Veda2.0 Released!


Disabling a process in a region
#1
Dear All,

I am aggregating regions in a multi-region model. 
I first test how I could disable a process in a region specific VT file. I included a dummy process (which is not supposed to exist in this region) and used PRC_NOFF to in BY_Trans to disable the process from this region but this changes the objective function. I also tested ACT_BND (UP) = 0, still the objective function changes. 
Could you help ?

Many thanks,
Reply
#2
If the value of the objective function changes after setting ACT_BND (UP) = 0, that would mean that you do have some activity for the process when you don't impose that bound. I cannot see how else the activity bound would affect the objective function, unless it is actually causing a change in the activity level.
Reply
#3
(24-09-2018, 12:25 AM)Antti-L Wrote: If the value of the objective function changes after setting ACT_BND (UP) = 0, that would mean that you do have some activity for the process when you don't impose that bound. I cannot see how else the activity bound would affect the objective function, unless it is actually causing a change in the activity level.

Dear Antti,
Thank you for your reply. I checked and this is not the case (there is no activity without imposing a bound)
What I am trying to figure out is how to disable some processes from certain regions in a multi-region model. In the SubRES we can do this using TFM_AVA however there is no such possibility for technologies defined in the BY templates.
Reply
#4
Ok, then I am totally confused. Are you really saying that the activity is originally zero, but when you set a zero bound for it, the value of the objective function changes? That would be quite inexplicable to me.

Of course, if you have a reproducible case that you can share, I would be happy to look at it.  Otherwise, please post the explanation to this anomaly if/when you find it!
Reply
#5
Let me comment on the mechanics part: Right, there is no equivalent of TFM_AVA in BY_Trans file. I would recommend using NCAP_BND(r,p,FX,yr=0)=2 (via an INS table in BY_Trans) for the r-p you want to disallow. EDIT: Note that Antti has proposed a better solution below.

Generally, regions within one super-region are expected to have very similar RES structures. So, if you want to disable only a few r-p combinations, it is OK. But if there are too many cases, then you probably need to revisit the super-region definition.

Related information: Note that a big difference between BY templates and SubRES files is that the BY recognize regions while the SubRES file is completely unaware of regions. Regions are introduced into SubRES only via their respective Trans files.
Reply
#6
The method I would recommend for disabling processes would be to use PRC_NOFF (or NCAP_START=year, which is an alias for PRC_NOFF(BOH,year-1)), and to remove any existing capacities. I am using that method all the time in my scenarios. But as I understood it, Aymane did not seem to be happy with PRC_NOFF.

Using NCAP_BND(r,p,FX,yr=0)=2 has otherwise a similar effect, but it does not prevent the process parameters from being processed, and therefore generates some unnecessary "garbage" in the GAMS work memory.

Neither of these methods disables existing capacities (if any), and so to disable the processes completely, one should also make sure they are removed, if such exist.
Reply
#7
Dear Antti and Amit,

Thanks a lot for your help. I managed to disable the processes using PRC_NOFF. The differences were due to a non uniform definition of the timeslice levels of certain processes across regions. 

Best,
Aymane
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)