Summer School 2026

29 June - 2 July 2026

Artificial Intelligence and Optimization:
From Mathematical Modelling to Hybrid and Learning-Based Approaches

Batiment K, Université de Haute-Alsace
18 rue des Frères Lumière
68093 MULHOUSE Cedex

All registrations are subject to prior approval by the organizers. A confirmation will be sent by email.
All registrations are final, binding, and non-refundable. By completing the registration, the participant expressly acknowledges and agrees that no cancellations, transfers, or refunds will be permitted.
Deadline for registration: 15 May 2026

Why Optimization Matters

Optimization is a fundamental pillar of numerous scientific and industrial applications, including logistics, planning, energy systems, transportation, and healthcare. However, real-world optimization problems are often large-scale, combinatorial, nonlinear, and highly constrained, which significantly limits the effectiveness of classical optimization techniques when applied in isolation.

In response to these challenges, hybrid approaches combining mathematical optimization, metaheuristics, and machine learning have emerged as powerful paradigms capable of leveraging both the theoretical guarantees of exact methods and the adaptability of learning-based techniques.

This summer school aims to provide a comprehensive and modern perspective on optimization, integrating the following methodological components:

mathematical modelling and exact optimization methods,
metaheuristics and approximate algorithms,
machine learning and data-driven optimization,
surrogate (approximation) models for computationally expensive functions,
selection and generation hyper-heuristics

The objective is to equip participants with the theoretical foundations and practical skills required to design robust, efficient, and scalable optimization strategies for complex real-world problems.

Overview

Understanding Modern Optimization Approaches

Optimization Problems in Real-World Practice

Optimization problems arise in many domains such as logistics, energy, transportation, and healthcare, where decisions must be made within large-scale, combinatorial, and often nonlinear search spaces under complex constraints. Mathematical modelling provides the framework to represent these systems and define objective-driven decision processes.

As models become more realistic and problem sizes increase, solving them becomes computationally demanding. This calls for solution methods that can effectively explore complex search spaces while maintaining the structure of the underlying model.

Exact, Hybrid, and Learning-Based Solution Methods

A range of solution approaches can address such problems. Exact methods provide rigorous formulations and solution guarantees, and are often the natural starting point, while metaheuristics offer practical ways to explore large, complex search spaces.

These approaches can be combined in different ways, including integrating exact methods with heuristics, combining several heuristic strategies, or designing solution processes that exploit their respective strengths. More recently, learning-based techniques have been introduced to guide or adapt the search using data, helping improve efficiency in certain settings. Together, these approaches form a coherent set of tools for tackling complex optimization problems.

What You'll Learn

Methodological Components:

✓ Mathematical modelling and exact optimization methods
✓ Metaheuristics and approximate algorithms
✓ Machine learning and data-driven optimization
✓ Surrogate (approximation) models for computationally expensive functions
✓ Selection and generation hyper-heuristics

Learning Objectives

By the end of this summer school, you will be able to:

Formulate Real-World Problems as mathematical optimization models with real case study: Electric Vehicle Charging Scheduling (EVCS)

Solve EVCS Using Exact Methods including MILP solvers such as Gurobi

Design Efficient Metaheuristics and understand algorithmic implementation

Develop Hybrid Strategies combining metaheuristics and machine learning

Integrate Surrogate Models to reduce computational costs

Conduct Experimental Comparisons between exact, heuristic, and hybrid approaches

Tentative Program

4 Days of Intensive Learning and Hands-On Practice

Day 1

From Modelling to Exact Methods

Pr. Ammar Oulamara and Dr. Abdennour Azerine

Introduction to combinatorial optimization and operations research
Mathematical modelling of the EVCS problem using MILP formulations
Problem solving with exact solvers (e.g., Gurobi)
Performance analysis, scalability issues, and methodological limitations

Day 2

Metaheuristics and Hybridization

Pr. Diego Oliva, Dr. Mahmoud Golabi and Pr. Lhassane Idoumghar

Overview of state-of-the-art metaheuristics
Metaheuristic design for the EVCS problem
Hybrid metaheuristic–machine learning approaches
Introduction to surrogate modelling in optimization

Day 3

Learning-Based and Automated Optimization

Pr. Ed. Keedwell

Selection and Generation Hyper-Heuristics for Optimisation
Sequence-based Selection Hyper-Heuristics
Sequence-based Selection Hyper-Heuristics applied to EVCS problem
Practical applications and real-world case studies

Day 4

Challenge / Competition

All Instructors

Apply all learned techniques to a new optimization problem
Work in teams or individually
Present your solution and methodology
Receive feedback from experts in the field

Why Attend?

Gain Knowledge and Network with Leading Experts

Coherent Framework

A comprehensive methodological framework integrating all modern optimization approaches

Hands-On Development

Strong emphasis on practical algorithmic development with real-world applications

State-of-the-Art Techniques

Learn cutting-edge optimization and AI techniques from industry leaders

Real Case Study

Realistic and industrially relevant Electric Vehicle Charging Scheduling problem

Expert Interaction

Direct interaction with experts and leaders in optimization and AI research

Professional Networking

Connect with PhD students, researchers, and professionals from around the world

Scientific Committee

Meet the Experts Leading This Summer School

Lhassane Idoumghar

University of Haute Alsace, France

Edward Keedwell

University of Exeter, UK

Diego Oliva

University of Guadalajara, Mexico

Ammar Oulamara

University of Lorraine, France

Pierrick Legrand

University of Bordeaux, France

Evelyne Lutton

INRAE, France

Nicolas Monmarché

University François Rabelais of Tours, France

Denis Pallez

Université Côte d'Azur, France

Abdennour Azerine

University of Haute Alsace, France

Mahmoud Golabi

University of Haute Alsace, France