Sorry for all the questions, and thx for the help so far !
I would like to perform a design of experiments to determine the server utilization versus a specific component mix. So changing the amount of product that enter in a specific time period and measure the server status. Since its ad DOE the amount of possibilities is rather large because I would like to vary 10 types of sources over 6 steps.
Is it possible to script this event and use ED to determine the server status without having to change all the sources manually ?
Thanks again !
Design of experiment
Re: Design of experiment
To clarify my problem what I think could fix my problem.
Would it be possible to perform the following function:
- use a 4D script to read the value from a excel file (I know how to do this),
- Execute an experiment to determine the server utilization,
- alter the 4D script to ad +1 to the value of the row indicator (so that the sources have a different interval time)
- Execute the experiment again,
- Loop until the the experiment count is equal to the amount of possibilities
Thanks !
Would it be possible to perform the following function:
- use a 4D script to read the value from a excel file (I know how to do this),
- Execute an experiment to determine the server utilization,
- alter the 4D script to ad +1 to the value of the row indicator (so that the sources have a different interval time)
- Execute the experiment again,
- Loop until the the experiment count is equal to the amount of possibilities
Thanks !
Re: Design of experiment
Hello Steven,
You could do the following:
- In the experiment wizard use the On end of run trigger to increase a value in a table (e.g. store the row indicator that you mentioned in this table).
- Run the experiment and instead of using the average results, select the option 'Show raw data in report'. This will give you the results for each individual run.
- This will give you a quick scan of the effects of your settings in the separate runs. It could be necessary to do some of the runs again, now without the end of run trigger and multiple runs for 1 setting.
Another option could be to use the Scenario Manager (you will find it in the tools menu). You could use the scenario manager to run multiple scenario's after each other automatically. However, you will have to save a model file for each individual scenario first.
Regards, Marlies
You could do the following:
- In the experiment wizard use the On end of run trigger to increase a value in a table (e.g. store the row indicator that you mentioned in this table).
- Run the experiment and instead of using the average results, select the option 'Show raw data in report'. This will give you the results for each individual run.
- This will give you a quick scan of the effects of your settings in the separate runs. It could be necessary to do some of the runs again, now without the end of run trigger and multiple runs for 1 setting.
Another option could be to use the Scenario Manager (you will find it in the tools menu). You could use the scenario manager to run multiple scenario's after each other automatically. However, you will have to save a model file for each individual scenario first.
Regards, Marlies
Re: Design of experiment
Marlies,
Fantastic suggestion !!!
After finding the 4D script to alter a table value i got it working !
This will safe me so much time !!
Thanks !!
Fantastic suggestion !!!
After finding the 4D script to alter a table value i got it working !
This will safe me so much time !!
Thanks !!