Startup & partial: startup variable and timestep length

[ojdo]

Just to not forget looking into it: constraint def_startup_capacity_rule states that the startup variable startup_pro must be greater or equal to the difference of online capacities of the current and previous time steps: cap_online[tm, ...] - cap_online[tm-1, ...]. Without at least one extensive whiteboarding session, I'm unsure whether this value should also depend on the time step length m.dt. Intuitive reasoning: an increase of cap_online by a certain value should be more expensive if it happens over a shorter time, basically increasing the gradient the process experiences.

Open question: should the new expression simply be (cap_online[tm, ...] - cap_online[tm-1, ..]) / m.dt?

[ojdo]

Conversely, I'm more an more convinced that m.dt has to be removed from the startup cost term m.costs['Startup']. Also here, some more conceptual thought has to occur.

[lodersky]

Partial load was redone with commit e5146ca03dcccc80cddbb591143000920d36a51d.

General mixed integer formulations (including startup) will be implemented in Q3 or Q4.