------------------------------------------------------------ To: Philipp Bachmann <[email protected]> Subject: Re: [Help-glpk] Can I combine "glp_mpl_read_model(, , 1)" Message-ID: <1328616049.3377.19.camel@corvax> From: Andrew Makhorin <[email protected]> Date: Tue, 07 Feb 2012 15:00:49 +0300 ------------------------------------------------------------
[snip - earlier part of post] > If your model has a simple structure, I'd suggest > build it with glpk api routines, i.e. not to use > MathProg description at all. This is quite easy and > would resolve most interface problems. Hello Philipp You can also build quite complex models programmatically. As an example from my current project, take the following high-voltage transmission line data (connectivity information omitted): capacity [W] f > 400e+06 resistance-per-metre [ohm/m] f > 67e-06 reactance-per-metre [ohm/m] f > 364e-06 length [m] f > 200e+03 voltage-angle-delta-upper [degrees] f > 30.0 discretization-steps [-] i > 10 nameplate-capacity [W] f > 400.0e+06 duty-specific-cost-financial [$/J] f > 0.0 size-specific-cost-financial [$/W/s] f > 8.0e-12 standing-cost-financial [$/s] f > 0.0 annual-discount-rate-decimal [-] f > 0.08 economic-life [y] i > 30 capex-initial [$] f > 1.6e+09 capex-terminal [$] f > 0.4e+09 current-age [y] i > 10 I have a couple of layers of abstraction (C++ class hierarchies) between this and the GLPK solver, but most of the above data does eventually end up in an MIP structural matrix somewhere deep within a GLPK problem object. best wishes --- Robbie Morrison PhD student -- policy-oriented energy system simulation Technical University of Berlin (TU-Berlin), Germany University email (redirected) : [email protected] Webmail (preferred) : [email protected] [from Webmail client] _______________________________________________ Help-glpk mailing list [email protected] https://lists.gnu.org/mailman/listinfo/help-glpk
