TUTORIALS

Prediction : An End-to-End Tutorial

Tutorial organizers

Lotfi Snoussi, Tunisia
Olfa Fakhfakh, Tunisia

Subject Area :

Artificial Intelligence for Green Energy
Machine Learning and Ensemble Methods for Energy Systems
Explainable AI for Sustainable Power Systems

Main Objectives :

The digital transformation of the energy sector is accelerating the adoption of Artificial Intelligence to enhance efficiency, reliability, and sustainability. Accurate energy output prediction has become a critical component of modern power systems, enabling better operational planning, performance optimization, and smarter integration of renewable resources.

This tutorial offers a hands-on, end-to-end introduction to Machine Learning for energy output prediction using a real-world dataset. Participants will build a complete predictive pipeline—from data exploration and preprocessing to advanced model development, evaluation, and comparison.

The session combines regression techniques, powerful ensemble learning methods, and Artificial Neural Networks to model complex nonlinear relationships in energy systems. Emphasis is placed on robust validation strategies, hyperparameter optimization, and generalization analysis to ensure reliable and scalable solutions.

In addition, the tutorial introduces practical Explainable AI (XAI) tools to interpret model behavior and understand the influence of key variables on energy output. This transparency is essential for deploying trustworthy AI systems in real-world energy environments.

Through guided demonstrations and reproducible Python implementations, participants will gain practical expertise in designing data-driven models tailored to green energy challenges. The tutorial bridges theoretical AI concepts with operational energy applications, equipping researchers and engineers with actionable skills for building intelligent and sustainable energy systems.

Brief Description of the tutorial proposal

This tutorial provides a comprehensive and practical introduction to applying Machine Learning techniques for energy output prediction using a real-world dataset. It covers the complete ML workflow while integrating regression models, ensemble learning methods, neural networks, and Explainable AI (XAI).

The main objectives are:

To understand the complete Machine Learning pipeline for energy prediction.
To implement and compare regression and ensemble models.
To apply Artificial Neural Networks for nonlinear energy modeling.
To optimize and validate models using cross-validation and hyperparameter tuning.
To interpret model behavior using Explainable AI techniques.
To design reliable and trustworthy AI-driven energy forecasting systems.

All regular contributions to IEEE ICAIGE26 & S4IoT26 must be submitted online, in electronic format to :

https://confcomm.ieee-ies.org/app/general/conferences/ICAIGE-S4IoT26/initial-submission

Tutorial Proposal

Title : Explainable AI for Sustainable Energy Output Prediction: An End-to-End Tutorial

Download Now

Intelligent Power Routing in Packetized Energy Networks

Tutorial organizers

Hadi Y. Kanaan, Lebanon
Amani Fawaz, Lebanon

Subject Area :

Artificial Intelligence for Smart Grids
Energy Internet
Packetized Power Networks
Intelligent Energy Management
Reinforcement Learning
Multi-Agent Systems
Smart Energy Routing
Green Energy Systems

Main Objectives :

his tutorial aims to introduce the emerging concept of the Energy Internet and its transformation of conventional power systems into intelligent, decentralized, and packetized energy networks. The tutorial focuses on Energy Internet, Energy Routers, packetized power transmission, and intelligent power routing protocols.

The tutorial will explain how Artificial Intelligence (AI), Reinforcement Learning (RL), Multi-Agent Systems (MAS), and Metaheuristic Optimization techniques can be integrated into power routing protocols to optimize energy distribution, reduce transmission losses, manage congestion, improve scalability, and support self-healing operation in decentralized energy networks.

Practical implementation methodologies based on MATLAB simulations and distributed Q-learning routing mechanisms will also be presented.

Brief Description of the tutorial proposal

The Energy Internet (EI) represents a major evolution of conventional smart grids toward intelligent, decentralized, and highly interconnected energy systems. Inspired by the operational principles of the Internet, the EI enables energy exchange in the form of packetized power transmission between distributed producers and consumers through Energy Routers (ERs). As renewable energy integration, peer-to-peer energy trading, and decentralized microgrids continue to expand, efficient and adaptive power routing protocols become essential to ensure reliable, scalable, and energy-efficient operation.
This tutorial presents recent advances in intelligent power routing methodologies for packetized energy networks, with a strong emphasis on Artificial Intelligence (AI)-based routing approaches. The tutorial introduces the architecture of Energy Internet systems, Energy Routers, packetized power concepts, and the formulation of the power routing problem under real-world constraints such as congestion, transmission losses, limited capacities, and dynamic topology changes.
Several routing methodologies will be discussed, including graph theory approaches, metaheuristic optimization techniques, Multi-Agent Systems, and Reinforcement Learning methods. Particular focus will be given to distributed Q-learning-based routing protocols and self-healing routing mechanisms for decentralized energy networks.
The tutorial also explores future research directions including Graph Neural Networks (GNNs), intelligent adaptive routing, cloud energy storage, and AI-native energy infrastructures for future green energy systems.

All regular contributions to IEEE ICAIGE26 & S4IoT26 must be submitted online, in electronic format to :

https://confcomm.ieee-ies.org/app/general/conferences/ICAIGE-S4IoT26/initial-submission

Tutorial Proposal

Title : Explainable AI for Sustainable Energy Output Prediction: An End-to-End Tutorial

Download Now

confcomm.ieee-ies.org