Constrain solutions over time to be near each other
Created by: aroffringa
Introduce a new constraint that forces new solutions to be near the previous solution, using either some sensible heuristic for how large TEC jumps are allowed to be, or by manually providing the maximum TEC jump. Whenever the solutions converge to a value that is outside the allowed bounds, it is placed back onto the bound, and converges to its appropriate value later on.
The first solve is a bit special. It might be good to force the first solve to take a minimum nr. of iterations to make sure it converged.
This approach was just discussed with @revoltek and might help with the TEC jumps that we continue to see. Currently he uses a script afterwards that takes out any jumps -- but that defeats the idea a little bit of enforcing the constraint inside the calibration, as if one station has made an incorrect jump during the solve, that jump can be corrected afterwards for that station, but there will be some (small) effect on other stations as well caused by the incorrect jump of that station during the solve.