Launching the workflow : <tt>xpansion trajectory</tt>

Usage

Usage of the launcher is :

-xpansion-launcher --trajectory --step <step> [-i <path/to/input/root>] [--input-file <path/to/user/input/file>] ...

Mandatory arguments

• --trajectory : Tells the program to run in trajectory context.
• --step: Step of the workflow to execute, can be:
  • full : Executes the full workflow, step by step (in the order described below).
  • input_translation : Parses and translates the input-trajectory.yaml file, and writes the master_merger_info.json file. See : User input parsing and translation for more details.
  • problem_generation : Runs the problem generation step on the studies in the tree, and writes the nodal_lp_info.json file. See : Multiple problem generation for more details.
  • merge_master : Merges the generated master problem and writes the merged_master.<mps/svf> and merged_structure.txt files. See : Master problem merging for more details.
  • merge_weights : Merges the weights file generated during the problem generation into a single weights file that accurately weighs each subproblem depending on the node's study data and the node's information in the tree. See : Generating a merged weights file for more details.

Optional arguments

Input location

• -i, --dataDir : Path to the folder that contains each node's corresponding antaresXpansion study. If not specified, defaults to the current directory.
• --input-file : Path to the user input file. If not specified, the program will look for a file named input-trajectory.yaml in the --dataDir directory. If this file is not found, an error will be raised.
• --cache_problems : If specified, tells the program to cache the generated optimization problems on disk during the benders step, instead of keeping them in memory. Prevents out of memory issue for large studies.

Execution options

• --memory : Tells the program wether to run the problem generation in memory mode. Only useful when --step is either problem_generation or full.
  NOTE : As of now, multiple problem generation is only implemented for memory mode, and thus this flag is necessary when running a part of the workflow that includes the problem generation.
• --method : Similarly to the annual xpansion context, sets the method used for the resolution.
  NOTE : As of now, only benders and merge_mps resolutions are compatible with the trajectory context. The outer_loop (adequacy_criterion) method still requires adaptations and probably cannot be used as-is.
• --np : See the eponymous annual xpansion argument.
• --problems-format : See the eponymous annual xpansion argument.
• --solver : Tells the program which solver to use when manipulating optimization problems in steps

  _master

  and resolution

Note that this does not apply to the step problem_generation, where the solver and file format used in instead determined by the solver entry in <study>/user/expansion/settings.ini.

A note on intermediary files

As of now, most intermediary files will be written using a hardcoded default name in a folder name

_files

situated at : <INPUT_ROOT>/intermediary_files, and the drivers of the different steps will only look at those hardcoded locations when looking for those files.
TODO : implement arguments in xpansion-trajectory-launcher to allow custom names and locations for the intermediary files.

• After --step input_translation, master_merger_info.json will be found at

  <INPUT_ROOT>/intermediary_files/master_merger_info.json

  .
• The input files of the multiple_problem_generation executable generated by

  _problem_generation_driver

  will be written at :
  • <INPUT_ROOT>/intermediary_files/mpg_input.txt for the study paths file.
  • <INPUT_ROOT>/intermediary_files/mpg_weights.txt for the custom weights reference file, if at least one study has a custom weights file.
  • <INPUT_ROOT>/intermediary_files/mpg_constraints.txt for the additional constraints reference file, if at least one study has additional constraints.
• After --step problem_generation, nodal_lp_info.json will be found at

  <INPUT_ROOT>/intermediary_files/nodal_lp_info.json

• The options_merge_master.json file used by the merge_master_mps executable is generated by

  _master_driver

  and written at <INPUT_ROOT>/intermediary_file/options_merge_master.json
• After --step merge_master, the output files will be found at :
  • <INPUT_ROOT>/merged_master.<mps/svf> for the merged master problem. Note that it is written directly at the <INPUT_ROOT>.
  • <INPUT_ROOT>/merged_structure.txt for the merged structure file.
• After --step merge_weights, the output file will be found at <INPUT_ROOT>/merged_weights.txt.
• The options_benders.json file required by the different resolution executables (benders or merge_mps mostly) is written by the trajectory_resolution_driver at

  <INPUT_ROOT>/intermediary_file/options_benders.json

  .