Optimization-based Methods

Introduction

In the context of forecast-based control (FBC), optimization-based methods determine the optimal operation of energy system components over multiple future timesteps by minimizing or maximizing an objective function while satisfying a set of constraints. These methods provide a structured and mathematically rigorous approach to decision-making in complex energy systems with forecasted inputs.

General Approach

The general approach to implementing optimization-based forecast-based controllers involves:

1. System modeling: Defining the mathematical representation of the energy system over multiple timesteps

2. Problem formulation: Specifying the objective function and constraints across the prediction horizon

3. Forecast integration: Incorporating forecasts of relevant parameters (e.g., weather, prices, loads)

4. Solver selection: Choosing an appropriate optimization solver for multi-period problems

5. Solution implementation: Applying the optimal control actions to the system (typically only the first timestep)

6. Receding horizon implementation: Re-solving the problem at each time step with updated forecasts and system states

Documentation Structure

This section is organized as follows:

• Mathematical Formulations
  • Introduction
  • Mathematical Formulation Structure
  • General Problem Structure
  • Key Concepts
• Implementations
  • Introduction
  • Available Implementations
• Building Your Own Optimization-Based Forecast-Based Controller
  • Introduction
  • Prerequisites
  • General Steps
  • Handling Forecasts
  • Implementing Receding Horizon Control
  • Extension Points
  • Best Practices

The Mathematical Formulation section focuses on the general understanding of the objective function and component models over a prediction horizon, independent of specific implementation details.

The Implementation section provides concrete implementations using different frameworks (Linopy and PyOptInterface) for forecast-based control.

The Build Your Own section provides instructions on how to extend or customize the optimization-based forecast-based controllers for specific needs.