|
Antares Xpansion
Investment simulations for Antares studies
|
The MultipleProblemGeneration executable will run the problem generation step for every study in the tree.
The most simple usage is :
Note that this executable uses options that extend the simple ProblemGenerationExeOptions parser used in the individual problem generation executable. Thus, every other option of the individual problem generation can be given to the MultipleProblemGeneration executable and will be duplicated for each individual node's problem generation.
Note : the C++ executable has to be launched at the --dataDir. (The python driver takes care of this, this should not be a concern to most users).
The study paths file (or archive paths file when called with --archive <archive_paths_file>) tells the executable where to find the relevant input for each node :
--study)--archive)The reference file is a simple text file with two columns : The first column contains the node's name, and the second the path to the relevant input (in this case the path to the study). Note : the paths have to be given as relatives paths, relative to the --input-root
The full input files & folders layout for this example is diplayed in this section for more clarity.
The weights (resp. additional constraints) reference file gives the MultipleProblemGeneration executable information about the potential custom weights file (resp. additional constraints file) given by the user.
The reference file is a simple text file with two columns : The first column contains the node's name, and the second the path to the custom weights file (resp. additional constraints file). Note : the paths have to be given as relatives paths, relative to the --dataDir
If a node does not appear in this file, it will be assumed to have no custom weights file (resp. no additional constraints file).
In this example, for node 2030, the result will be the same as calling :
.
And for node 2040, the result will be the same as calling :
We give an example of a corresponding nodal_lp_info.json. This file is written to the path given in the --nodal-file argument when calling the executable.
This gives the following information :
lp_folder is the path to the folder containing the output of the problem generation.master_fileis the name of the file containing this node's master problem (no file extension, file extension is implicitely deduced from PROBLEMS_FORMAT input option later.)structure_file is the corresponding structure file (mostly useful if the name is not the default weights_file points to the weights file outputted by the problem generation if it exists (optional).