Re: Redefine parameter in Monte Carlo analysis

[Thiago Neves]

Dear Guidati,

Complementing Michael's answer, instead of creating a single DAT file containing all the parameters, you can run GLPK with 2 (or more) dat files, where in one of then you put the fixed parameters and in the other one you can put the variable (monte carlo sampled) parameters, like below:

FIXED_PARAMS.dat:
{all the fixed parameters you already use in the model}

MONTE_CARLO_PARAMETERS.dat:
data;
 param Xnormal := {...} ;
 param Xadjust := {...} ;
end;

and them run:
glpsol -m [mod file] -d FIXED_PARAMS.dat -d MONTE_CARLO_PARAMETERS.dat

The advantage is that you can create 1000 samples, put them in 1000 different "MONTE_CARLO_PARAMETERS.dat" files, and them run 1000 glpsol processes in parallel.

Att,

Thiago H. Neves

(31) 98608-0666

Em ter., 12 de nov. de 2019 às 14:53, Michael Hennebry <address@hidden> escreveu:

On Tue, 12 Nov 2019, Guidati  Gianfranco wrote:

> I am using GLPK to optimize an energy system model for Switzerland. Lots of parameters such as costs or efficiencies are defined in a dat file. Now I want to systematically vary some of these parameters using a Monte Carlo approach. The easiest way to do this would be to just create a small addendum to the dat file that redefines / overrides the parameter values that were previously defined in the dat file, and to paste these two together. Unfortunately, that doesn't work, the solver just complains that a parameter was already defined.

Include parameters Xnormal and Xadjust along with parameter X.
Read in Xnormal and Xadjust instead of X.
Have X computed from Xnormal and Xadjust.

--
Michael   address@hidden
"Sorry but your password must contain an uppercase letter, a number,
a haiku, a gang sign, a heiroglyph, and the blood of a virgin."
                                                              --  someeecards