Hi Carlos, MOST does not have a mechanism for adding variables or constraints without modifying existing code (like you can with MATPOWER’s OPF), but it should be straightforward to add variables and constraints by modifying most.m (or, even better, a copy of it). Simply add your own om.add_var(…) and om.add_lin_constraint(…) calls at the appropriate places.
If you are including integer variables for non-unit-commitment problems, you will also have to modify the logic used to select the mixed integer (miqps_matpower) versus continuous (qps_matpower) solver, since it currently only calls the mixed integer solver when doing unit commitment. Ray > On Oct 16, 2019, at 9:04 AM, Carlos Ferrandon Cervantes > <[email protected]> wrote: > > Hello Doctor: > > I am aware that unit commitment uses MILP techniques to solve the > optimisation problem. For frequency stability reasons in the power system, I > am in the process of adding new constraints to the problem, but some of them > are nonlinear. My approach is using separable programming to linearise them. > Basically the solver will have to deal with new constraints (let's call them > λ) that meet SOS2 conditions, i.e., at most two of the variables can be non > zero, mapping these r-variables in a sum from the real numbers 0 to 1. Also, > another set of binary r-constraints (let's call them y) must be met such as > only one of them can be nonzero. (I am adding the picture of the conditions > in the email) > <image.png> > > My question more specifically is: Does adding these new constraints to MOST > is possible? These constraints will be linked to reserves and the power > output of the generators. And they will be different to the already existing > binary operators u,v,w. > > As always, thank you very much in advance, > > Carlos Ferrandon > -- > Carlos Ferrandon
