Hello everyone,
I'm working on an interesting problem and I'm curious if I can use GLPK
to solve it.
Suppose I am trying to solve a linear program, but for some reason I
have some prior knowledge.. i.e. a good guess as to the optimal values
of my variable. In fact let's say it's a very good guess.. I know that
the optimal values can only be each off my some small error epsilon.
Based on my (admittedly rough) understanding of the LP algorithms, it
seems like this knowledge could only help. Please correct me if I'm wrong.
Is there any way I could inform simplex or an interior point of this
prior knowledge, e.g. tell it to start looking at a particular part of
the feasible region with the values I have?
(For some context: I have a problem where I need to incrementally solve
a sequence of LPs, where each LP in the sequence has the same form as
the one before it and only a small change a smaller number of the
constraints' coefficients. I'd like to solve the first LP once from
scratch and then hopefully iteratively re-use the values to the next ones. )
Thanks,
Marc
_______________________________________________
Help-glpk mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/help-glpk
