Talk:Parametric programming

So what is big Theta?
Is the $$\subset$$ in $$\theta \in \Theta \subset R^m$$ strict? I.e., is the parameter space always (or usually) constrained in some way apart from the constraints $$g$$? Or can I safely change that to a $$\subseteq$$ for clarity? Q VVERTYVS (hm?) 21:22, 4 February 2015 (UTC)

Yes indeed, the $$\subset$$ is strict. The reasoning behind it is the boundedness of the parameter space. If you have an unbounded space, then you would need to prove that nowhere in this unbounded space there is at least one point where another set of constraints is active (which denotes a different solution). While you might be able to construct a very simple example (say 1 parameter, 1 constraint), where you can exhaustively enumerate all the options and make this statement, in general $$\subset$$ has to hold. In particular when you look at nonlinear systems (see the citation to Fiacco's sensitivity theorem for that), then you see that the results only hold in a neighborhood of the initial solution point. So, to summarize: you might find some small examples where you could say $$\subseteq$$, but to include this in the general problem formulation is in my opinion misleading. What do you think?