Monte-Carlo mean included in expressions

Svilen,

Assuming you have an existing expression for the particular measurement you want, say called existing_measurement, you can add another output and set the evalType column to "sweeps" (this will measure across the Monte Carlo sweep). The expression can then be:

average(existing_measurement ?continuous nil ?overall t)

The ?overall t should get added automatically to the expression when you change the evalType to sweeps.

Andrew

Andrew, thanks for replying. Let's see if I understood what you are saying.

I have defined an expression Gain for which I want to obtain the mean value out of MC simulation.
If I understood well, you are saying I should define another expression, say Gain_avg, for which I set the evalType to sweep and define that Gain_avg as

average(Gain ?continuous nil ?overall t)

Have I got it right?

Svilen

Svilen64 said:

Have I got it right?

Yes, spot on.

Andrew

Svilen64 said:

Have I got it right?

Yes, spot on.

Andrew

Svilen64 I assume you spotted the typo (I was about to point it out) and then deleted the post - presumably you got it working.

Andrew

Yes, I spotted the typo and deleted the message because of that, thinking that was the reason it was not working. But even after correcting it, I still couldn't get it to work. Simulations go fine, it just doesn't evaluate that specific expression at the end.

Svilen

I need to correct myself. It works but the result is not displayed in the yield summary. I have to switch to detail view to see it.
Now comes another question. Can I combine in a new expression that average result with any of the other expressions' results that are not of the sweep evalType? For example to calculate the percentage variation wrt average.

At the moment, all you can do is add another "sweeps" expression and compute (say) Gain-Gain_Avg (or some similar computation). If you do this, you'll get a waveform you can plot which will plot the computation versus the Monte Carlo iteration; you just can't get the numbers into the output table (in other words, you can't have a Point expression that refers to a Sweeps result - it just doesn't work properly if you do this - I just get 0 for the difference in all points). I even tried a few newer approaches using Measure Across Any - but this didn't work either.

I can see this would be useful from time to time, so if you want this you should contact customer support and request an enhancement (if you do so, please let me know the case number as I'd be interested in tracking it and encouraging R&D to implement it).

Regards,

Andrew