question

jacopo-r avatar image
1 Like"
jacopo-r asked jacopo-r commented

Problem with AGVs allocation

Hi,

I am working on a model for my master thesis and I encountered some problems that I am not able to fixed. According to my understanding they are all associated with path and control point allocation.
Can someone try to help me while keeping the model as private? Is there a specific place where I should upload the model?

PS: I am also using Cplex on python for running a dynamic tasks dispatching algorithm, however the problems exists also while running the model without the Dynamic algorithm (so withouth the python connection).

Thanks in advance

FlexSim 23.1.2
agvallocation
· 4
5 |100000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 12 attachments (including images) can be used with a maximum of 23.8 MiB each and 47.7 MiB total.

jacopo-r avatar image jacopo-r commented ·

UPDATE:

I just made this post public and I will now share a new version of the model where the "Assign task" code has been changed a bit. However, the model still creates problems with deadlock and wrong allocation with some replication when running the model manually for many times with random streams.

04_All_Cases_Model_1_16.fsm

python-cplex-model.py

1 Like 1 ·
04-all-cases-model-1-16.fsm (3.7 MiB)
python-cplex-model.py (2.2 KiB)
Kavika F avatar image Kavika F ♦ jacopo-r commented ·
Hey @jacopo_r, have you looked at the " Preventing Collisions and Deadlock" section of building AGV Network Logic? There are some helpful guides that may help you eliminate your deadlocks. For example, avoid using accumulating paths and instead have AGVs stop on actual Control Points. If you want to use accumulating paths, then we can figure out a work around using Control Areas and such to ensure deadlocks won't happen.


I'll continue looking into the issue.

1 Like 1 ·
Kavika F avatar image Kavika F ♦ Kavika F ♦ commented ·

@jacopo_r, after turning all paths to "No Accumulation", there were no more deadlocks for the entire run; however, this probably affected the spacing and timing of events. You would need to add more control points to ensure the desired spacing at intersections, allocations, etc. Here's the modified model where I programmatically changed all paths to non-accumulating.

attempted_fix_1.fsm

1 Like 1 ·
attempted-fix-1.fsm (3.7 MiB)
Show more comments

0 Answers

·

Write an Answer

Hint: Notify or tag a user in this post by typing @username.

Up to 12 attachments (including images) can be used with a maximum of 23.8 MiB each and 47.7 MiB total.