hurol Posted February 13, 2015 Share Posted February 13, 2015 I constructed a simple model with a single server where I used a rate table for arrivals. The server block capacity was fixed at 2. When I run this model by hand, I see that maximum holding time for the input buffer of the server is just about 24 minutes every time (but the same exact number), just as it would be if the arrival rate was constant throughout a 24-hour period. However, when I run an experiment where I defined the Server1.InputBuffer.Contents.MaximumTimeWaiting as the response variable, the result is 40 minutes for the scenario where the server capacity of 2 is 40 minutes.(I am using version 5.91) I am wondering 1) if I am confusing two different quantities (maybe the maximum holding time result is averaged by dividing by capacity, but that would be a bug, because the capacity should have no effect on the input buffer) or 2) if the randomization seed is the same for every run that I do by hand, which must be the case, because when I reduce the required replication count to 1 for the experiment, I get the same exact result for the single run that I do by hand. If that is the true reason, then I must ask where to change the randomization seed for every run. Link to comment Share on other sites More sharing options...
gocken Posted February 14, 2015 Share Posted February 14, 2015 The problem is not related to randomization seed. I think, in experiment window you do 10 replications for each run by default. In the interactive run (you call this by hand) you have to do only one replication for each run (you are forced to do like this for some unknown reasons???). This is it, it is the way simio take. A number of users asked for this property (being able to do more than one replication in each run in interactive mode), but it is not considered yet... Link to comment Share on other sites More sharing options...
hurol Posted February 17, 2015 Author Share Posted February 17, 2015 Thank you gocken, but if the problem is not randomization seed, why do I get the same results for, say average or maximum waiting time, from every interactive run? For that specific example model, my intention was to show a single run could give different results for the waiting times but running an experiment with replications could give more reliable estimates. Link to comment Share on other sites More sharing options...
MikeKk Posted February 18, 2015 Share Posted February 18, 2015 The Facility view of the model has a random seed number which you can change from the Run Tab, under Advanced Options>Randomness>Replication Number. This number corresponds to a scenario's replication number in the experimenter. If you export the "Details" csv From the response results tab of the experiment you can inspect each replications returned value for any of your responses As long as no changes have been made to the model, and the properties from the scenario are input identically, using this replication number, you should be able to have have the facility view of the model return the same value as the experiment allowing you to inspect particular experiment runs in the 3D view. Link to comment Share on other sites More sharing options...
Recommended Posts