EVPI (Expected Value of Perfect Information)
(16-04-2020, 12:49 PM)Pernille.S Wrote:
(14-04-2020, 11:26 PM)Deju Wrote: Dear all,

Is it realistic to have a negative EVPI if the investments prior to uncertainty resolution is similar for both perfect information and  stochastic scenarios?

Any suggestions is much appreciated.


Dear Dejene, 

As I undertstand, EVPI is in theory larger or equal to zero. 

But more interesting, why are you calculating this value? What is your uncertainty? In my opinion, EVPI has a limited practical value in long-term TIMES models.


Dear Pernille,

Thanks a lot for your help and also for your contribution in the forum. I also have the expectation that EVPI should non-negative but i was getting it the other way. Now i have managed to figure out my mistake, it was just inconsistent scenarios formulation. 

Yes, its practical value is comparable with that of the system cost. Nevertheless, I'm calculating it to compare the impact that each uncertain parameter has in mitigating wrong decisions prior to resolution periods (hedging stage). Also, to cross check the healthiness of my model as i did here. I got a negative value when it should always be non-negative.

I'm looking at current off-grid power demands' accessibility to strong grid and storage battery cost development as long term uncertain parameters in my model. 


Messages In This Thread
RE: EVPI (Expected Value of Perfect Information) - by Deju - 16-04-2020, 02:18 PM

Forum Jump:

Users browsing this thread: 1 Guest(s)