Antares-Xpansion simulation procedure

An Antares-Xpansion simulation is currently run in four separate steps, managed by a Python orchestrator.

The basic idea behind an Antares-Xpansion simulation is to let Antares-Simulator read the Antares study and build the initial supply-demand optimisation problem without knowledge of the investment candidates. Then the investment optimization problem is built and solved within Antares-Xpansion.

There are four steps in a full Antares-Xpansion simulation. Each section title is followed by the corresponding value of the variable --step when launching antares-xpansion-launcher from the command line.

1- Antares-Simulator optimization problems retrieval: antares

The solver of Antares-Simulator, antares-solver, runs the Antares study with a specific configuration set by Antares-Xpansion, see Antares optimization problem retrieval for more details.

The weekly supply-demand optimization problems are written as .mps files. In this mode Antares-Simulator writes some additional files that add a business context of the linear problems (variable.txt, area.txt and interco.txt). These files allow to map the variable numbers in the .mps files to their physical meaning. For more details, see Antares optimization problem retrieval.

2- Investment problem generation: problem_generation

This stage is responsible for building the investment optimization problem, using the Benders reformulation, which results in the creation of a master problem and of satellite problems. The master problem takes investment decisions whereas satellite problems solve the Antares supply-demand problems once the investment decisions are taken.

Using the data on investment candidates specified in the user/expansion/candidates.ini file, Antares-Xpansion modifies the .mps problems provided by Antares-Simulator in the previous step to add information on the candidates. These modified problems, still written as .mps files, are the so-called satellite problems. They correspond to the supply-demand problem to be solved once the investment has been fixed to a given value (that changes during the Antares-Xpansion optimization).

This stage is also responsible for building the initial master problem used for Benders decomposition. If there are additional constraints linking the investment candidates (defined in the additional-constraints field of the user/expansion/settings.ini file), then they are added to the master problem.

For more details on how the master and the satellite problems are generated, see Investment problem generation.

3- Resolution stage with Benders decomposition: benders

The core step of Antares-Xpansion problem consists in solving the investment problem generated in the previous step. Antares-Xpansion uses a Benders decomposition algorithm as it is well-suited to the structure of the investment problem: the satellite problems are independent, i.e. the weekly problems in an Antares study are independent (using water values is not allowed in Antares-Xpansion). Details of the methodology are given in the optimization principles page.

4- Update of antares study: study_update

The resolution stage provides, for each candidate, the investment that minimizes the overall cost of the investment problem. However, it does not allow access the details provided by an Antares-Simulator simulation. In order to access these details, a "standard" Antares-Simulator simulation is prepared by updating the original Antares study with the values obtained from the resolution stage. Each investment is associated to a link and its investment value is added to the link direct and indirect transfer capacity.

Antares-Xpansion package executables

Antares-Xpansion consists in four executables and a Python orchestrator that is responsible for calling these executables with the correct options for each step.

Executable	Simulation step
antares-x.x-solver	Antares-Simulator optimization problems retrieval, where x.x stands for the version number of Antares-Simulator, for example antares-8.1-solver.
lp_namer	Investment problem generation.
benders	Benders decomposition.
xpansion-study-updater	Update of antares study.