any more flow items because it is past the break task in the task sequence.

The general idea that you need to follow is that you create an extra task sequence that preempts the current task sequences with a delay and you only create this task sequence with the first flow item that the TE loads.

Brandon

Dear Brandon,

I am sorry for not clear description.

Attached please find a model that I tried to explain logic for what I want to achieve. I added a new model logic flow using a "regular processor" instead of "robot". That is a simple source - queue - processor - queue - sink logic flow. There is no waiting for the second flow item (arriving 0.5 sec later) due to the capacity of two for the processor. Completion time of two flow items is 3.5 (3 sec each parallel processing). Simply I want robot do the same thing. Robot is not used as a transporter for this particular flowitem (No kinematic or travel is necessary)

It could be simply done by replacing picture for the processor, only if all required processing of the robot is whatever the 3 sec's predefined processing at fixed position (Flow item type 1).

However, there might be a different flow item that requires actually being transported by the robot (Flow item type 2). Please note that the demo model only considers the flow items type 1 that requires processing rather than being transported. It is just because I want to focus and solve this parallel processing for flowitem 1 first.

Anyway, for this reason, I don't think I can use a regular processor for the robot just replacing the picture. The machine should be a "robot (TE)".

I hope this clarifies what I want to have.

Thank you.
Sung

Sung,

I can not think of any way to get a TE to behave as a processor with a content greater than 1. Because of the way task sequences work it would be almost impossible to pull this off. The only way I can think of would be to not give the object complete task sequences but to send messages to create partial task sequences to load and unload the items. However, doing this would not give you the states data that you need and would require even more work to try and get that right.

Being that you have two types of flow items, one that has to be processed like above and another that has to be moved. I would try to do this in one of two ways. The first way would be to use two different objects and the second way would be to use one processor and make the process time for the second item type include the travel time. In both cases I would use the on entry and on exit triggers to control the flow of the item types.

Good Luck,
Brandon

Dear Brandon,

I like your second suggestion. The first way might be easy, but later on, summary of one representative stats including utilization from two different objects might become tricky for me. With careful control of ports like you advised, I think the second idea will work just fine. However, I understand that using a processor object can not take advantage of built in TE transfer kinematic movement for the type 2. I believe we can certainly add kinematics for the processor object later, can't we? For now, I don't want to worry about the kinematics, and simply will take your second advice and use the processor object.

Thank you very much again for your help!

Sung