Skip to content

File Overview : Model, System, Optimization, Business

The GEMS architecture enforces a structured approach, separating modelling logic, system configuration, optimization workflow, and business intelligence into four distinct "bounded domains" (see the following definition diagram):

File structure


Types of files and roles

The different types of files describing a GEMS study case are:

Type of File Domain File Description & Role
Model Libraries Abstract modelling YAML (e.g., basic-models-library.yml, antares-models-library.yml) Defines Models: Abstract representations of system components to be simulated. Models are defined in a library file and specifies its ports, parameters, and internal behavior.These definitions can also include optional constraint and objective contributions used in simulation.
Taxonomy Abstract modelling YAML (e.g., taxonomy.yml) Model Structure & Categories: Specifies mandatory parameters, variables, ports, or extra outputs per category. Useful for structuring the UI (user interface) and simulation outputs.
System System YAML (system.yml) Defines Components: Numerical instantiation of models, linking to model IDs (e.g., example_library_id.example_model_id). Specifies parameter values and connections between components via ports, forming the system graph.
Timeseries System Dataseries (e.g., wind_generation.csv, solar_generation.csv) Time-dependent Data: Numerical data for parameters varying by time and scenario. Stored as .csv or .tsv files, typically in a data-series folder.
Solution Workflow Solution Workflow YAML (optim-config.yml) Workflow Definition: Describes calculation block processing (sequential, parallel, Xpansion frontale, Benders decomposition) and master problem constraints, especially for investment variables.
Optimization Parameters Solution Workflow YAML (parameters.yml) Solver & Configuration Settings: Contains solver parameters and configuration required for running Modeler.
Business Views Configurations Business Intelligence YAML (e.g., business-view-def.yml, business-metric.yml) Business Metrics Logic: Calculates business metrics from simulation results in two phases: Step 1 (component scope, complex arithmetic), Step 2 (global scope, aggregation/filtering).

Files Interaction

The following scheme shows the interaction of the different core concepts presented previously. It is based on the basic-model-library included in this documentation.

GEMS Architecture Diagram



Illustration with an example

To get familiar with these concepts, see the table below for a correspondence between theoretical concepts and examples from a thermal plant use case from the basic_models_library:

Concept Example
Library Library (basic_models_library) where generator model is defined
Model The generator model
Component A specific thermal plant, like a 300MW CCGT Thermal plant from generator model
Variable Actual dispatched power ; generation from generator model
Parameter The generation_cost (specific for each thermal plant, to be entered by the users)
Field The field flow is exchanged through balance_port
Port The balance_port let the power flow be transferred
Connection A link between a generation unit and a node representing the injection from the plant to the energy system's node
Constraint Value interval accepted for power generation
Binding Constraints Energy balance inside an area
Scenario dependency The maximum power output p_max can depend on the scenario chosen by the users (and also depends on the time)
Time dependency p_max is a time dependent parameter (and also depends on the scenario)

Outputs File (generated by a Gems interpreter)

The outputs of GEMS consist of three main categories of objects: Optimization Problem, Simulation Table and Business Views. Their structure is detailed in the User Guide section.

  • Optimization Problem represents the global mathematical formulation of the energy system simulation/optimization.
  • Simulation Table contains the raw results of the simulation or optimization, including the optimal values of decision variables and the values of expressions computed from them.
  • Business Views provide curated representations of the simulation or optimization results from a business-intelligence perspective, tailored to users specific needs.