Problem with branch probabilities
Brought to you by:
danieledipompeo,
g_serazzi
In the attached model, part of the examples distributed with JMT, the fork sends 2 jobs to Queue 1 and 1 job to the others. If plotting response times it looks fine, but plotting queue length at Queue 1 shows that it is unstable and grows linearly, even though the utilization at that station is only 80%. It may be possible that the join is not waiting for all 4 tasks, only for 3, so one remains at Queue 1.