Help getting plot_feasibilities to run with fixed/performance variables

[1ozturkbe]

@alisaab I want to generate the flight envelope of an aircraft (eg. range vs. payload). But I want the performance variables of a model not to be fixed (i.e. aircraft can fly at different CL etc.), and I want all of the fixed variables to have the same values (all variables with a given fix = True attribute). The current code keeps everything fixed except for the two variables of interest. (Right?) Perhaps we can meet at some point to discuss, and improve plot_feasibilities to be able to know which variables should be fixed and varied.

[alisaab]

@1ozturkbe actually both scenarios can be done. setting design_feasibility to True would fix the variables with attribute fixed only, while setting it to false would fix everything. We should meet and talk more about that!

[1ozturkbe]

Whaaat, that is pretty non-intuitive. We should definitely meet in short order to talk about this stuff. We have been floating the idea of having a hackathon for gpkit-related stuff, so you and I can perhaps have a robust hackathon.

[1ozturkbe]

@alisaab I wanted to ask for a clarification. If you do designed_model.localsolve() that solves the robust solution in the nominal case, right?

[alisaab]

@1ozturkbe the designed_model takes a solution (possibly the robust solution), and uncertain parameters substitutions (possibly a random realization). The free variables marked as fixed=True will be substituted using the given solution, and the uncertain parameters will be substituted using the substitutions given.