Flavien Lucas
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Flavien Lucas has a total of 16 co-authors. Here are the top co-authors by number of shared publications:
- Bachtiar Herdianto
IMT Atlantique Lab-STICC (UMR 6285 CNRS) Brest France
4 shared publications - Marc Sevaux
Université Bretagne Sud, Lab-STICC, UMR 6285, CNRS, Lorient, France
4 shared publications - Romain Billot
IMT Atlantique Lab-STICC (UMR 6285 CNRS) Brest France
4 shared publications - Kenneth Sörensen
University of Antwerp
2 shared publications - Ying Lian
ANT/OR - Operations Research Group, Department of Engineering Management, University of Antwerp, Belgium
2 shared publications - Alexis Lebis
Univ. Lille, Centre for Digital Systems
1 shared publication
Latest Publications
- 14 October 2026
- Published by Springer Nature in Journal of Heuristics
- Vol. 32 (4) , 33
Study of robust features in formulating guidance for heuristic algorithms for solving the vehicle routing problem
- 14 October 2026
- Published by Springer Nature in Journal of Heuristics
- Vol. 32 (4) , 33
Study of robust features in formulating guidance for heuristic algorithms for solving the vehicle routing problem
- 4 April 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
Résolution de problèmes réalistes de tournées à flotte hétérogène en milieu urbain : vers un solveur adaptatif mêlant recherche opérationnelle et apprentissage automatique
Dans le domaine de la recherche opérationnelle, la résolution des problèmes de tournées constitue un défi emblématique depuis les années 1960. Initialement développées sur des problèmes simples, de nombreuses variantes ont permis d'améliorer le réalisme des instances étudiées, en particulier en milieu urbain où la congestion dépend du véhicule utilisé. Cependant, encore peu de méthodes sont compatibles avec l'ensemble des variantes existantes, celles-ci perdant de l'efficacité avec l'augmentation du nombre de clients. C’est dans ce contexte que nous avons développé un solveur robuste et adaptatif, obtenant de bonnes solutions dans un très bref délai sur un nombre élevé d’instances aussi difficiles que variées (plusieurs milliers de clients, véhicules hétérogènes, milieu urbain, etc.). Ce nouveau solveur permet de générer des solutions issues de multiples instances, caractérisées par un ensemble de variables, dont l'étude permet de mieux comprendre ce que constitue une bonne solution. Dans un contexte où les travaux interdisciplinaires croisant recherche opérationnelle et apprentissage automatique sont de plus en plus fréquents, nous nous sommes intéressés à l’utilisation d’arbres de décision pour cartographier les aires contenant a priori l’ensemble des solutions de bonne qualité. La combinaison entre les arbres de décision et ce nouveau solveur restreint la recherche de bonnes solutions aux zones prometteuses, afin d’augmenter les chances d’obtenir de bons résultats et de réduire la durée de cette exploration. Les résultats expérimentaux indiquent que ces travaux interdisciplinaires sont prometteurs pour générer un ensemble de méthodes hybrides performantes.
- 4 April 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
Résolution de problèmes réalistes de tournées à flotte hétérogène en milieu urbain : vers un solveur adaptatif mêlant recherche opérationnelle et apprentissage automatique
Dans le domaine de la recherche opérationnelle, la résolution des problèmes de tournées constitue un défi emblématique depuis les années 1960. Initialement développées sur des problèmes simples, de nombreuses variantes ont permis d'améliorer le réalisme des instances étudiées, en particulier en milieu urbain où la congestion dépend du véhicule utilisé. Cependant, encore peu de méthodes sont compatibles avec l'ensemble des variantes existantes, celles-ci perdant de l'efficacité avec l'augmentation du nombre de clients. C’est dans ce contexte que nous avons développé un solveur robuste et adaptatif, obtenant de bonnes solutions dans un très bref délai sur un nombre élevé d’instances aussi difficiles que variées (plusieurs milliers de clients, véhicules hétérogènes, milieu urbain, etc.). Ce nouveau solveur permet de générer des solutions issues de multiples instances, caractérisées par un ensemble de variables, dont l'étude permet de mieux comprendre ce que constitue une bonne solution. Dans un contexte où les travaux interdisciplinaires croisant recherche opérationnelle et apprentissage automatique sont de plus en plus fréquents, nous nous sommes intéressés à l’utilisation d’arbres de décision pour cartographier les aires contenant a priori l’ensemble des solutions de bonne qualité. La combinaison entre les arbres de décision et ce nouveau solveur restreint la recherche de bonnes solutions aux zones prometteuses, afin d’augmenter les chances d’obtenir de bons résultats et de réduire la durée de cette exploration. Les résultats expérimentaux indiquent que ces travaux interdisciplinaires sont prometteurs pour générer un ensemble de méthodes hybrides performantes.
- Open Access
Optimal target search: modeling complex evasive behavior with automata
In this study, we address the problem of target search, specifically focusing on the search allocation problem which involves optimizing the distribution of a search budget to detect a target. Previous studies in the scientific literature have made oversimplifications about target behavior, typically assuming either that the target has perfect knowledge of the searchers and makes optimal decisions, or that the target's movements are entirely predictable and will not change in response to the search efforts. To overcome these limitations, we propose a new mathematical model that represents the target as a deterministic automaton. This model allows for the representation of complex target behavior, closer from a real human behavior, where the target's strategy can evolve with partial information gained about the searchers and where suboptimal decisions are modeled. The mathematical formulation is a mixed-integer non-linear program, and we propose an adaptation of a Forward-And-Backward heuristic algorithm to provide solutions to this problem.Experimental studies evaluate the feasibility of this approach in terms of computation time and quality of the solutions and demonstrate the potentials and limits of such an approach.
Open AccessOptimal target search: modeling complex evasive behavior with automata
In this study, we address the problem of target search, specifically focusing on the search allocation problem which involves optimizing the distribution of a search budget to detect a target. Previous studies in the scientific literature have made oversimplifications about target behavior, typically assuming either that the target has perfect knowledge of the searchers and makes optimal decisions, or that the target's movements are entirely predictable and will not change in response to the search efforts. To overcome these limitations, we propose a new mathematical model that represents the target as a deterministic automaton. This model allows for the representation of complex target behavior, closer from a real human behavior, where the target's strategy can evolve with partial information gained about the searchers and where suboptimal decisions are modeled. The mathematical formulation is a mixed-integer non-linear program, and we propose an adaptation of a Forward-And-Backward heuristic algorithm to provide solutions to this problem.Experimental studies evaluate the feasibility of this approach in terms of computation time and quality of the solutions and demonstrate the potentials and limits of such an approach.
- 16 December 2025
- Published by Elsevier in European Journal of Operational Research
- Vol. 333 (2) , 414-428
Feature-guided metaheuristic with diversity management for solving the capacitated vehicle routing problem
- 16 December 2025
- Published by Elsevier in European Journal of Operational Research
- Vol. 333 (2) , 414-428
Feature-guided metaheuristic with diversity management for solving the capacitated vehicle routing problem
- Open Access
Hybrid Node-Destroyer Model with Large Neighborhood Search for Solving the Capacitated Vehicle Routing Problem
In this research, we propose an iterative learning hybrid optimization solver developed to strengthen the performance of metaheuristic algorithms in solving the Capacitated Vehicle Routing Problem (CVRP). The iterative hybrid mechanism integrates the proposed Node-Destroyer Model, a machine learning hybrid model that utilized Graph Neural Networks (GNNs) such identifies and selects customer nodes to guide the Large Neighborhood Search (LNS) operator within the metaheuristic optimization frameworks. This model leverages the structural properties of the problem and solution that can be represented as a graph, to guide strategic selections concerning node removal. The proposed approach reduces operational complexity and scales down the search space involved in the optimization process. The hybrid approach is applied specifically to the CVRP and does not require retraining across problem instances of different sizes. The proposed hybrid mechanism is able to improve the performance of baseline metaheuristic algorithms. Our approach not only enhances the solution quality for standard CVRP benchmarks but also proves scalability on very large-scale instances with up to 30,000 customer nodes. Experimental evaluations on benchmark datasets show that the proposed hybrid mechanism is capable of improving different baseline algorithms, achieving better quality of solutions under similar settings.
Open AccessHybrid Node-Destroyer Model with Large Neighborhood Search for Solving the Capacitated Vehicle Routing Problem
In this research, we propose an iterative learning hybrid optimization solver developed to strengthen the performance of metaheuristic algorithms in solving the Capacitated Vehicle Routing Problem (CVRP). The iterative hybrid mechanism integrates the proposed Node-Destroyer Model, a machine learning hybrid model that utilized Graph Neural Networks (GNNs) such identifies and selects customer nodes to guide the Large Neighborhood Search (LNS) operator within the metaheuristic optimization frameworks. This model leverages the structural properties of the problem and solution that can be represented as a graph, to guide strategic selections concerning node removal. The proposed approach reduces operational complexity and scales down the search space involved in the optimization process. The hybrid approach is applied specifically to the CVRP and does not require retraining across problem instances of different sizes. The proposed hybrid mechanism is able to improve the performance of baseline metaheuristic algorithms. Our approach not only enhances the solution quality for standard CVRP benchmarks but also proves scalability on very large-scale instances with up to 30,000 customer nodes. Experimental evaluations on benchmark datasets show that the proposed hybrid mechanism is capable of improving different baseline algorithms, achieving better quality of solutions under similar settings.
- Open Access
Study of Robust Features in Formulating Guidance for Heuristic Algorithms for Solving the Vehicle Routing Problem
The Vehicle Routing Problem (VRP) is a complex optimization problem with numerous real-world applications, mostly solved using metaheuristic algorithms due to its $\mathcal{NP}$-Hard nature. Traditionally, these metaheuristics rely on human-crafted designs developed through empirical studies. However, recent research shows that machine learning methods can be used the structural characteristics of solutions in combinatorial optimization, thereby aiding in designing more efficient algorithms, particularly for solving VRP. Building on this advancement, this study extends the previous research by conducting a sensitivity analysis using multiple classifier models that are capable of predicting the quality of VRP solutions. Hence, by leveraging explainable AI, this research is able to extend the understanding of how these models make decisions. Finally, our findings indicate that while feature importance varies, certain features consistently emerge as strong predictors. Furthermore, we propose a unified framework able of ranking feature impact across different scenarios to illustrate this finding. These insights highlight the potential of feature importance analysis as a foundation for developing a guidance mechanism of metaheuristic algorithms for solving the VRP.
Open AccessStudy of Robust Features in Formulating Guidance for Heuristic Algorithms for Solving the Vehicle Routing Problem
The Vehicle Routing Problem (VRP) is a complex optimization problem with numerous real-world applications, mostly solved using metaheuristic algorithms due to its $\mathcal{NP}$-Hard nature. Traditionally, these metaheuristics rely on human-crafted designs developed through empirical studies. However, recent research shows that machine learning methods can be used the structural characteristics of solutions in combinatorial optimization, thereby aiding in designing more efficient algorithms, particularly for solving VRP. Building on this advancement, this study extends the previous research by conducting a sensitivity analysis using multiple classifier models that are capable of predicting the quality of VRP solutions. Hence, by leveraging explainable AI, this research is able to extend the understanding of how these models make decisions. Finally, our findings indicate that while feature importance varies, certain features consistently emerge as strong predictors. Furthermore, we propose a unified framework able of ranking feature impact across different scenarios to illustrate this finding. These insights highlight the potential of feature importance analysis as a foundation for developing a guidance mechanism of metaheuristic algorithms for solving the VRP.
- 18 June 2024
- Published by Springer Nature
- p. 112-125
Combining Neighborhood Search with Path Relinking: A Statistical Evaluation of Path Relinking Mechanisms
- 18 June 2024
- Published by Springer Nature
- p. 112-125
Combining Neighborhood Search with Path Relinking: A Statistical Evaluation of Path Relinking Mechanisms
- 11 March 2024
- Published by Elsevier in European Journal of Operational Research
- Vol. 317 (3) , 696-717
Instance generation tool for on-demand transportation problems
- 11 March 2024
- Published by Elsevier in European Journal of Operational Research
- Vol. 317 (3) , 696-717
Instance generation tool for on-demand transportation problems
- Open Access
- 20 November 2023
- Published by Elsevier in International Journal of Artificial Intelligence in Education
- Vol. 34 (3) , 1102-1137
Fully Individualized Curriculum with Decaying Knowledge, a New Hard Problem: Investigation and Recommendations
- 20 November 2023
- Published by Elsevier in International Journal of Artificial Intelligence in Education
- Vol. 34 (3) , 1102-1137
Open AccessFully Individualized Curriculum with Decaying Knowledge, a New Hard Problem: Investigation and Recommendations
- 12 October 2023
- Published by Elsevier in Transportation Research Part C: Emerging Technologies
- Vol. 157
The electric on-demand bus routing problem with partial charging and nonlinear function
- 12 October 2023
- Published by Elsevier in Transportation Research Part C: Emerging Technologies
- Vol. 157
The electric on-demand bus routing problem with partial charging and nonlinear function