eta-ctrl 0.3.0

A modular framework for rolling-horizon operational control using a Gym-compatible agent–environment abstraction. The framework enables integration of optimization-based agents (Pyomo), reinforcement learning agents (Stable-Baselines3), heuristic search methods, and rule-based controllers. Environments can be instantiated from FMI-compliant simulations, mathematical models, or real-world system interfaces.

The ETA Ctrl framework provides a standardized interface for developing digital twins of factories or machines in a factory. It is designed to facilitate rolling horizon optimization, simulation, and interaction with factory systems. The framework is based on the Gymnasium environment and integrates seamlessly with tools like FMUs, Pyomo models, and live connections to real-world assets.

Documentation

Full Documentation can be found on the Documentation Page.

Warning

This is beta software. APIs and functionality might change without prior notice. Please fix the version you are using in your requirements to ensure your software will not be broken by changes in ETA Ctrl.

Overview

Core

• `EtaCtrl`: Central controller for managing optimization workflows, including learning and execution processes.

Configuration

• `Config`: Represents the configuration for an optimization run.

• `RunInfo`: Handles paths and metadata for optimization runs.

Environment

• Base Classes:

  • `BaseEnv`: Abstract base class for creating custom environments.

  • `LiveEnv`: Extends BaseEnv for live environments interacting with real-world systems.

  • `PyomoSimEnv`: Extends BaseEnv for environments using a pyomo model for simulation.

  • `SimEnv`: Extends BaseEnv for environments using FMU-based simulations.

• Vectorization:

  • `NoVecEnv`: Custom vectorizer for environments that handle multithreading internally.

Simulation

• `FMUSimulator`: Provides functionality for simulating FMUs (Functional Mock-up Units).

Time Series

• `scenario_from_csv`: Imports and processes scenario data from CSV files.

• `df_from_csv`: Reads time series data from a CSV file and returns it as a pandas DataFrame.

• `df_resample`: Resamples the time index of a DataFrame to a specified frequency.

• `df_interpolate`: Interpolates missing values in a DataFrame with a specified frequency.

State Management

• `StateVar`: Represents a single variable in the state of an environment.

• `StateConfig`: Configures the action and observation spaces based on StateVar instances.

Contributing

Please read the development guide before starting development on ETA Ctrl

Citing this Project

For referencing this package in academic work, please refer to CITATION.cff.

See AUTHORS.rst for a list of further contributors.