You might try using a small overlap in the intervals covered by subsequent
runs. E.g Use a 27-hour planning horizon run every 24 hours.
Ray
> On Oct 26, 2018, at 5:47 PM, Kaspar Mueller <[email protected]> wrote:
>
> Hi Matpower community,
>
> We are solving long time(year) large system(10k nodes) multiperiod dcopf
> problems. In order to do so, we split the whole time interval into smaller
> mp-dcopf runs. This because we run into memory issues caused by the size of
> the system.
> As discussed here:
> https://www.mail-archive.com/[email protected]/msg05683.html
> <https://www.mail-archive.com/[email protected]/msg05683.html>
> we enforce the ramp constrain by adding a profile that fixes the PMIN and
> PMAX for the first hour.
>
> I was wondering if there is any other way to solve large systems or is there
> another way to connect intervals of multi-period runs to enforce the ramp
> constrain.
>
> Thanks for your help.
>
> Best,
> Kaspar
