Understanding of Experiment randomness

Hi

I have a question with regards to the random stream in Experiment.

I have repeat running the same set of experiment with a number of replication. I found the raw data that I have from each repeat run of the replication are the same. As I understand, the random stream is used during the replication and therefore the result in each replication is different. But why is the raw data between each repeat run of the replication are the same? Is the sequence of the random stream used in each replication the same when the replication is run again?

Thanks

FlexSim initializes random streams the same for a given replication across all scenarios. For example replication 1 will use the same random stream initialization in scenario 1, scenario 2, scenario 3, etc. This, however, will be different than replication 2, which gets its own stream initialization for scenarios 1, 2, 3, etc. This is a commonly used variance reduction technique. If you want to turn this off, you can check the box Initialize random streams based on system time in your global preferences, but I think most math wonks would tell you that's a bad idea.

Anthony,

I run the same experience twice. The raw data that I have from the first run is the same as the second run with respect to each replication. Therefore am I right to say that the random stream use for each replication for each re-run is the same.

Regards

David

should be correct David, it will allow you to replicate an Experiment run (great when bug fixing/chasing) and means that the same Random stream is used for Rep 1 in all experiments - good when you identify a particular stream is giving you an outlier.